APPARATUS AND METHODS FOR PHENOTYPING PLANTS
20180276818 ยท 2018-09-27
Inventors
Cpc classification
H04N5/2624
ELECTRICITY
H04N23/684
ELECTRICITY
H04N5/2625
ELECTRICITY
H04N23/6812
ELECTRICITY
International classification
H04N5/262
ELECTRICITY
Abstract
An apparatus and methods for imaging and phenotyping of plants. Plants are placed in an enclosure having an access door and a floor, A turntable is disposed on the floor of the enclosure. A side view camera assembly attached to a frame of the apparatus is configured to capture images of a plant placed on the turntable. An overhead view camera assembly attached to the frame above the turntable is configured to capture images of a plant placed on the turntable.
Claims
1. An apparatus for capturing one or more images of one or more trays of plants comprising: a frame; an enclosure having a floor; a drawer attached to the floor of the enclosure; a camera assembly attached to the frame above the drawer, the camera assembly configured to capture one or more images of a tray of plants placed on the drawer; a computer electronically connected to the camera assembly; and a monitor connected to the computer, the monitor configured to display information related to the tray of plants placed on the drawer.
2. The apparatus of claim 1 wherein the overhead view camera assembly comprises a vibration sensor configured to detect vibration of a movable camera.
3. A method for capturing one or more images of one or more trays of plants comprising: providing a tray of plants to be imaged; placing the tray on a platform of a plant imaging device; reading a unique identifier disposed on the tray; positioning a camera having an image sensor installed on the plant imaging device such that the camera is directly above a tray section; capturing an image of the tray section; repeating the steps of positioning the camera and capturing an image until all tray sections have been imaged; and creating a composite image comprising the captured images.
4. The method of claim 3 further comprising displaying the composite image to a user of the plant imaging device.
5. The method of claim 4 further comprising recapturing the overhead view image.
6. The method of claim 3 further comprising analyzing the composite image.
7. The method of claim 6 wherein the analysis comprises color analysis.
8. The method of claim 6 further comprising presenting results of analyzing the composite image to an operator of the plant imaging device immediately after analyzing the composite image.
Description
BRIEF DESCRIPTION OF THE DRAWINGS
[0008] Having thus described the invention in general terms, reference will now be made to the accompanying drawings, which are not necessarily drawn to scale, and wherein:
[0009]
[0010]
[0011]
[0012]
[0013]
[0014]
[0015]
[0016]
[0017]
[0018]
[0019]
[0020]
[0021]
[0022]
[0023]
[0024]
[0025]
[0026]
[0027]
[0028]
[0029]
[0030]
DETAILED DESCRIPTION
[0031] Some embodiments of the present invention will now be described more fully hereinafter with reference to the accompanying drawings, in which some, but not all, embodiments of the invention are shown. Indeed, various embodiments of the invention may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided so that this disclosure will satisfy applicable legal requirements. Like reference numerals refer to like elements throughout. Some components of the apparatus are not shown in one or more of the figures for clarity and to facilitate explanation of embodiments of the present invention.
[0032] In accordance with one embodiment,
Frame 10
[0033] As shown in
[0034] The frame 10 may be constructed in two or more separable parts to enable the apparatus 1 to be transported more easily. For example, the frame 10 may comprise a base section, center section, and top section that can be separated from each other to allow for movement through a standard doorway.
Panels 20
[0035] As shown in
Access Door 30
[0036] As shown in
Turntable 100
[0037] As shown in
[0038] Turntable 100 may comprise a rotatable platform driven by a motor that is in communication with the computer 600. Turntable 100 may be mounted on the floor of the apparatus 1. Markings of various shapes may be drawn on the platform of turntable 100 to allow for consistent plant placement. For example, concentric circles may be drawn on the platform of turntable 100 to allow for consistent placement of plants in round pots, or rectangular or square markings may be present to allow for consistent placement of plants in rectangular or square pots.
Side View Camera Subassembly 200
[0039] As shown in
[0040] Side view camera subassembly 200 further comprises a camera platform 215 that provides support for components that provide vertical movement for side view camera subassembly 200. Camera platform 215 comprises a first end that engages horizontal rail 205 and horizontal lead screw 220. Camera platform 215 is movable along the length of horizontal rail 205 and horizontal lead screw 220. Rotation of the shaft of motor 225 in a first direction causes camera platform 215 to move horizontally along the length of horizontal rail 205 and horizontal lead screw 220 toward the plant. Rotation of the shaft of motor 225 in a second direction that is opposite from the first direction causes camera platform 215 to move horizontally along the length of horizontal rail 205 and horizontal lead screw 220 away from the plant.
