Camera Gamma Curve Adaptation for Motor Vehicle Obstacle Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Camera systems for motor vehicles face issues with image quality due to wide opening angles, where vehicle elements like the number plate influence exposure and white balance, leading to incorrect brightness assumptions and contrast issues in HDR images, particularly affecting important image regions like obstacle recognition.
Innovation Solution
A method to dynamically adapt the gamma curve by determining a region of interest and generating a new gamma curve with a steeper output range within the preferred input range, prioritizing contrast improvement for critical image regions while maintaining stability and avoiding oversteepening.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If a wide opening angle lens (fish-eye lens) is used to capture the entire environment, then the coverage area is improved, but vehicle elements (body, number plate) are also captured which interfere with exposure and white balance accuracy
Solution Approach 1:
The image is divided into multiple regions with different gamma correction curves applied to each. The region of interest (excluding vehicle elements) receives a first gamma curve optimized for accurate obstacle detection, while regions containing vehicle elements receive a second gamma curve that compensates for their interfering effect on exposure and white balance measurements.
Solution Approach 2:
Different gamma correction characteristics are applied to different spatial regions of the image. The region of interest has enhanced contrast and dynamics optimized for obstacle recognition, while vehicle element regions have adjusted gamma characteristics that prevent them from skewing overall exposure and white balance calculations.
2Adaptability or versatility
If automatic exposure adjustment is performed based on captured images, then the exposure adapts to scene brightness, but vehicle elements skew the brightness measurement leading to incorrect exposure values
Solution Approach 1:
The image processing separates the region of interest from vehicle element regions. Exposure and white balance are calculated exclusively from the region of interest using the first gamma curve, eliminating the skewing effect of vehicle elements while maintaining automatic adaptation to actual scene brightness.
3Illumination intensity
If a steep gamma curve is applied to dark image regions to improve contrast, then dark structures become more visible, but contrast deteriorates in other image regions including the region of interest
Solution Approach 1:
Different gamma curves are applied to different regions: a first gamma curve with appropriate steepness is applied to the region of interest to maintain contrast for obstacle detection, while a second gamma curve with enhanced steepness is applied to vehicle element regions to improve dark structure visibility without compromising the region of interest.
Solution Approach 2:
The gamma correction is localized to specific regions with different characteristics. The region of interest receives gamma correction optimized for its specific requirements (obstacle detection), while vehicle element regions receive customized gamma correction that addresses their specific needs (improving dark structure visibility) without affecting other regions.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method for adapting a gamma curve (G2) of a camera system by providing an input image of an environmental region of a motor vehicle and correcting the input image to an output image based on a current gamma curve (G2), by determining at least one region of interest in the output image by an image processing device, by determining a minimum and a maximum pixel value within the at least one region of interest of the output image, wherein the minimum and maximum pixel values define a current output range of values, by calculating associated minimum and maximum input pixel values (Amin, Amax) of the input image from the minimum and maximum pixel values using the current gamma curve (G2), wherein the minimum and maximum input pixel values (Amin, Amax) define a preferred input range of values (36), by generating a new gamma curve (G2') by defining a new output range of values (35') to the preferred input range of values (36), such that the new output range of values (35') is greater than the preferred input range of values (36), and applying the new gamma curve (G2') as current gamma curve (G2) for a subsequent input image.