Implementations are described herein for localizing individual plants using high-elevation images at multiple different resolutions. A first set of high-elevation images that capture the plurality of plants at a first resolution may be analyzed to classify a set of pixels as invariant anchor points. High-elevation images of the first set may be aligned with each other based on the invariant anchor points that are common among at least some of the first set of high-elevation images. A mapping may be generated between pixels of the aligned high-elevation images of the first set and spatially-corresponding pixels of a second set of higher-resolution high-elevation images. Based at least in part on the mapping, individual plant(s) of the plurality of plants may be localized within one or more of the second set of high-elevation images for performance of one or more agricultural tasks.

In some examples, a system includes a memory configured to store a first image and a second image captured by one or more cameras mounted on one or more vehicles and store locations and orientations of the one or more cameras at times when the first and second images were captured. The system also includes processing circuitry configured to identify an existing landmark in the first and second images. The processing circuitry is also configured to determine a latitude, a longitude, and an altitude of the existing landmark based on the locations and orientations of the one or more cameras at the times when the images were captured. The processing circuitry is configured to create a file including the location of the existing landmark and pixel coordinates of the existing landmark in the first and second images.

A first 2D recording of a specified reference view of an object is captured by a camera and, starting from the first 2D recording, a user's starting location relative to the object is ascertained by a computer vision module. Starting from the origin of a coordinate system as the starting location of the camera, one or more specified and/or settable relative positions in the vicinity of the object and/or in the object are determined as one or more locations for the respective perspective of the camera for taking at least one second 2D recording. The respective location in an object view on a display of the camera is displayed by a respective first augmented reality marker on the ground and/or on the object. The alignment of the camera with regard to angle and rotation with the perspective corresponding to the respective location is performed in this case by second augmented reality markers as auxiliary elements.

A matching support apparatus 10 has a generation unit 11 that, in the case where a feature region indicating a facial feature of a person targeted for matching visible on the skin surface of the person targeted for matching is designated, using a reference face development image display region, displayed on the screen of a display device 30, for displaying a reference face development image generated based on three-dimensional data of a head serving as a reference, generates feature information relating to the designated feature region, a matching unit 12 that matches the feature information against matching information in which a matching-use face development image and matching-use feature region for each person registered in advance are associated with each other, and a selection unit 13 for selecting a person to serve as a candidate, based on a matching result.

A matching support apparatus 10 has a generation unit 11 that, in the case where a feature region indicating a facial feature of a person targeted for matching visible on the skin surface of the person targeted for matching is designated, using a reference face development image display region, displayed on the screen of a display device 30, for displaying a reference face development image generated based on three-dimensional data of a head serving as a reference, generates feature information relating to the designated feature region, a matching unit 12 that matches the feature information against matching information in which a matching-use face development image and matching-use feature region for each person registered in advance are associated with each other, and a selection unit 13 for selecting a person to serve as a candidate, based on a matching result.

The present invention makes it possible to cause an action recognizer capable of recognizing actions with high accuracy and with a small quantity of learning data to learn. An input unit 101 receives input of a learning video and an action label indicating an action of an object, a detection unit 102 detects a plurality of objects included in each frame image included in the learning video, a direction calculation unit 103 calculates a direction of a reference object, which is an object to be used as a reference among the plurality of detected objects, a normalization unit 104 normalizes the learning video so that a positional relationship between the reference object and another object becomes a predetermined relationship, and an optimization unit 106 optimizes parameters of an action recognizer to estimate the action of the object in the inputted video based on the action estimated by inputting the normalized learning video to the action recognizer and the action indicated by the action label.

An AI-based object recognition method is provided to recognize an object in a first image captured by a photographic device at a first shooting angle. The method includes: Step A, determining a difference value between the first shooting angle and a preset second shooting angle; Step B, converting the first image into a second image with a view angle of the second shooting angle when the difference value is greater than a preset value; and Step C, sending the second image to an artificial intelligence model for recognition.

When captured videos transmitted from n units of cameras are received, a server analyzes position information and characteristic information on a person appearing in each of captured videos as analysis results. A client terminal displays, on a screen, each of a k number of captured videos that are selected by a user operation. According to a selection of a search target person, the client terminal transmits, to the server, a search request for the search target person in a (k−1) number of captured videos other than the captured video in which the search target person appears. The server extracts the search target person and transmits the analysis result to the client terminal. The client terminal displays cut-out images of one or more persons identical with or similar to the search target person based on the analysis result.

A method for labeling an image, an electronic device, and a storage medium are provided. The method includes the following operations. A remote sensing image is acquired. A local binary image respectively corresponding to at least one building in the remote sensing image and direction angle information of a contour pixel located on a building contour in the local binary image are determined based on the remote sensing image. The direction angle information includes information of an angle between a contour edge where the contour pixel is located and a preset reference direction. A labeled image labeled with a polygonal contour of the at least one building in the remote sensing image is generated based on the local binary image respectively corresponding to the at least one building and the direction angle information.

A method for labeling an image, an electronic device, and a storage medium are provided. The method includes the following operations. A remote sensing image is acquired. A local binary image respectively corresponding to at least one building in the remote sensing image and direction angle information of a contour pixel located on a building contour in the local binary image are determined based on the remote sensing image. The direction angle information includes information of an angle between a contour edge where the contour pixel is located and a preset reference direction. A labeled image labeled with a polygonal contour of the at least one building in the remote sensing image is generated based on the local binary image respectively corresponding to the at least one building and the direction angle information.