2D Template Object Detection for Vehicle Front
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for automatically detecting objects in front of a motor vehicle are inefficient due to high memory requirements and processing time, especially on hardware with limited capabilities.
Innovation Solution
The use of flat two-dimensional template objects with reduced memory needs, arranged orthogonally to a sensing plane and standing on a ground plane, allows for faster processing and reduced memory usage, employing disparity or depth maps to simplify image processing calculations by calculating average values for regions-of-interest and integrating them for efficient matching.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If three-dimensional template objects are used for object detection, then detection accuracy is improved, but memory requirements and processing time increase significantly
Solution Approach 1:
The patent extracts only the essential geometric information from three-dimensional objects and represents them using flat two-dimensional templates. This extraction approach retains the core identification capability while eliminating the memory-intensive three-dimensional data storage, directly resolving the contradiction between detection accuracy and memory requirements.
Solution Approach 2:
The patent transforms the representation of objects from three-dimensional to two-dimensional. By changing the dimensional representation, the system maintains object identification capability through projected silhouettes while dramatically reducing memory footprint and processing complexity, thus resolving the contradiction between accuracy and resource consumption.
2Measurement precision
If three-dimensional template objects are used for object detection, then detection accuracy is improved, but processing time increases
Solution Approach 1:
The patent extracts only the essential geometric information from three-dimensional objects and represents them using flat two-dimensional templates. This extraction approach retains the core identification capability while eliminating the memory-intensive three-dimensional data storage, directly resolving the contradiction between detection accuracy and memory requirements.
Solution Approach 2:
The patent transforms the representation of objects from three-dimensional to two-dimensional. By changing the dimensional representation, the system maintains object identification capability through projected silhouettes while dramatically reducing memory footprint and processing complexity, thus resolving the contradiction between accuracy and resource consumption.
3Measurement precision
If complex image processing calculations are performed for every pixel, then detection precision is improved, but processing speed decreases
Solution Approach 1:
The patent segments the image processing task by first calculating disparity or depth values only for specific regions of interest where objects are likely to be present, rather than processing every pixel uniformly. This segmentation strategy maintains detection precision for object regions while significantly improving overall processing speed by reducing unnecessary calculations in empty areas.
Solution Approach 2:
The patent applies different processing qualities to different regions of the image. By identifying regions of interest and applying detailed analysis only to those areas, the system maintains high detection precision where needed while using simpler, faster processing methods in other areas, thus resolving the contradiction between precision and speed.
Data Source
AI summary
A method of automatically detecting objects in front of a motor vehicle comprises the steps of pre-storing template objects representing possible objects in front of the motor vehicle, detecting images from a region in front of the vehicle by a vehicle mounted imaging means, generating a processed image containing disparity or vehicle-to-scene distance information from the detected images, comparing the pre-stored template objects with corresponding regions-of-interest of the processed image, and generating a match result relating to the match between the processed image and the template objects. Each of the pre-stored template objects is a flat two-dimensional multi-pixel area of predetermined shape.