Vehicle Light Spot Detection Using Multi-Exposure Image Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image processing systems for vehicles face challenges in accurately detecting the distance of vehicles at night using cameras with low dynamic range, particularly due to the large intensity differences between headlights and taillamps, and struggle to differentiate vehicle light spots from noise sources like traffic lights and vending machines, leading to reduced detection accuracy.
Innovation Solution
An image processing system that uses a single camera to analyze images with varying exposures, allowing for precise detection of vehicle light spots from near to far distances, and employs exposure control and color information to distinguish vehicle lights from noise sources, enhancing detection accuracy and reducing the influence of noise lights.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If exposure is adjusted to the intensity of far tail lamps, then far tail lamps become visible, but near headlights cause blooming
Solution Approach 1:
The patent divides the image processing into multiple exposure segments. It captures images at different exposure levels (first exposure for near objects, second exposure for far objects) and processes them separately. The image processing unit selectively combines information from these segmented exposure images to detect light spots at different distances without interference, thus resolving the blooming problem while maintaining detection accuracy for both near and far vehicles.
Solution Approach 2:
The patent changes the exposure parameter to adapt to different distance scenarios. By adjusting exposure time or gain between first and second exposure images, the system optimizes detection for near headlights in the first exposure and far tail lamps in the second exposure. This parameter change allows the same imaging device to capture both near and far light spots effectively without blooming interference.
2Measurement precision
If exposure is adjusted to the intensity of near headlights, then near headlights become visible, but far tail lamps become dim
Solution Approach 1:
The patent segments the detection task into two exposure-based image sets: first exposure images optimized for near headlights and second exposure images optimized for far tail lamps. By segmenting the detection process this way, the system can accurately detect near headlights without compromising far tail lamp visibility, as each segment is processed with appropriate exposure settings.
Solution Approach 2:
The patent applies partial exposure action by taking multiple exposure images with different exposure levels. The first exposure uses settings appropriate for near objects, while the second exposure uses settings for far objects. This partial action approach ensures that near headlights are detected with optimal exposure without overexposing and losing detail in far tail lamps.
3Measurement precision
If two imaging devices with different filters are used, then difference of intensity between headlights and tail lamps is absorbed, but difference of intensity due to distance cannot be absorbed
Solution Approach 1:
Instead of using different filters to separate headlight and tail lamp detection, the patent inverts the approach by using different exposure levels. The first exposure image captures near light spots (headlights) with shorter exposure, while the second exposure image captures far light spots (tail lamps) with longer exposure. This inverted strategy effectively addresses the distance-related intensity difference that filter-based methods cannot resolve.
4Ease of manufacture
If a single camera is used, then system cost is reduced, but detection accuracy for vehicles at different distances deteriorates
Solution Approach 1:
The patent makes a single imaging device perform multiple detection functions by capturing images at different exposure levels. The same camera unit detects both near headlights and far tail lamps by processing first and second exposure images differently. This universal approach eliminates the need for multiple specialized imaging devices while maintaining detection accuracy across different distances through intelligent image processing.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The system achieves precise vehicle detection and light distribution control at night, improving the accuracy of vehicle distance calculation and reducing errors caused by noise lights, resulting in enhanced vehicle detection performance and safer driving conditions.
Implementation Method 1
a camera (imaging means) mounted on a vehicle and an image analysis means that analyzes plural images photographed by the camera
Data Source
AI summary
The invention provides an image processing system capable of precisely locating the positions of the light spots covering near headlights through far tail lamps by using one camera. The invention, discriminating the headlights and tail lamps from noise lights such as a traffic light, streetlight, and vending machine, enhances the vehicle detection performance at night. The system includes an image input means that inputs images in front of a vehicle, an image analysis means that analyzes the images inputted by the image input means. The image input means has a means of photographing more than two images with different exposures. The image analysis means analyzes the images photographed by the image input means to transform them into the position information of the vehicles traveling in front.