Camera System HDR Gradation Compression for Automated Driving
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing in-vehicle camera systems face challenges in managing the increased data volume and resolution requirements due to dynamic range expansion, leading to degraded image quality and object recognition performance, particularly in medium to high luminance regions, which can affect automated driving accuracy.
Innovation Solution
A camera system that performs dynamic range expansion by alternating short and long exposure periods, followed by gradation compression using mode-specific input/output conversion parameters for viewing and sensing, optimizing image quality and recognition performance based on driving modes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If dynamic range expansion is performed by synthesizing multiple images with different exposure periods, then image quality with high dynamic range is improved, but the amount of data increases enormously
Solution Approach 1:
The patent extracts only the necessary luminance information from the synthesized HDR image by converting to a compressed luminance representation. Instead of transmitting the full high-bit-depth HDR image data, the system extracts essential luminance values and transmits them in a compressed format, reducing data amount while preserving critical visual information for sensing applications.
Solution Approach 2:
The patent changes the parameter representation of luminance data by converting from high-bit-depth luminance values to a compressed luminance scale. This parameter transformation reduces the data volume significantly while maintaining the dynamic range information needed for automated driving sensing, resolving the contradiction between image quality and data amount.
2Quantity of substance
If resolution is reduced to decrease data amount, then data transmission load is reduced, but image quality performance and object recognition performance deteriorate
Solution Approach 1:
The patent applies local quality by differentiating the treatment of different luminance regions. The compressed luminance representation preserves high precision for medium to high luminance regions where traffic lights and road signs are located, while using appropriate precision for other regions. This selective precision maintenance ensures object recognition performance is not degraded despite overall data compression.
3Illumination intensity
If more gradations are allocated to low luminance pixels to improve dark part visibility, then dark part gradation is improved, but gradation in medium to high luminance regions is reduced
Solution Approach 1:
The patent implements local quality by allocating different gradation precision to different luminance regions based on their importance for automated driving. Medium to high luminance regions containing critical objects like traffic lights and road signs receive sufficient gradation precision, while low luminance regions receive appropriate but reduced precision. This selective allocation prevents degradation of object recognition performance while managing overall data compression.
Data Source
Figure 1
Figure 2
Figure 3A~3C
AI summary
In order to provide a camera system capable of performing improved gradation compression processing in accordance with an imaging mode, the camera system includes an imaging system configured to image an object to generate an image signal, performs gradation compression processing on the image signal based on at least one of two or more input/output conversion parameters, and includes an imaging system control unit setting the input/output conversion parameter, and an image processing system that acquires an image from the imaging system in at least one of two or more imaging modes, and the imaging system control unit sets the input/output conversion parameter in a gradation compression unit according to the imaging modes in the image processing system.