Image processing techniques are described to receive bounding box information that describes a bounding box located around a detected object in an image, determine one or more positions of one or more reference points on the bounding box, determine, for each reference point, 3D world coordinates of a point of intersection of the reference point and the road surface, and assign the 3D world coordinates of the one or more reference points to a location of the detected object.

A volumetric data structure models a particular volume representing the particular volume at a plurality of levels of detail. A first entry in the volumetric data structure includes a first set of bits representing voxels at a first level of detail, the first level of detail includes the lowest level of detail in the volumetric data structure, values of the first set of bits indicate whether a corresponding one of the voxels is at least partially occupied by respective geometry, where the volumetric data structure further includes a number of second entries representing voxels at a second level of detail higher than the first level of detail, the voxels at the second level of detail represent subvolumes of volumes represented by voxels at the first level of detail, and the number of second entries corresponds to a number of bits in the first set of bits with values indicating that a corresponding voxel volume is occupied.

A system for navigating a host vehicle may include a at least one processing device. The at least one processing device may be programmed to receive, from an image capture device, at least one image representative of an environment of the host vehicle. The at least one processing device may also be programmed to analyze the at least one image to identify an object in the environment of the host vehicle. The at least one processing device may also be programmed to determine a location of the host vehicle. The at least one processing device may also be programmed to receive map information associated with the determined location of the host vehicle, wherein the map information includes elevation information associated with the environment of the host vehicle. The at least one processing device may also be programmed to determine a distance from the host vehicle to the object based on at least the elevation information. The at least one processing device may further be programmed to determine a navigational action for the host vehicle based on the determined distance.

A system and method are arranged to provide recommendations to a user of a vehicle. In one aspect, the vehicle navigates in an autonomous mode and the sensors provide information that is based on the location of the vehicle and output from sensors directed to the environment surrounding the vehicle. In further aspects, both current and previous sensor data is used to make the recommendations, as well as data based on the sensors of other vehicles.

A method for determining objects of interest using a vehicular vision system includes providing at least one camera and providing a control having an image processor that processes image data captured by the camera. An object present in the field of view of the camera is detected via processing at the control of a first frame of captured image data. Hypotheses filtering and hypotheses merging may be utilized when processing additional frames of captured image data to discern the detected object being an object of interest from the detected object being not an object of interest. Hypotheses filtering includes tracking the detected object over multiple additional frames of captured image data, and hypotheses merging includes comparing outputs of hypotheses filtering for the multiple additional frames of captured image data. The detected object may be determined to be an object of interest responsive to results of hypotheses merging.

A computer vision system is provided that includes an image generation device configured to generate consecutive two dimensional (2D) images of a scene, and a dense optical flow engine (DOFE) configured to determine a dense optical flow map for pairs of the consecutive 2D images, wherein, for a pair of consecutive 2D images, the DOFE is configured to perform a predictor based correspondence search for each paxel in a current image of the pair of consecutive 2D images, wherein, for an anchor pixel in each paxel, the predictor based correspondence search evaluates a plurality of predictors to select a best matching pixel in a reference image of the pair of consecutive 2D images, and determine optical flow vectors for each pixel in a paxel based on the best matching pixel selected for the anchor pixel of the paxel.

A driver assistance apparatus being configured to determine an object position sequence for each of a plurality of objects and generate an object track to approximate each respective object position sequence. The apparatus also sorts the object tracks in to at least one object group according to the value of at least one parameter of each of the object tracks. For each object group, a swarm function is generated to approximate the object position sequences of the object tracks that are members of the respective object group. A swarm lane is generated according to each the swarm function, the swarm lane portion representing a portion of a lane. A corresponding method is also provided.

In various examples, a deep neural network (DNN) is trained—using image data alone—to accurately predict distances to objects, obstacles, and/or a detected free-space boundary. The DNN may be trained with ground truth data that is generated using sensor data representative of motion of an ego-vehicle and/or sensor data from any number of depth predicting sensors—such as, without limitation, RADAR sensors, LIDAR sensors, and/or SONAR sensors. The DNN may be trained using two or more loss functions each corresponding to a particular portion of the environment that depth is predicted for, such that—in deployment—more accurate depth estimates for objects, obstacles, and/or the detected free-space boundary are computed by the DNN. In some embodiments, a sampling algorithm may be used to sample depth values corresponding to an input resolution of the DNN from a predicted depth map of the DNN at an output resolution of the DNN.

A system for processing images captured by a movable object includes one or more processors individually or collectively configured to process a first image set captured by a first imaging component to obtain texture information in response to a second image set captured by a second imaging component having a quality below a predetermined threshold, and obtain environmental information for the movable object based on the texture information. The first imaging component has a first field of view and the second imaging component has a second field of view narrower than the first field of view.

A system for navigating a host vehicle may include at least one processing device. The at least one processing device may be programmed to receive, from an image capture device, images representative of an environment of the host vehicle. The at least one processing device may also be programmed to analyze the images to identify features within the environment of the host vehicle. The at least one processing device may also be programmed to: obtain map data corresponding to the environment of the host vehicle, the map data comprising information of the features. The at least one processing device may also be programmed to match the features identified from the images with information of the features within the environment, obtained from the map data; and localize a position of the host vehicle within a roadway using the matched features and elevation information associated with the roadway.