Vehicle Camera ROI Illuminance Detection for Low-Light Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous driving systems face limitations in image recognition due to low illuminance conditions, such as insufficient lighting at night or in poorly lit indoor/underground parking areas, which can lead to unreliable performance and safety concerns.
Innovation Solution
A vehicle system that includes a camera, preprocessor, image processor, and determination unit to assess illuminance levels by generating a cumulative distribution function for illuminance values and probability values, setting a region of interest, and considering emergency light blinking periods to determine low illuminance conditions, thereby adjusting image recognition reliability accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If image recognition is used in low illuminance conditions, then autonomous driving system can operate at night or in indoor parking, but image recognition performance becomes unreliable
Solution Approach 1:
The system performs preliminary assessment of illuminance conditions before executing image recognition. By evaluating illuminance values and generating cumulative distribution functions in advance, the system determines whether image recognition results are reliable, preventing unreliable recognition in low illuminance scenarios
Solution Approach 2:
The patent introduces an intermediary assessment mechanism that mediates between the image recognition system and the autonomous driving decision-making. This intermediary evaluates illuminance conditions and provides reliability information, allowing the system to appropriately weight or disregard image recognition results based on environmental conditions
2Measurement precision
If multiple multi-cameras are mounted for autonomous driving, then object recognition capability is improved, but system complexity and cost increase
Solution Approach 1:
Instead of uniformly processing images from all cameras with equal complexity, the system applies local quality assessment by evaluating illuminance conditions specifically for regions of interest. The cumulative distribution function is generated for specific pixel regions, allowing selective processing based on local lighting conditions rather than global uniform processing
Solution Approach 2:
The system performs partial action by selectively assessing illuminance for relevant regions rather than processing all image data from multiple cameras with full complexity. By focusing computational resources on evaluating illuminance in critical areas and using cumulative distribution functions only where necessary, the system achieves reliable assessment without processing every pixel at maximum complexity
Data Source
AI summary
An embodiment vehicle includes a camera, a preprocessor configured to set a region of interest (ROI) in an image of an area outside the vehicle obtained by the camera, an image processor configured to obtain an illuminance value of a pixel belonging to the ROI in each frame of the image, and a determination unit configured to determine whether or not each frame of the image has low illuminance based on the illuminance value.