[0041] Mounted to camera platform 215 is a vertical rail 230 and a vertical lead screw 235. Vertical lead screw 235 is situated adjacent to and generally parallel to vertical rail 230. One end of vertical lead screw 235 engages a motor 240 that is controlled by computer 600. Vertical rail 230, vertical lead screw 235, and motor 240 provide for vertical motion for side view camera subassembly 200. A camera mount 245 provides a stable base for mounting side view camera 210. Camera mount 245 comprises a first surface that engages vertical rail 230 and vertical lead screw 235, and a second surface upon which side view camera 210 is securely connected. Camera mount 245 is movable along the lengths of vertical rail 230 and vertical lead screw 235. Rotation of the shaft of motor 240 in a first direction causes camera mount 245 to move vertically along the lengths of vertical rail 230 and vertical lead screw 235 toward horizontal rail 205 and horizontal lead screw 220. Rotation of the shaft of motor 240 in a second direction that is opposite from the first direction causes camera mount 245 to move vertically along the lengths of vertical rail 230 and vertical lead screw 235 away from horizontal rail 205 and horizontal lead screw 220.
[0042] Side view camera 210 may be a digital RGB camera, and may comprise an industrial camera or a consumer camera. While the above description of side view camera subassembly 200 anticipates that a digital RGB camera, other instruments may be substituted for side view camera 210. For example, a hyperspectral, near infrared or thermal imagers may be substituted for side view camera 210. The interchangeable nature of instruments in the apparatus 1 allows for all measurements to be performed in one enclosure, which prevents inconsistencies that would be caused by moving the plant or placement in a different environment for imaging. Because the same view can be captured using multiple instruments without moving the plant, a researcher can be sure that the same feature is examined across multiple instruments.
[0043] A vibration sensor, such as an accelerometer, may be connected to side view camera 210 to detect vibration after movement of side view camera 210 to ensure motion does not impact the clarity of image. Motion can come from intended motion of the camera and also vibrations from the environment. Measurement from the vibration sensor can be used to delay imaging and to correct for motion through image analysis methods.
[0044] Each camera subassembly can be made more independent by co-locating a computer to control the motors and camera, abstracting the implementation. The interface with the main computer is thus abstracted and the replacement and upgrading of camera systems simplified.
Angle Mounted Camera Subassembly 400
[0045] As shown in
[0046] As in embodiments using the side view camera subassembly 200, the angle mounted camera subassembly 400 allows images of the side of the plant to be captured, which enables measurement of the width of the plant as well as measurement of other properties of the plant. However, the angle mounted camera subassembly 400 components are mounted at an angle, providing both horizontal and vertical camera movement while moving the camera in a single axis. By reducing motion to a single axis, fewer components are needed, the design is simpler, and cost is reduced.
[0047] Angle mounted camera subassembly 400 comprises a rail 405 and a lead screw 410. Rail 405 and lead screw 410 are securely mounted at an angle between zero and 90 degrees to frame 10. For example, rail 405 and lead screw 410 may be mounted to frame 10 at a 45 degree angle. Lead screw 410 is situated adjacent to and generally parallel to rail 405. One end of lead screw 410 engages a motor 415 that is controlled by computer 600. Rail 405, lead screw 410, and motor 415 provide for both horizontal and vertical motion for angle mounted camera subassembly 400. A camera mount 420 provides a stable base for mounting a camera 425. Camera mount 420 comprises a first surface that engages rail 405 and lead screw 410, and a second surface upon which camera 425 is securely connected. Camera mount 420 is movable along the lengths of rail 405 and lead screw 410. Rotation of the shaft of motor 415 in a first direction causes camera mount 420 to move along the lengths of rail 405 and lead screw 410 down and toward the plant. Rotation of the shaft of motor 415 in a second direction that is opposite from the first direction causes camera mount 420 to move along the lengths of rail 405 and lead screw 410 upward and away from the plant.
[0048] A vibration sensor, such as an accelerometer, may be connected to camera 425 to detect vibration after movement of camera 425.
Corner View Camera Subassembly 450
[0049] As shown in
[0050] Corner view camera subassembly 450 comprises a corner camera support 460 that provides a stable base for mounting corner view camera 470. In one embodiment, corner camera support 460 comprises a stationary structure that supports corner view camera 470 in a fixed position near an upper corner of apparatus 1. When corner view camera 470 is stationary, the complexity, weight, and cost of apparatus 1 are reduced, and vibration that may occur with a movable camera is eliminated. In other embodiments, corner view camera subassembly 450 comprises a structure similar to angle mounted camera subassembly 400 that is mounted near an upper corner of apparatus 1, allowing corner view camera 470 to move toward or away from the plant to be imaged.
Overhead View Camera Subassembly 300
[0051] As shown in
[0052] A vibration sensor, such as an accelerometer, may be connected to overhead view camera 325 to detect vibration after movement of overhead view camera 325.
