Tail Lamp Detection Using Directional Scanning and Luminance Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing preceding vehicle detection systems face instability in detecting tail lamps, especially when integrated with light from oncoming vehicles, due to limitations in distinguishing between different light sources in gray scale images, leading to incorrect identification and unstable detection.
Innovation Solution
A system that scans pixel lines vertically and horizontally to detect tail lamps by setting a reference pixel line and determining the decrease in pixel number, calculating central points, and adjusting detection based on luminance thresholds, allowing for accurate discrimination of tail lamp areas from other light sources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If tail lamp detection is performed in gray scale images, then detection capability is maintained, but discrimination between tail lamps and oncoming vehicle headlamps deteriorates
Solution Approach 1:
The patent segments the detection process into multiple stages: first detecting bright regions above the vehicle, then filtering these regions based on luminance distribution characteristics. By dividing the detection into these segments, the system can maintain gray scale compatibility while improving discrimination through systematic analysis of luminance patterns across different regions.
Solution Approach 2:
The patent applies local quality analysis by examining luminance characteristics at different locations within the detected bright regions. Specifically, it compares luminance values between the upper region (where tail lamps appear) and lower regions, and analyzes the distribution of luminance across horizontal and vertical directions to distinguish tail lamps from headlamps based on their characteristic luminance patterns.
2Productivity
If bright regions are detected without directional scanning, then detection speed is improved, but detection stability deteriorates due to integration of multiple light sources
Solution Approach 1:
The patent segments the scanning process into horizontal and vertical directional scans, analyzing bright regions along different axes. This segmentation allows the system to detect the characteristic rectangular shape of tail lamps through systematic scanning patterns, improving detection stability by filtering out spurious bright regions that do not exhibit the expected geometric characteristics.
Solution Approach 2:
The patent introduces dimensional analysis by scanning bright regions both horizontally and vertically to determine their geometric characteristics. By examining the extent of bright regions in multiple dimensions (horizontal width, vertical height), the system can distinguish actual tail lamps from transient or atypical bright spots, thereby improving detection stability without significantly increasing processing time.
3Measurement precision
If no minimum pixel number threshold is applied, then detection sensitivity is improved, but false detection from noise increases
Solution Approach 1:
The patent applies a minimum pixel number threshold as a filtering criterion to eliminate false detections from noise. By requiring that detected bright regions contain a minimum number of pixels, the system filters out spurious small bright spots that are likely noise, while still maintaining sensitivity to actual tail lamps which naturally occupy sufficient pixel areas. This partial filtering action effectively removes harmful false detections without significantly impacting legitimate detection capability.
Data Source
AI summary
A preceding vehicle detection system includes an image capture module for capturing an image of surroundings of a subject vehicle, a preceding vehicle detection module for detecting a preceding vehicle from an image captured and a tail lamp detection module for detecting a pixel area having a luminance which is larger than or equal to a threshold value on a pixel line in the image and detecting the pixel area in each pixel line while scanning pixel lines on the image in leftward and rightward directions from the pixel line set as a reference to detect areas where tail lamps of the preceding vehicle are captured from the image.