Image Processing Device Optical Noise Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing imaging technologies face challenges in suppressing optical noise in captured images without modifying the mechanical configuration of imaging elements or apparatuses, as stray light and other sources of noise contaminate the signal during exposure and reading processes.
Innovation Solution
An image processing device that reads and corrects image data in divided regions, allowing for the identification and subtraction of optical noise by comparing adjacent regions and using optical noise characteristic information to improve detection accuracy and precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If electric charges are stored in the electric charge accumulation unit until readout after exposure, then global shutter function is achieved, but optical noise is superimposed on the accumulated charges
Solution Approach 1:
The pixel array is divided into a first pixel array and a second pixel array. The first pixel array reads out image signals during the exposure period, while the second pixel array accumulates optical noise signals during the same period. This segmentation allows simultaneous acquisition of both image data and noise data without interfering with each other, resolving the contradiction between maintaining global shutter function and preventing optical noise contamination.
2Measurement precision
If optical noise removal signal is acquired by reading out only noise electric charges, then noise correction is enabled, but additional reading operations increase processing time
Solution Approach 1:
The patent combines the acquisition of image signals and noise signals into a unified reading process. The first pixel array reads out image signals while the second pixel array reads out noise signals simultaneously during the exposure period. By merging these operations, the system acquires both types of data in parallel without requiring separate reading operations, thus maintaining high noise correction accuracy while minimizing additional processing time.
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 approach effectively suppresses optical noise in captured images without altering the imaging element or apparatus configuration, enhancing image quality by accurately identifying and correcting noise sources.
Implementation Method 1
A CMOS image sensor generally uses a sequential reading method called a rolling shutter. An image sensor that implements an electronic global shutter by, while using the CMOS image sensor, disposing an electric charge accumulation unit adjacent to pixels and transferring electric charges to the electric charge accumulation unit at the same time from all pixels is also known.
Data Source
AI summary
An image processing device includes a image processor that reads out image data which is captured by an imaging element and transferred to a memory and on which optical noise is superimposed, as region image data for each of a plurality of divided regions of the memory, and that reads out data of a predetermined region again after reading for each region image data is finished, and an display processor that outputs corrected image data obtained by correcting captured image data for each of the plurality of regions in accordance with optical noise decided in accordance with the data read out again by the image processor, the captured image data being captured by the imaging element and stored in the memory.