Vehicle Position Calculation Using 3D Wheel and Bumper Keypoints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle position calculation methods face challenges in accurately determining the distance and heading angle between vehicles due to difficulties in obtaining three-dimensional world coordinate values of key points, particularly when the bumper is displaced from the ground, making it difficult to calculate the distance between vehicles using two-dimensional image pixel values.
Innovation Solution
A method and system that projects three-dimensional coordinate values of front and rear wheels and bumpers of a target vehicle onto a two-dimensional plane by converting them into two-dimensional coordinate values, using conditions where bumper Z values are common and the bumper-line is perpendicular to the wheel-line, and corrects positions based on movement in consecutive images to calculate distance and heading angle.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If two-dimensional image pixel values are used to calculate vehicle position, then the calculation process is simple, but the position accuracy deteriorates because the bumper is displaced from the ground (z>0)
Solution Approach 1:
The patent transitions from two-dimensional image pixel coordinates to three-dimensional world coordinates by introducing the Z-axis (height dimension). It establishes a coordinate transformation relationship that maps 2D image points to 3D space points, allowing accurate representation of the bumper's position above the ground. This dimensional expansion resolves the fundamental limitation of 2D imaging when dealing with elevated objects.
Solution Approach 2:
The patent introduces additional parameters (Z-coordinate, height information) to the calculation model. By changing from a 2D parameter system (x, y) to a 3D parameter system (x, y, z), the model can accurately represent the bumper's spatial position. The transformation involves multiple parameters including camera intrinsic parameters, extrinsic parameters, and the height offset of the bumper from the ground.
2Measurement precision
If three-dimensional coordinate values are calculated using camera parameters, then position accuracy improves, but the calculation complexity increases due to coordinate transformation requirements
Solution Approach 1:
The patent pre-calculates and stores the transformation matrix between the camera coordinate system and the world coordinate system. By performing the coordinate transformation preparation in advance (pre-computing rotation and translation matrices), the actual real-time calculation is simplified. The transformation relationship is established beforehand, reducing the computational burden during vehicle position calculation.
Solution Approach 2:
The patent introduces an intermediate coordinate system (camera coordinate system) as a mediator between the image pixel coordinates and the world coordinates. The transformation is performed in two steps: first from image coordinates to camera coordinates, then from camera coordinates to world coordinates. This intermediate step simplifies the overall transformation process and makes the calculation more manageable.
3Loss of information
If key points of bumpers are used for position calculation, then the distance calculation becomes possible, but the calculation accuracy deteriorates because bumpers are at different heights from the ground
Solution Approach 1:
The patent segments the vehicle into different key point components (bumper, wheels, body) and processes each segment with appropriate transformation rules. The bumper key points are specifically handled by applying the height offset transformation, while other points may use different reference levels. This segmentation allows each part to be transformed according to its specific spatial characteristics.
Solution Approach 2:
The patent applies different transformation parameters to different parts of the vehicle based on their local characteristics. The bumper, being at a different height from the ground compared to wheels, receives a specific height offset adjustment in the transformation process. This local quality approach ensures that each key point is transformed according to its actual spatial position, maintaining accuracy throughout the vehicle structure.
Data Source
AI summary
Disclosed are a vehicle and a vehicle position calculation method. The vehicle includes an image sensor. The image may acquire, using the image sensor, a video image of a target vehicle, determine, based on the video image, a set of location points of the target vehicle, project, based on the determined set of location points, first three-dimensional coordinate values of location points corresponding to front and rear wheels of the target vehicle onto a two-dimensional plane by converting the first three-dimensional coordinate values into first two-dimensional coordinate values in the video image, project, based on the determined set of location key points, second three-dimensional coordinate values of location points corresponding to bumpers of the target vehicle onto the two-dimensional plane by converting the second three-dimensional coordinate values into second two-dimensional coordinate values in the video image, and control, based on the three-dimensional coordinate values and the second three-dimensional coordinate values, the host vehicle.