Lighting Subassembly 500
[0053] As shown in
[0054] In one embodiment, each lighting unit 515 comprises a single row of light vertically arranged light emitting diodes (LEDs). Though use of a single row of vertically arranged LEDs per lighting unit 515, it is possible to avoid lighting artifacts within the enclosure of the apparatus 1 and achieve uniform lighting conditions. If multiple rows of LEDs were used, constructive and deconstructive effects would occur, causing bright or dark bands across the area to be imaged. In other embodiments, incandescent bulbs or a mix of LEDs and incandescent lighting may be used. The choice of lighting can be optimized for the camera sensor.
[0055] As shown, lighting subassembly 500 comprises two lighting units 515 situated adjacent to side view camera subassembly 200. However, lighting units 515 may be mounted in other locations of apparatus 1. If shadows are present in the area to be imaged, additional lighting units 515 may be mounted adjacent to the overhead view camera subassembly 300.
[0056] Lighting subassembly 500 may further comprise one or more baffles 510 attached to each lighting unit 515. Baffles 510, often referred to as barndoors by photographers, are configured to partially cover lighting units 515, focusing the light provided by each lighting unit 515 onto a particular part of the area to be imaged.
[0057] Lighting subassembly 500 can be controlled by computer 600. For example, computer 600 may control the on/off functionality or light intensity of the lighting units 515.
Computer 600
[0058] As shown in
[0059] The computer 600 may communicate electronically with a remote server that contains experiment information. For example, the remote server may contain the species, images captured in the past, last measured height and width, growing conditions, and other properties for each plant under study.
[0060] Each major subassembly of the apparatus 1 may have a separate controller in communication with the computer 600. A power switch controlled by computer 600 capable of cycling power to motors 225, 240, 315, and 415 and cameras 210, 325, and 425 may be included. If a component of the apparatus 1 experiences a fault, the computer 600controlled power switch can be cycled to return apparatus 1 to a known state.
[0061] In one embodiment, the computer 600 comprises a commercially available personal computer. In another embodiment, the computer 600 comprises a commercially available panel computer with an integrated monitor 700 and touchscreen
Monitor 700
[0062] As shown in
Barcode Reader
[0063] A unique barcode may be placed on the pot of each plant under study, providing a way to identify each plant under study and retrieve images, measurements, and other data associated with each plant. A barcode reader may be provided to identify plants being placed in the apparatus 1 for imaging. In one embodiment, barcode reader is a handheld reader that a user can use to manually scan each plant as it is placed in the apparatus 1. In an alternative embodiment, the side view camera 210 may capture a unique barcode disposed on the plant, and determine the identity of the plant using image processing techniques.
[0064] The barcode could be a 1D or 2D barcode, or QC code, or RFID, or other similar device for uniquely identifying a plant for tracking purposes.
Alternative Lighting and Imaging Subassemblies
[0065] As shown in
[0066] As shown in
[0067] As shown in
[0068] In accordance with another alternative embodiment,
Drawer 105
[0069] As shown in FIG, 19, drawer 105 is configured to accept a tray of plants to be placed inside apparatus 1 for imaging. Drawer 105 may comprise a generally rectangular platform mounted on the floor of the apparatus 1 using mounting hardware that allows drawer 105 to be slid between fully extended and fully retracted positions. For ease of loading and operator ergonomics, drawer 105 may be fully extended out of the apparatus 1 for loading of the tray prior to imaging. After the tray is placed on drawer 105, drawer 105 is pushed into the apparatus 1 until the fully retracted position is reached. A sensor 106 detects when drawer 105 has reached the fully retracted position, ensuring consistency in tray position. Optional inserts or placement guides may be inserted into drawer 105 to accept and ensure consistent positioning of small pots or petri dishes. In the case of petri dishes, a back illuminated holder may be inserted into drawer 105 to allow imaging of the contents of the petri dishes.
Overhead View Camera Subassembly 330
[0070] As shown in
[0071] Rails 345 engage a camera mount 350 that is capable of moving along the length of rails 345. Camera mount 350 provides a stable base for mounting overhead view camera 355. Motion of camera mount 350 may be accomplished using one or more motor-driven belts that engage one or both ends of camera mount 350. Actuating the motor in a first direction causes camera mount 350 to move in a first direction along rails 345, and actuating the motor in a second and opposite direction causes camera mount 350 to move in a second and opposite direction along rails 345. The distance of camera mount 350 to drawer 105 remains generally the same as camera mount 350 travels from one end of rails 345 to the other end, ensuring consistency as all of the plants in the tray are imaged.
[0072] During imaging of the tray, camera 355 moves to directly above each flat insert of petri dish to maximize coverage of the image sensor of camera 355. By maximizing sensor coverage, resolution of the measurements that can be obtained from the resulting images is also maximized. Each insert or petri dish may separately imaged, and the images may be processed to form one composite image of potentially hundreds of megapixels. Camera 355 may use exchangeable camera lenses that accommodate larger and smaller inserts.
[0073] A vibration sensor, such as an accelerometer, may be connected to overhead view camera 355 to detect vibration after movement of overhead view camera 355.
