A system and a method are disclosed for planning a path within a field for a farming machine. The system predicts boundaries of the field using a previously captured image of the field and generates a suggested route to be taken by a data collection device based on the predicted boundaries. As the data collection device travels along the suggested route, the data collection device collects location data associated with a current layout of the field. The location data may be labeled with obstructions encountered along the way. Based on the location data, the system identifies current boundaries of the field, which may be different from the predicted boundaries. The current boundaries are sent to a verification device to be verified. After the current boundaries have been verified, the path for the farming machine is planned.

An automatic steering system for a work vehicle includes: a steering travel apparatus for performing leftward turning based on a leftward steering amount with respect to a straight-forward direction and performing rightward turning based on a rightward steering amount with respect to a neutral position; a subject vehicle location calculator for calculating a location of a subject vehicle; a locational shifting calculator for calculating locational shifting from the travel route and the location of the subject vehicle; a vehicle body direction calculator for calculating a vehicle body direction that indicates a direction of a vehicle body; a directional shifting calculator for calculating directional shifting from the travel route and the vehicle body direction; a steering amount calculator for calculating a first steering amount, which is a steering amount for correcting the locational shifting and the directional shifting, based on the locational shifting and the directional shifting; and a steering amount limiter for limiting the first steering amount based on a second steering amount, which is the steering amount at present.

Implementations are disclosed for selectively operating edge-based sensors and/or computational resources under circumstances dictated by observation of targeted plant trait(s) to generate targeted agricultural inferences. In various implementations, triage data may be acquired at a first level of detail from a sensor of an edge computing node carried through an agricultural field. The triage data may be locally processed at the edge using machine learning model(s) to detect targeted plant trait(s) exhibited by plant(s) in the field. Based on the detected plant trait(s), a region of interest (ROI) may be established in the field. Targeted inference data may be acquired at a second, greater level of detail from the sensor while the sensor is carried through the ROI. The targeted inference data may be locally processed at the edge using one or more of the machine learning models to make a targeted inference about plants within the ROI.

Methods, apparatus, and systems for sensing or tracking relative position between objects or locations. A digital camera or imager captures one or more fiducials in its field of view. By calibration and processing, the imaged fiducials can be identified and distinguished from other objects and background in camera space, and position of imaged fiducials in camera space relative to a reference can be translated to position of the actual fiducials in physical space. In one example, the fiducials are IR LEDs.

A control system for a tillage implement broadly includes front and rear sensors, a leveling assembly, and a controller. The front sensor is positioned on a front of a central section, wherein the front sensor is configured to obtain height information indicative of a height of the front of the central section above a ground. The rear sensor is positioned on a rear of the central section, wherein the rear sensor is configured to obtain height information indicative of a height of the rear of the central section above the ground. The leveling assembly is configured to adjust a front to rear orientation of the central section. The controller is configured to receive the height information from the front sensor and the height information from the rear sensor, and to provide instructions to the leveling assembly to adjust the front to rear orientation of the central section based on the received height information.

System and techniques for adaptive color transformation to aid computer vision are described herein. Colors from an image are mapped into a multi-dimensional space to create a distribution of colors in the image. A line can be fit to the distribution. Here, the line includes an angle relative to a coordinate system of the multi-dimensional space. A transformation to colors can then be applied to the image based on the angle of the line. The transformation producing a reduced image where a color complexity of the original image is reduced.

Various embodiments of an apparatus, methods, systems and computer program products described herein are directed to an agricultural observation and treatment system and method of operation. The agricultural treatment system determines a vehicle pose of a vehicle as the vehicle moves along a path. The system identifies a first target agricultural object for treatment. Based on the determined vehicle pose, the system positions a treatment head of a first treatment unit such that a first projectile fluid may be emitted by the first treatment unit at the identified first target agricultural object. The system then causes an emitter to emit a fluid from the treatment head at the first target agricultural object.

In an embodiment, autonomous vehicles with global positioning systems (GPS) are used for field inspection to reduce fuel and labor costs and improve reliability with increased consistency in field crop inspection. A vehicle may be programmed to traverse a field while using sensors to detect objects and operating in a first image capture mode, for example, capturing low-resolution images of objects in the field, typically crops. Under program control, machine vision techniques are used with the low-resolution images to recognize crops, non-crop plant material or undefined objects. Under program control, location data is used to correlate recognized objects with digitally stored field maps to resolve whether a particular object is in a location at which crop planting is expected or not expected. Under program control, depending on whether an object in a low-resolution digital image is recognized as a crop, and whether the object is in an expected geo-location for crops, the vehicle may cease traversing temporarily and switch to a second image capture mode, for example, capturing a high-resolution image of the object, for use in disease analysis or classification, weed analysis or classification, alert notifications or other messages, or other processing. In this manner, a field may be rapidly traversed and imaged using coarse-level, rapid techniques that require lower processing resources, storage or memory, while automatically switching to execute special processing only when necessary to resolve unexpected objects or to perform operations such as disease classification that benefit from high-resolution images and more intensive use of processing resources, storage or memory.

A system that includes an agricultural tractor that includes a steering mechanism for steering at least one surface-engaging member so as to cause changes in a direction of movement of the agricultural tractor; and an agricultural implement that is towed behind the agricultural tractor. The system further includes at least one forward sensor for sensing one or more objects or conditions located forwardly of the agricultural tractor and at least one lateral sensor for sensing one or more objects or conditions that when sensed are located laterally of the agricultural tractor or implement. The system further includes a controller that acts in dependence on at least one output of the at least one lateral sensor to take account of a presence of the one or more objects or conditions sensed by the at least one lateral sensor.

An object identification method is disclosed. The method includes obtaining images of a target geographical area and telemetry information of an image-collection vehicle at a time of capture, analyzing each image to identify objects, and determining a position of the objects. The method further includes determining an image capture height, determining a position of the image using the capture height and the telemetry information, performing a transform on the image based on the capture height and the telemetry information, identifying the objects in the transformed image, determining first pixel locations of the objects within the transformed image, performing a reverse transform on the first pixel locations to determine second pixel locations in the image, and determining positions of the objects within the area based on the second pixel locations within the captured image and the determined image position.