Methods and systems for determining a path include detecting objects within a perspective image that shows a scene. Depth is predicted within the perspective image. Semantic segmentation is performed on the perspective image. An attention map is generated using the detected objects and the predicted depth. A refined top-down view of the scene is generated using the predicted depth and the semantic segmentation. A parametric top-down representation of the scene is determined using a relational graph model. A path through the scene is determined using the parametric top-down representation.

An apparatus including an interface and a processor. The interface may be configured to receive first video frames of a first field of view captured by a first capture device and second video frames of a second field of view captured by a second capture device. The first capture device and the second capture device may have a symmetrical orientation with respect to a vehicle. The fields of view may have an overlapping region. The processor may be configured to select the capture devices to operate as a stereo pair of cameras based on the symmetrical orientation, receive the video frames from the interface, detect an object located in the overlapping region, perform a comparison operation on the object based on the symmetrical orientation with respect to the video frames and determine a distance of the object from the vehicle in response to the comparison operation.

An AGV navigation device is provided, which includes a RGB-D camera, a plurality of sensors and a processor. When an AGV moves along a target route having a plurality of paths, the RGB-D camera captures the depth and color image data of each path. The sensors (including an IMU and a rotary encoder) record the acceleration, the moving speed, the direction, the rotation angle and the moving distance of the AGV moving along each path. The processor generates training data according to the depth image data, the color image data, the accelerations, the moving speeds, the directions, the moving distances and the rotation angles, and inputs the training data into a machine learning model for deep learning in order to generate a training result. Therefore, the AGV navigation device can realize automatic navigation for AGVs without any positioning technology, so can reduce the cost of automatic navigation technologies.

Systems, and method and computer readable media that store instructions for distance measurement, the method may include obtaining, from a camera of a vehicle, an image of a surroundings of the vehicle; searching, within the image, for an anchor, wherein the anchor is associated with at least one physical dimension of a known value; and when finding the anchor, determining a distance between the camera and the anchor based on, (a) the at least one physical dimension of a known value, (b) an appearance of the at least one physical dimension of a known value in the image, and (c) a distance-to-appearance relationship that maps appearances to distances, wherein the distance-to-appearance relationship is generated by a calibration process that comprises obtaining one or more calibration images of the anchor, and obtaining one or more distance measurements to the anchor.

A vehicle may include a camera configured to obtain an image of at least one surrounding vehicle; and a controller configured to determine object recognition data including at least one of full area data, wheel area data, and bumper area data of the surrounding vehicle from the image of the at least one surrounding vehicle, based on the object recognition data, to set a reference point in the at least one of the full area data, the wheel area data, and the bumper area data of the surrounding vehicle from the image of the surrounding vehicle, and to predict a driving speed of the surrounding vehicle based on a change in a position of the reference point.

A vehicle may include a camera; a storage; and a controller electrically connected to the camera and the storage. The controller may be configured to obtain a front image of the vehicle from the camera by controlling the camera, in a response to obtaining the front image, to identify image data corresponding to a license plate of a front vehicle positioned in front of the vehicle from the front image, to identify a decrease in a distance between the vehicle and the front vehicle based on the image data and reference distance information corresponding to at least one reference image data of at least one reference license plate stored in the storage, to identify a change in height of the license plate of the front vehicle based on the image data, and based on the decrease in the distance between the vehicle and the front vehicle and the change in height of the license plate of the front vehicle, to identify a speed bump located in front of the vehicle.

The obstacle detection system for a work vehicle enables an obstacle, present in the surroundings of a work vehicle, to be accurately detected. This obstacle detection system for a work vehicle has: a plurality of imaging devices that take images of the surroundings of a work vehicle; and an image processing device that performs, according to a time division system, an obstacle identification process for identifying an obstacle on the basis of the images from the plurality of imaging devices. The image processing device changes a to-be-processed cycle per unit time for the plurality of imaging devices in the time division system in accordance with the vehicle speed and the traveling direction of the work vehicle.

The present disclosure provides an obstacle detection method as well as an apparatus and a robot using the same. The method includes: obtaining, through the sensor module, image(s); detecting an obstacle image of an obstacle from the image(s) according to characteristic(s) of the obstacle; extracting image feature(s) of the obstacle; obtaining, through the sensor module, a position of the obstacle; associating the image feature(s) of the obstacle with the position of the obstacle; calculating a motion state a the obstacle based on the position information of the obstacle at different moments; and estimating the position of the obstacle in a detection blind zone of the robot based on the motion state. In such a manner, it is capable of providing more accurate position information of the obstacle in the detection blind zone, which is beneficial to the robot to plan a safe and fast moving path.

A method for obstacle identification for a vehicle. In order to provide means for reliable obstacle identification for a vehicle, the method includes: performing a first obstacle search with a radar sensor using a radio frequency signal; performing a second obstacle search with a night vision camera; and identifying an obstacle if a first detection of the first obstacle search corresponds to a second detection of the second obstacle search, wherein a reliability of the first detection is assessed and if the reliability is above a first threshold, the obstacle is identified independently of the second obstacle search.

A vehicular vision system includes a camera disposed at an in-cabin side of a windshield of a vehicle. Responsive to image processing at an ECU of captured frames of image data, the vehicular vision system detects an object present in a field of view of the camera. As the vehicle moves relative to the detected object, captured frames of image data are processed to determine a point of interest, present in multiple captured frames of image data, on the detected object. The vehicular vision system, via processing of captured frames of image data as the vehicle moves relative to the detected object, and based on the determined point of interest present in multiple captured frames of image data, estimates a location in three dimensional space of the determined point of interest and determines distance to the determined point of interest on the detected object.