Barcode Reader
[0074] A unique barcode may be placed on each tray of plants under study, providing a way to identify each tray of plants under study and retrieve images, measurements, and other data associated with each tray of plants. A barcode reader may be provided to identify plants being placed in the apparatus 1 for imaging. In one embodiment, barcode reader 800 is a handheld reader that a user can use to manually scan each tray as it is placed in the apparatus 1.
[0075] The barcode could be a 1D or 2D barcode, or QC code, or MD, or other similar device for uniquely identifying a tray of plants for tracking purposes.
Methods
[0076] As shown in
[0077] Alternatively, if the identity of the plant is to be identified using the side view camera 210, the access door 30 is closed once the plant is in place inside the apparatus 1, and the computer 600 then transmits instructions to rotate turntable 100 until the barcode on the pot of the plant is visible to the side view camera 210. The unique plant identifier is communicated to the computer 600.
[0078] Alternatively, the plant could be identified with a barcode reader permanently attached to the apparatus by rotating the plant until the barcode is found, or with MD immediately upon entering imaging chamber without any need for rotation.
[0079] At step 1030, once the unique plant identifier has been determined, either by handheld barcode reader or side view camera 210, the computer 600 displays information corresponding to the unique plant identifier on the monitor 700. The most recently captured image and analysis for the plant may be displayed on the monitor 700, If the access door 30 is still open at this time, the computer 600 displays a message on the monitor 700 to close the access door 30.
[0080] With the access door 30 closed and the plant placed on the turntable 100 and identified, the apparatus 1 can commence imaging the plant at step 1040. Imaging may begin within obtaining a side view image of the plant or an overhead view image of the plant or a corner view of the plant, or obtaining the side view, overhead view ,and corner view images, and potentially multiple angled views, may occur simultaneously.
[0081] After each image is obtained, the raw image collected and analyzed results are immediately presented to a user of the apparatus 1 at step 1050. If a problem has occurred with apparatus 1 or the plant, the user can immediately fix the problem and obtain a new image. The ability to immediately remedy issues that occur with plant imaging is important since plants continue to grow between imaging sessions, and at the end of an experiment may be destroyed. If missing or flawed images or analysis are not discovered immediately, the opportunity to correct such issues is lost. Processing of images may be performed on the apparatus 1 or sent to a server or cluster of computers at step 1060 for processing with the goal of the perception of immediate feedback to the apparatus 1 operator. At step 1070, the images and other analysis may be viewed on a reporting website.
[0082] After the plant has been identified in step 1020, the operator can enter notes about the current plant being imaged at the apparatus 1.
Obtaining a Side View Image
[0083] As shown in
[0084] In one embodiment, to obtain a side view image using side view camera subassembly 200, the horizontal and vertical position of side view camera 210 are adjusted at step 1120. To position side view camera 210 horizontally, computer 600 instructs a motor controller controlling side view horizontal motor 225 to rotate the shaft of side view horizontal motor 225. Rotation of side view horizontal motor 225 causes rotation of side view horizontal lead screw 220, which causes side view camera platform 215 to move along the lengths of side view horizontal rail 205 and side view horizontal lead screw 220. If the shaft of side view horizontal motor 225 is rotated in a first direction, then side view camera platform 215 and side view camera 210 move toward the plant. If the shaft of side view horizontal motor 225 is rotated in a second direction that is opposite to the first direction, the side view camera platform 215 and side view camera 210 move away from the plant.
[0085] To position side view camera 210 vertically, computer 600 instructs a motor controller controlling side view vertical motor 240 to rotate the shaft of side view vertical motor 240. Rotation of side view vertical motor 240 causes rotation of side view vertical lead screw 235, which causes side view camera mount 245 to move along the lengths of side view vertical rail 230 and side view vertical lead screw 235. If the shaft of side view vertical motor 240 is rotated in a first direction, then side view camera mount 245 and side view camera 210 moves toward horizontal rail 205 and horizontal lead screw 220. If the shaft of side view vertical motor 240 is rotated in a second direction that is opposite to the first direction, the side view camera mount 245 and side view camera 210 moves away from the horizontal rail 205 and horizontal lead screw 220.
[0086] In an alternative embodiment, a side view image is obtained using angle mounted camera subassembly 400. Through use of angle mounted camera subassembly 400, the optimum horizontal and vertical position of camera 425 is achieved by moving camera 425 along a single axis of motion. To position camera 425 using angle mounted camera subassembly 400, computer 600 instructs a motor controller controlling angle mounted motor 415 to rotate the shaft of angle mounted motor 415. Rotation of angle mounted motor 415 causes rotation of angled lead screw 410, which causes angled camera mount 420 to move along the lengths of angled rail 405 and angled lead screw 410. If the shaft of angle mounted motor 415 is in a first direction, camera mount 420 and camera 425 move along the lengths of rail 405 and lead screw 410 down and toward the plant. Rotation of the shaft of motor 415 in a second direction that is opposite from the first direction causes camera mount 420 and camera 425 to move along the lengths of rail 405 and lead screw 410 upward and away from the plant.
