A method of complementing a map of a scene with 3D reference points including four steps. In a first step, data is collected and recorded based on samples of at least one of an optical sensor, a GNSS, and an IMU. A second step includes initial pose generation by processing of the collected sensor data to provide a track of vehicle poses. A pose is based on a specific data set, on at least one data set re-coded before that dataset and on at least one data set recorded after that data set. A third step includes SLAM processing of the initial poses and collected optical sensor data to generate keyframes with feature points. In a fourth step 3D reference points are generated by fusion and optimization of the feature points by using future and past feature points together with a feature point at a point of processing. This second and fourth steps provides significantly better results than SLAM or VIO methods known from prior art, as the second and the fourth steps are based on recorded data. Wherein a normal SLAM or VIO algorithm only can access data of the past, in these steps, processing may also be done by looking at positions ahead, by using the recorded data.

Systems and techniques are described herein for performing optical flow estimation for one or more frames. For example, a process can include determining an optical flow prediction associated with a plurality of frames. The process can include determining a position of at least one feature associated with a first frame and determining, based on the position of the at least one feature in the first frame and the optical flow prediction, a position estimate of a search area for searching for the at least one feature in a second frame. The process can include determining, from within the search area, a position of the at least one feature in the second frame

The present disclosure relates to a computer-implemented method for object tracking, the method including the steps of defining a state-space of interest based on a class of objects subject to tracking. Further, the method includes the step of representing the state-space of interest using a FEM representation partitioning the state-space of interest in elements. Further, the method includes initiating a state-space distribution defining a probability density for different states of at least one tracked object in the state-space of interest. Moreover, the method updates the state-space distribution based on evidence, wherein the evidence being at least one of sensor data and external data of at least one tracked object in said class of objects. Furthermore, the method propagates the state-space distribution of the at least one tracked object for a time period.

A navigation system for a vehicle comprises onboard sensors including a vision sensor, and an onboard map database of terrain maps. An onboard processer, coupled to the sensors and map database, includes a position PDF filter, which performs a method comprising: receiving image data from the vision sensor corresponding to terrain images captured by the vision sensor of a given area; receiving map data from the map database corresponding to a terrain map of the area; generating a first PDF of image features in the image data; generating a second PDF of map features in the map data; generating a measurement vector PDF by a convolution of the first PDF and second PDF; estimating a position vector PDF using a non-linear filter that receives the measurement vector PDF; and generating statistics from the estimated position vector PDF that include real-time measurement errors of position and angular orientation of the vehicle.

A navigation system for a vehicle comprises onboard sensors including a vision sensor, and an onboard map database of terrain maps. An onboard processer, coupled to the sensors and map database, includes a position PDF filter, which performs a method comprising: receiving image data from the vision sensor corresponding to terrain images captured by the vision sensor of a given area; receiving map data from the map database corresponding to a terrain map of the area; generating a first PDF of image features in the image data; generating a second PDF of map features in the map data; generating a measurement vector PDF by a convolution of the first PDF and second PDF; estimating a position vector PDF using a non-linear filter that receives the measurement vector PDF; and generating statistics from the estimated position vector PDF that include real-time measurement errors of position and angular orientation of the vehicle.

A system for tracking a target object across a plurality of image frames. The system comprises a logic machine and a storage machine. The storage machine holds instructions executable by the logic machine to calculate a trajectory for the target object over one or more previous frames occurring before a target frame. Responsive to assessing no detection of the target object in the target frame, the instructions are executable to predict an estimated region for the target object based on the trajectory, predict an occlusion center based on a set of candidate occluding locations for a set of other objects within a threshold distance of the estimated region, each location of the set of candidate occluding locations overlapping with the estimated region, and automatically estimate a bounding box for the target object in the target frame based on the occlusion center.

A method can include receiving (1) images of at least one subject and (2) at least one total mass value for the at least one subject. The method can further include executing a first machine learning model to identify joints of the at least one subject. The method can further include executing a second machine learning model to determine limbs of the at least one subject based on the joints and the images. The method can further include generating three-dimensional (3D) representations of a skeleton based on the joints and the limbs. The method can further include determining a torque value for each limb, based on at least one of a mass value and a linear acceleration value, or a torque inertia and an angular acceleration value. The method can further include generating a risk assessment report based on at least one torque value being above a predetermined threshold.

The invention relates to a method and a device for fusion of measurements from various information sources (I 1, I 2, . . . , I m) in conjunction with filtering of a filter vector, wherein the information sources (I 1, I 2, . . . , I m) comprise one or more environment detection sensor(s) of an ego vehicle, wherein in each case at least one measured quantity derived from the measurements is contained in the filter vector, wherein the measurements from at least one individual information source (I 1; I 2; . . . , I m) are mapped nonlinearly to the respective measured quantity, wherein at least one of these mapping operations depends on at least one indeterminate parameter, wherein the value to be determined of the at least one indeterminate parameter is estimated from the measurements of the different information sources (I 1, I 2, . . . , I m) and wherein the filter vector is not needed for estimating the at least one indeterminate parameter.

The present disclosure provides an object tracking method, an object tracking device, an electronic device and a computer-readable storage medium, and relates to the field of computer vision technology. The object tracking method includes: detecting an object in a current image, so as to obtain first information about an object detection box, the first information being used to indicate a first position and a first size; tracking the object through a Kalman filter, so as to obtain second information about an object tracking box in the current image, the second information being used to indicate a second position and a second size; performing fault-tolerant modification on a predicted error covariance matrix in the Kalman filter, so as to obtain a modified covariance matrix; calculating a Mahalanobis distance between the object detection box and the object tracking box in the current image in accordance with the first information, the second information and the modified covariance matrix; and performing a matching operation between the object detection box and the object tracking box in the current image in accordance with the Mahalanobis distance.

The present disclosure provides an object tracking method, an object tracking device, an electronic device and a computer-readable storage medium, and relates to the field of computer vision technology. The object tracking method includes: detecting an object in a current image, so as to obtain first information about an object detection box, the first information being used to indicate a first position and a first size; tracking the object through a Kalman filter, so as to obtain second information about an object tracking box in the current image, the second information being used to indicate a second position and a second size; performing fault-tolerant modification on a predicted error covariance matrix in the Kalman filter, so as to obtain a modified covariance matrix; calculating a Mahalanobis distance between the object detection box and the object tracking box in the current image in accordance with the first information, the second information and the modified covariance matrix; and performing a matching operation between the object detection box and the object tracking box in the current image in accordance with the Mahalanobis distance.