Image Sensor HDR Capture Using Parallel Multi-Exposure Subframes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image sensing devices struggle to generate high dynamic range (HDR) images in a short period of time, especially in environments with rapidly changing illumination levels, leading to partial or complete overexposure/underexposure issues that hinder accurate object detection in self-driving cars.
Innovation Solution
The proposed image sensing device employs an image sensing array with multiple sensing units that generate pixel data at different frame rates and exposure periods. This allows for the simultaneous capture of multiple pieces of pixel data, including first and second sub-frame data, where the first sub-frame data is generated at a longer exposure period and the second sub-frame data is generated at a shorter exposure period, enabling real-time HDR image fusion during each image capturing operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Illumination intensity
If multiple images are captured continuously at different exposures to generate HDR images, then the dynamic range representation is improved, but the time required for image generation increases because it must wait for the image with the longest exposure period
Solution Approach 1:
The image sensor array is divided into multiple pixel groups, where each pixel group captures images at different exposure periods simultaneously. This segmentation allows parallel capture of multiple exposure frames without sequential waiting, resolving the contradiction between comprehensive dynamic range capture and extended generation time.
Solution Approach 2:
Multiple pixel groups perform preliminary capture actions at different exposure periods concurrently during the same time window. By preparing all necessary exposure frames in parallel rather than sequentially, the system eliminates the waiting time for the longest exposure while maintaining complete dynamic range coverage.
2Illumination intensity
If multiple independently captured photographs are merged to capture HDR images, then the dynamic range is improved, but the process fails in low-light environments because short exposure periods cannot receive sufficient signal
Solution Approach 1:
Different pixel groups are assigned different exposure periods tailored to their specific function: some pixel groups use longer exposure periods to capture sufficient signal in low-light conditions, while others use shorter exposure periods for bright region capture. This local differentiation of exposure quality ensures reliable HDR generation across varying light conditions.
Solution Approach 2:
The system varies the exposure period parameter across different pixel groups simultaneously, with some groups using extended exposure times for low-light reliability and others using standard exposure times for dynamic range coverage. This parameter diversification resolves the contradiction between dynamic range improvement and low-light reliability.
3Manufacturing precision
If the image fusion time is extended to allow complete image generation, then the quality of HDR images is improved, but the real-time processing capability deteriorates
Solution Approach 1:
The image sensor maintains continuous operation with multiple pixel groups capturing images at different exposure periods simultaneously throughout the same time window. This continuous parallel capture eliminates idle waiting time while ensuring complete image data is available for immediate fusion, maintaining both quality and real-time processing capability.
Solution Approach 2:
The system dynamically assigns different exposure periods to different pixel groups within the same operational cycle, allowing flexible adaptation to varying scene conditions while maintaining simultaneous capture. This dynamic configuration enables complete image fusion without extending total processing time, preserving real-time performance.
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
This solution enables the generation of clear image information under varying light conditions, overcoming the limitation that image fusion must be shorter than the exposure time, and allows for the rapid production of HDR images, enhancing the efficiency and reliability of image sensing in dynamic environments.
Implementation Method 1
a photodiode; a transmission circuit, coupled to the photodiode
Data Source
AI summary
The present invention relates to an image sensing device comprising: an image sensing array and an image processing circuit. The image sensing array includes sensing units, and the sensing units respectively generate multiple pieces of pixel data. The multiple pieces of pixel data are generated according to different frame rates under different exposure periods, and include a first pixel data of a first subframe and a second pixel data of a second subframe. The first pixel data is generated by exposing a first exposure period for a first frame rate, and the second pixel data is generated by exposing a second exposure period for a second frame rate. The first frame rate is less than the second frame rate. The first exposure period is greater than the second exposure period, and multiple pieces of the second pixel data are generated during one image capturing operation.