[0087] Once the optimum horizontal and vertical position of side view camera 210 is achieved, the side view camera 210 is focused at step 1130. The side view camera 210 may be focused using an auto-focus feature of side view camera 210 or camera 425 may be used to obtain a clear side view image of the plant. Focusing can also be implemented using a mechanized lens, the movement of the side view camera platform 215, modification of a standard lens to turn the focus ring, or use of a lens such as a liquid lens.
[0088] Movement of side view camera 210 or camera 425 can cause vibration that would result in blurry images. If side view camera 210 or camera 425 is equipped with a vibration sensor, such as an accelerometer, to detect vibration, and vibration is detected by the vibration sensor at step 1140, capture of the side view image may be delayed while vibration is present (detected by accelerometer or other vibration sensor) or for a period of time at step 1150 to allow vibration to cease, or the motion measured by the vibration sensor in each axis can be used to correct motion artifacts during image analysis using image processing software. If side view camera 210 or camera 425 is not equipped with a vibration sensor to detect vibration, capture of the side view image may be delayed for a period of time after each movement of side view camera platform 215, side view camera mount 245, or angled camera mount 420 to allow for any vibration to cease before imaging is performed at step 1160.
[0089] By adjusting the horizontal and vertical position of side view camera 210 or camera 425 as described above, side view camera 210 or camera 425 is placed at a distance from the plant that ensure that the entire width and height of the plant are captured in resulting side view images, and that maximum coverage of the image sensor of side view camera 210 or camera 425 is achieved. By maximizing sensor coverage, resolution of the measurements that can be obtained from the resulting images is also maximized.
[0090] Several methods may be used to ensure that optimal positioning of side view camera 210 or camera 425 has been achieved. In one embodiment, the user may instruct computer 600 via the touchscreen on monitor 700 to move side view camera 210 or camera 425 based on a visual inspection of an image of the plant. In another embodiment, side view camera 210 or camera 425 may be moved away from the plant if image processing software running on computer 600 determines that plant pixels are present along the outermost edges of the resulting image. Likewise, side view camera 210 or camera 425 may be moved toward the plant if image processing software running on computer 600 determines that there are no plant pixels present within a predefined distance from the outermost edge of the resulting image. In another embodiment, prior image stored for the plant may be retrieved by computer 600 and used to obtain the optimal position of side view camera 210 or camera 425. In this embodiment, the age, species, approximate size, and prior positioning information is used to adjust the position of side view camera 210 or camera 425. Further, the position of side view camera 210 can be further adjusted if visual inspection or image processing software indicates that the side view camera 210 is now too close due to increased plant growth that has occurred. The position of the camera 210 is stored and because of previous calibration can be used to convert measurement of pixels into consistent units across any number of images and positions.
[0091] After obtaining a side view image of the widest part of the plant, turntable 100 may be rotated to a different position, and the steps above repeated to obtain additional side views of the plant. For example, the plant may be rotated 90 degrees from the widest view and imaged. Alternatively, repeated side view images may be captured as turntable 100 is rotated to obtain a 360 degree view of the plant.
[0092] After each side view image is obtained, the raw image collected and analyzed results are immediately presented to a user of the apparatus 1. If a problem has occurred with apparatus 1 or the plant, the user can immediately fix the problem and obtain a new side view image. The ability to immediately remedy issues that occur with plant imaging is important since plants continue to grow between imaging sessions, and at the end of an experiment may be destroyed. If missing or flawed images or analysis are not discovered immediately, the opportunity to correct such issues is lost. Processing of images may be performed on the apparatus 1 or sent to a server of cluster of computers processing with the goal of the perception of immediate feedback to the apparatus 1 operator.
[0093] A time series of images and measurements can be more useful to a researcher than a single image or set of measurements. Therefore, the process of obtaining a side view image and associated measurements described above may be repeated multiple times during the life of the plant. As new images and measurements for each plant are captured, previous images and measurements may be presented to the operator using monitor 700, and the operator may use monitor 700 to review current and past images and measurements. By accessing previous images from the server, a researcher can more consistently orient the plant with respect to previous times, and more appropriately compare growth of the plant. In addition, the operator is able to review the experiment's progress, identify plants with higher or lower performance than the rest of the experiment population, review treatments for those showing improved or degraded performance compared to others or controls, and gain a better understanding of the progression of the experiment. Final results are improved by discovering mistakes early, and the operator may decide to alter or end the experiment, because a conclusion is possible early or a flaw is discovered, either result saving time. Instead of reviewing the results of the experiment potentially weeks after the experiment is complete, the operator is reviewing data as new data is collected. Time that the operator would normally be waiting is now used to consider the experiment possibly resulting an earlier conclusion than would be possible if results were only reviewed after experiment completion. Time during which the operator would otherwise be waiting is productively used.
