Stereo Vision Edge Detection for Object Width Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Pre-crash sensing systems face challenges in accurately determining the width of objects, particularly vehicles and pedestrians, due to similar height distributions and unreliable length features, which hinders robust classification and collision detection.
Innovation Solution
A method that computes horizontal image derivatives to detect edges, integrates column sums for edge strength plotting, and selects the pair of edges with the maximum score to estimate object width, utilizing stereo vision and image processing techniques to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If height distribution is used for object classification, then classification can be performed using readily available sensor data, but classification accuracy is insufficient due to similar height distributions between vehicle and pedestrian classes
Solution Approach 1:
The patent transitions from using only height (vertical dimension) to incorporating width (horizontal dimension) for object classification. By computing horizontal image derivatives and detecting edge pairs, the system extracts width information that provides an additional discriminative dimension, enabling better separation between vehicle and pedestrian classes that have similar height distributions.
2Reliability
If length is used as a classification feature, then some discrimination between objects can be achieved, but reliability is poor because length is only observable for objects in side lanes at close distances
Solution Approach 1:
The patent applies local quality by computing horizontal image derivatives at each pixel location and integrating edge strengths locally before detecting dominant edges. This local processing approach enables width measurement for objects at various positions and distances in the field of view, making the feature observable and reliable across all locations rather than only in side lanes at close distances.
3Measurement precision
If edge detection algorithms are implemented to measure object width, then classification accuracy can be improved, but device complexity increases due to additional image processing requirements
Solution Approach 1:
The patent segments the image processing task into distinct stages: computing horizontal derivatives, integrating column sums to create edge strength plots, detecting dominant edges through peak detection, and evaluating edge pairs. This segmentation of the complex edge detection process into manageable components improves measurement precision while making the overall system more tractable and implementable.
Data Source
AI summary
The present invention provides a system and method for detecting edges of an object in a pre-crash sensing environment. The method includes extracting vertical edges from peaks in an array of column sums of absolute values of normalized horizontal derivates computed in the intensity image. The pixels with corresponding range image depths proximate the estimated object's closest surface contribute the column sums. The search is conducted for all likely combinations of left/right edge pairs preferably within a pre-determined horizontal range about the object's detected center. The left/right edge pairs that delimit a width within the pre-determined range of possible widths for the object are selected. Alternatively, an additional search is performed on each pair of edges to support any angular or inward slant on the edges of the object.