Omnidirectional Camera Object Position Measurement Using Vehicle Motion Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle imaging systems struggle to accurately detect and track the position of objects of interest, such as trailer hitches, using omnidirectional cameras, due to non-linear effects and uncertainties in image processing, which hinders driver-assist and automated maneuvers.
Innovation Solution
An object position measurement system that employs an electronic control unit (ECU) to process omnidirectional images from a fisheye lens camera, transform them into rectilinear images, and utilize vehicle movement data to estimate the position and distance of objects by aligning optical and ground coordinate systems, reducing non-linear effects and uncertainties through calibration and bundle adjustment algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If omnidirectional camera is used to detect objects, then wide field of view is achieved, but measurement precision deteriorates due to non-linear effects
Solution Approach 1:
The system segments the object detection process into multiple discrete points of interest on the object. By tracking multiple points rather than treating the object as a single entity, the system can more accurately compensate for non-linear distortions in the omnidirectional camera field, thereby maintaining measurement precision across the wide field of view.
Solution Approach 2:
The system transforms the 2D image coordinates from the omnidirectional camera into 3D spatial coordinates by incorporating vehicle motion data and odometric information. This dimensional transformation allows the system to overcome the non-linear effects of the omnidirectional lens and achieve accurate position measurement while maintaining the wide field of view advantage.
2Device complexity
If single camera is used for object detection, then device complexity is reduced, but measurement precision deteriorates due to uncertainties
Solution Approach 1:
The system merges multiple data sources including omnidirectional camera images, vehicle motion data from controllers, and odometric information from the vehicle communication bus. By combining these diverse information sources through bundle adjustment algorithms, the system achieves high measurement precision for object position and distance while maintaining a simple single-camera configuration.
Solution Approach 2:
The system implements feedback through bundle adjustment algorithms that continuously refine object position estimates by incorporating vehicle motion data and comparing expected versus actual object positions in the camera field. This feedback mechanism compensates for uncertainties in single-camera measurements and improves overall measurement precision.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
A method and system of locating a position of an object with an omnidirectional camera affixed to a vehicle. An electronic control unit receives movement data of the vehicle from a vehicle communication bus. Simultaneously, the electronic control unit receives a plurality of omnidirectional image frames of a stationary object with the omnidirectional camera. Based on the received omnidirectional image frames, the electronic control unit identifies an object of interest in the omnidirectional image frames and tracks the object of interest in the omnidirectional image frames while the vehicle is moving. The electronic control unit determines a change in position of the object of interest in the omnidirectional image frames as the vehicle is moving and determines a distance to the object of interest based on the change of position of the vehicle and the change in position of the object of interest in the omnidirectional image frames.