[0094] After capturing the side view image, phenotypic measurements, including width, height, leaf count, leaf length, leaf width, leaf angle, tassel length, tassel angle, silk count, flower count, flower size, seed count, organ size, organ color, plant damage, plant health, plant disease, pest damage, pest infestation, chemical damage, presence of non-target plant species, ratio of plant species, or any other characteristic of agronomic, ornamental, or commercial interest. can be obtained from the captured image at step 1170. These traits may be derived from a single image or instrument type, or may be derived from a combination of image or instrument types. Such phenotypic measurements may be determined using software running on computer 600 or a remote server.
Obtaining a Corner Image
[0095] If corner view camera subassembly 450 is present, a corner view image may be captured in a similar manner to the method of obtaining a side view image 1100.
Obtaining an Overhead Image
[0096] As shown in
[0097] In one embodiment, to obtain an overhead view image using overhead view camera subassembly 300, the vertical position of overhead view camera 325 is adjusted at step 1220. To position overhead view camera 325, computer 600 instructs a motor controller controlling overhead view motor 315 to rotate the shaft of overhead view motor 315. Rotation of overhead view motor 315 causes rotation of overhead view lead screw 310, which causes overhead camera mount 320 to move vertically along the lengths of overhead view rail 305 and overhead view lead screw 310. If the shaft of overhead view motor 315 is rotated in a first direction, then overhead camera mount 320 and overhead camera 325 move toward the plant. If the shaft of overhead view motor 315 is rotated in a second direction that is opposite to the first direction, the overhead view camera mount 320 and overhead view camera 325 move away from the plant.
[0098] Once the optimum position of overhead view camera 325 is achieved, the overhead view camera 325 is focused at step 1230. To focus overhead view camera 325, an auto-focus feature of overhead view camera 325 may be used to obtain a clear overhead view image of the plant. Focusing can also be implemented using a mechanized lens, the movement of the overhead camera mount 320, modification of a standard lens to turn the focus ring, or use of a lens such as a liquid lens.
[0099] Movement of overhead view camera 325 can cause vibration that would result in blurry images. If overhead view camera 325 is equipped with a vibration sensor, such as an accelerometer, to detect vibration, and vibration is detected by the vibration sensor at step 1240, capture of the overhead view image may be delayed at step 1250 while vibration is present (detected by accelerometer or other vibration sensor) or for a period of time to allow vibration to cease, or the motion measured by the vibration sensor in each axis can be used to correct motion artifacts during image analysis using image processing software. If overhead view camera 325 is not equipped with a vibration sensor to detect vibration, capture of the overhead view image may be delayed for a period of time after each movement of overhead camera mount 320 to allow for any vibration to cease before imaging is performed.
[0100] By adjusting the position of overhead view camera 325 as described above, overhead view camera 325 is placed at a distance from the plant that ensure that the entire plant is captured in resulting overhead view images, and that maximum coverage of the image sensor of overhead view camera 325 is achieved. By maximizing sensor coverage, resolution of the measurements that can be obtained from the resulting images is also maximized.
[0101] Several methods may be used to ensure that optimal positioning of overhead view camera 325 has been achieved. In one embodiment, the user may instruct computer 600 via the touchscreen on monitor 700 to move overhead view camera 325 based on a visual inspection of an image of the plant. In another embodiment, overhead view camera 325 may be moved away from the plant if image processing software running on computer 600 determines that plant pixels are present along the outermost edges of the resulting image. Likewise, overhead view camera 325 may be moved toward the plant if image processing software running on computer 600 determines that there are no plant pixels present within a predefined distance from the outermost edge of the resulting image. In another embodiment, prior image stored for the plant may be retrieved by computer 600 and used to obtain the optimal position of overhead view camera 325. In this embodiment, the age, species, approximate size, and prior positioning information is used to adjust the position of overhead view camera 325. Further, the position of overhead view camera 325 can be further adjusted if visual inspection or image processing software indicates that the overhead view camera 325 is now too close due to increased plant growth that has occurred.
[0102] After each overhead view image is obtained at step 1260, the raw image collected and analyzed results are immediately presented to a user of the apparatus 1. If a problem has occurred with apparatus 1 or the plant, the user can immediately fix the problem and obtain a new overhead view image. The ability to immediately remedy issues that occur with plant imaging is important since plants continue to grow between imaging sessions, and at the end of an experiment may be destroyed. If missing or flawed images or analysis are not discovered immediately, the opportunity to correct such issues is lost.
[0103] A time series of images and measurements can be more useful to a researcher than a single image or set of measurements. Therefore, the process of obtaining an overhead view image and associated measurements described above may be repeated multiple times during the life of the plant. As new images and measurements for each plant are captured, previous images and measurements may be presented to the operator using monitor 700, and the operator may use monitor 700 to review current and past images and measurements. By accessing previous images from the server, a researcher can more consistently orient the plant with respect to previous times, and more appropriately compare growth of the plant.
