A method of adjusting the drag on a trailing air vehicle (3) flying behind a leading air vehicle (1), the method comprising the steps of: (i) detecting a wingtip vortex (5) shed from the leading air vehicle (1), for example using background oriented schlieren; (ii) determining the position of the wingtip vortex (5) for example using photogrammetry; and (iii) modifying the flight path of the trailing air vehicle (3) in dependence on the determined position. This may enable the trailing air vehicle (3) to efficiently interact with the wingtip vortex (5) and reduce drag.

The present disclosure is directed to automatically generating a 360 Virtual Photographic Representation (“spin”) of an object using multiple images of the object. The system uses machine learning to automatically differentiate between images of the object taken from different angles. A user supplies multiple images and/or videos of an object and the system automatically analyzes and classifies the images into the proper order before incorporating the images into an interactive spin. The system automatically classifies the images using features identified in the images. The classifications are based on predetermined classifications associated with the object to facilitate proper ordering of the images in the resulting spin.

Vehicle repair workflow automation with natural language processing is disclosed. One computer-implemented method comprises: providing images of a damaged vehicle as first input to a computer vision machine learning model, wherein the computer vision machine learning model has been trained with images of other damaged vehicles and corresponding vehicle repair operations; receiving first output of the computer vision machine learning model responsive to the first input, wherein the first output represents a plurality of the vehicle repair operations; providing the first output of the computer vision machine learning model to a natural language processing (NLP) machine learning model, wherein the NLP machine learning model has been trained with vehicle repair content comprising a plurality of vehicle repair procedures; and receiving second output of the NLP machine learning model responsive to the second input, wherein the second output comprises a recommended one of the plurality of the vehicle repair procedures.

Disclosed is a vehicle protection fence repair plating system and method using artificial intelligence. The system includes a data management module that collects video data about a vehicle protection fence and pre-processes images per frame, a data prediction module that receives the data of the pre-processed image and performs machine learning for a corrosion level of the vehicle protection fence according to a preset labeling standard to detect a work area, and a process management module that standardizes customized work instructions according to a determination result of an image state of the vehicle protection fence, which has been machine-learned, wherein the data prediction module specifies a repair range of the vehicle protection fence and a work method for each repair range according to the labeling standard.

This description provides an autonomous or semi-autonomous excavation vehicle that is capable of navigating through a dig site and carrying out an excavation routine using a system of sensors physically mounted to the excavation vehicle. The sensors collect any one or more of spatial, imaging, measurement, and location data representing the status of the excavation vehicle and its surrounding environment. Based on the collected data, the excavation vehicle executes instructions to carry out an excavation routine. The excavation vehicle is also able to carry out numerous other tasks, such as checking the volume of excavated earth in an excavation tool and helping prepare a digital terrain model of the site as part of a process for creating the excavation routine.

The disclosed video processing device contains: a video acquisition unit that acquires surroundings information including video taken of the surroundings of a vehicle; a line-of-sight acquisition unit that acquires the origin and direction of the line of sight of the driver of the aforementioned vehicle; a line-of-sight video generation unit which generates, from the surroundings information, line-of-sight video corresponding to the origin of the line of sight; a blocking-information computation unit that computes, on the basis of the origin of the line of sight, blocking information including video or a region of the body of the aforementioned vehicle that blocks the driver's line of sight; and a display-video generation unit that generates display video on the basis of the line-of-sight video and the blocking information.

Systems and methods provide for an automated system for generating one or more three dimensional (3D) images of a vehicle and/or a baseline image for that vehicle. The system may receive 3D images of a plurality of vehicles of a same type (e.g., same make, model, year, etc.) and generate a 3D image of a baseline vehicle for vehicles of that same type based on 3D images of the plurality of vehicles of the particular type. The system may use a 3D image of the baseline vehicle to determine a characteristic of another vehicle, such as a modification made to the vehicle, damage to the vehicle, cost to repair the vehicle or replace parts of the vehicle, a value of the vehicle, an insurance quote for the vehicle, etc. In some aspects, the 3D images may optionally comprise 3D point clouds, and 3D laser scanners may be used to capture 3D images of vehicles.

An in-vehicle camera is provided which includes an imager equipped with a lens, a control board on which a circuit is mounted to control an operation of the imager, a flexible cable connecting the imager and the control board to be communicable therebetween, and an electrically conductive housing in which the imager, the flexible cable, and the control board are arranged. A spacer which is designed to have neither electromagnetic wave absorption nor electromagnetic wave reflection properties is arranged away from the control board in a contactless manner and located between the flexible cable and an inner surface of the housing to form a gap therebetween. This minimizes generation of noise transmitted from the housing to inside the vehicle without sacrificing the quality of electrical signals transmitted through a circuit in the flexible cable.

A method for monocular depth/pose estimation in a camera agnostic network is described. The method includes training a monocular depth model and a monocular pose model to learn monocular depth estimation and monocular pose estimation based on a target image and context images from monocular video captured by the camera agnostic network. The method also includes lifting 3D points from image pixels of the target image according to the context images. The method further includes projecting the lifted 3D points onto an image plane according to a predicted ray vector based on the monocular depth model, the monocular pose model, and a camera center of the camera agnostic network. The method also includes predicting a warped target image from a predicted depth map of the monocular depth model, a ray surface of the predicted ray vector, and a projection of the lifted 3D points according to the camera agnostic network.

A method for estimating a plurality of camera, comprising using at least one processor executing a code for: extracting a plurality of image features of a plurality of landmarks from a plurality of images captured by at least one camera from at least one pose, the plurality of landmarks calibrated with respect to a certain coordinate system; identifying among the plurality of image features at least one image feature documented in at least some of the images; producing scale values of at least one common image feature by analyzing the at least some of the images; determining a plurality of estimated poses of the at least one camera with respect to the certain coordinate system by using the scale values in calculating a minimal re-projection error between the plurality of image features and a plurality of predicted image features; and outputting the plurality of estimated poses.