[0104] After capturing the overhead view image, phenotypic measurements, including width, height, leaf count, leaf length, leaf width, leaf angle, tassel length, tassel angle, silk count, flower count, flower size, seed count, organ size, organ color, plant damage, plant health, plant disease, pest damage, pest infestation, chemical damage, presence of non-target plant species, ratio of plant species, or any other characteristic of agronomic, ornamental, or commercial interest can be obtained from the captured image at step 1270. These traits may be derived from a single image or instrument type, or may be derived from a combination of image or instrument types. Such phenotypic measurements may be determined using software running on computer 600 or a remote server.
Imaging Using Apparatus 1
[0105] As shown in
[0106] At step 2030, once the unique plant identifier has been determined, the computer 600 displays information corresponding to the unique tray identifier on the monitor 700. The most recently captured image and analysis for the tray may be displayed on the monitor 700. If the access door 30 is still open at this time, the computer 600 displays a message on the monitor 700 to close the access door 30.
[0107] With the access door 30 closed and the tray placed on drawer 105 and identified, the apparatus 1 can commence imaging the tray or collecting other measurements at step 2040.
[0108] At step 2045, the collected image or images or other measurements are analyzed. Analysis may occur at the computer 600, or images or other data may be sent to a server or cluster of servers for analysis. During analysis, all images associated with the tray are assembled into a composite image. Image processing techniques may be used to remove background material from the images.
[0109] After the image is obtained, the raw image collected and analyzed results are immediately presented to a user of the apparatus 1 at step 2050. If a problem has occurred with the apparatus 1 or the tray, the user can immediately fix the problem and obtain a new image. The ability to immediately remedy issues that occur with plant imaging is important since plants continue to grow between imaging sessions, and at the end of an experiment may be destroyed. If missing or flawed images or analysis are not discovered immediately, the opportunity to correct such issues is lost. Processing of images may be performed on the apparatus 1 or sent to a server or cluster of computers at step 2060 for processing with the goal of the perception of immediate feedback to the apparatus 1 operator. The images and other analysis may be viewed on a reporting website.
[0110] After the tray has been identified in step 2020, the operator can enter notes about the current tray being imaged at the apparatus 1.
[0111] A time series of images and measurements can be more useful to a researcher than a single image or set of measurements. Therefore, method 3000 may be repeated multiple times during the life of the plants. As new images and measurements for each plant are captured, previous images and measurements may be presented to the operator using monitor 700, and the operator may use monitor 700 to review current and past images and measurements. By accessing previous images from the server, a researcher can more consistently orient the tray with respect to previous times, and more appropriately compare growth of the plants in the tray. In addition, the operator is able to review the experiment's progress, identify plants with higher or lower performance than the rest of the experiment population, review treatments for those showing improved or degraded performance compared to others or controls, and gain a better understanding of the progression of the experiment. Final results are improved by discovering mistakes early, and the operator may decide to alter or end the experiment, because a conclusion is possible early or a flaw is discovered, either result saving time. Instead of reviewing the results of the experiment potentially weeks after the experiment is complete, the operator is reviewing data as new data is collected. Time that the operator would normally be waiting is now used to consider the experiment possibly resulting an earlier conclusion than would be possible if results were only reviewed after experiment completion. Time during which the operator would otherwise be waiting is productively used.
[0112] After capturing the overhead view image, phenotypic measurements, including width, height, leaf count, leaf length, leaf width, leaf angle, tassel length, tassel angle, silk count, flower count, flower size, seed count, organ size, organ color, plant damage, plant health, plant disease, pest damage, pest infestation, chemical damage, presence of non-target plant species, ratio of plant species, or any other characteristic of agronomic, ornamental, or commercial interest can be obtained from the captured image. Measurements may also include biomass accumulation or color change (e.g. greenness or yellowing), and these measurements can be used as an indication of stress or treatment effect. These traits may be derived from a single image or instrument type, or may be derived from a combination of image or instrument types. Such phenotypic measurements may be determined using software running on computer 600 or a remote server.
Obtaining an Overhead Image Using Apparatus 1
[0113] In one embodiment, step 2040 may be accomplished using an overhead view camera subassembly 330. As shown in
[0114] Once the camera 355 is positioned, camera 355 is focused at step 3020. To focus camera 355, an auto-focus feature of camera 355 may be used to obtain a clear overhead view image of the plant. Focusing may also be implemented using a mechanized lens, modification of a standard lens to turn the focus ring, or use of a lens such as a liquid lens.
[0115] Movement of camera 355 can cause vibration that would result in blurry images. If camera 355 is equipped with a vibration sensor, such as an accelerometer, to detect vibration, and vibration is detected by the vibration sensor at step 3030, capture of the image may be delayed at step 3040 while vibration is present (detected by accelerometer or other vibration sensor) or for a period of time to allow vibration to cease, or the motion measured by the vibration sensor in each axis can be used to correct motion artifacts during image analysis using image processing software. If camera 355 is not equipped with a vibration sensor to detect vibration, capture of the overhead view image may be delayed for a period of time after each movement of camera mount 350 to allow for any vibration to cease before imaging is performed.
[0116] Once the camera is positioned, focused, and stable, camera 355 captures an image of the tray section at step 3050. Steps 3010, 3020, 3030, 3040, and 3050 are repeated until all sections of the tray positioned on drawer 105 have been imaged.
Color Analysis
[0117] Specialized instruments such as spectrometers including hyperspectral cameras are ideally suitable for measurement of color; however, these instruments and cameras are often prohibitively expensive. Instead, sometimes people attempt to measure color using an RGB camera. The challenge with using an RGB camera is several ranges of overlapping wavelengths are merged into each of the three (or four) color channels. The next challenge is how to perform the comparison after the convolution of color measurements. One could perform a measurement in RGB space, often Euclidean or other distance metrics within the RGB space, which may not relate to intuitive sense of color difference, and many will choose to convert to HSV color space and compare color by comparing the single dimension of Hue. However, both approaches include colors that are not seen in any plant, let alone a particular species of plant. By including colors that will never be measured the discrimination power is reduced. There are only so many color bins available, each channel may have 8-bit, 10-bit, 12-bit, 16-bit, or other discrete number of bins. To maximize the use of those bins for comparison, we develop color spaces that are tailored to a particular plant species and this can include the subtle effects of instrument lighting and optics. By collecting a number of images of the species of interest in various stages of development, stress, and diseases, we can measure the colors that do occur. Through datamining methods including PCA and PLS, we can develop models of color-spaces that improve the discriminatory power of the color analysis system.
[0118] Images collected for color analysis are normalized against reference images collected earlier. After the lights have warmed and stabilized, verified with the installed cameras, the reference images are used to normalize the colors in the images. This reduces the effect of changes in lighting and the environment.
Data Interface
[0119] All collected images are stored to a server. The server uses the same methods and additional methods as the instrument to analysis the images and convert the images to data. While the apparatus 1 may only be able to perform a subset of the analyses due to processing time constraints, the server can spend additional time. The goal of performing the analysis on the apparatus 1 is to provide immediate feedback to the operator while the goal of performing the analysis on the server is maximize the data quality even at the expense of time. However, faster processing may be required if the results from the instrument are needed for feeding back into the experiment. The raw image is stored to the server in addition to processed images. Storing the raw image allows future analysis to have full fidelity for reprocessing.
[0120] Results (images, data, and statistical summaries) are presented in a web interface accessible on desktop or mobile devices by researchers. Web service interfaces allow other system in an integrated laboratory easy access to the data to inform decisions by this and other instruments and informatics systems.
[0121] Operators are able to collect notes at the time of imaging or subsequently associate notes with images and data to better inform researchers at a future date of the state of the plant, apparatus 1, and environment at the time of imaging. Experiment notes need not be restricted to imaging but include any sort of free text data. The notes can be mined for additional structure or trends.
Integration of Handheld Accessory Instruments
[0122] To provide high resolution spectral or color data at a more affordable cost and speed than is possible with a hyperspectral camera, a spectrometer may be used. The challenges with a spectrometer include blocking out the ambient light and being consistent with the placement of the spectrometer probe.
[0123] To address the challenge of blocking ambient lighting, a handheld clip device comprising directed controlled source illumination and the spectrometer probe. Fiber cables may be used to separate the light and spectrometer from the handheld clip device. The clip effectively isolates the point of measurement from the ambient light, and bathes the area only in the controlled light. To address the challenge of consistent placement of the spectrometer, one might use a robot but this shifts the challenge to a more complicated motion control problem. Instead, a handheld spectrometer is monitored by the cameras already present in the apparatus 1 enclosure. The current location of the handheld spectrometer is shown superimposed on a blend of an image of the plant as it is at this point and an image of the how the plant was at the last point a tracked handheld spectrometer reading was taken. The combination of previous recording and tracking of current position allows the operator to effectively make decisions of placement that would be difficult for an automated machine. While a person would not remember the exact placement of the handheld spectrometer on each plant, apparatus 1 is particularly well suited to remember the placement by taking an image at the time when the spectrometer reading is taken. The combination of apparatus 1 and human operator results in speed and accuracy.
[0124] Many modifications and other embodiments of the invention will come to mind to one skilled in the art to which this invention pertains having the benefit of the teachings presented in the foregoing descriptions and the associated drawings. Therefore, it is to be understood that the invention is not to be limited to the specific embodiments disclosed and that modifications and other embodiments are intended to be included within the scope of the appended claims. Although specific terms are employed herein, they are used in a generic and descriptive sense only and not for purposes of limitation.