Based on the principle and design of a novel CMOS image sensor 3

2. Preprocessing process and methods for images

In order to obtain good image quality, it is necessary to process the collected raw image data. Generally, image preprocessing is done in a coprocessor. Recently, with the development of SoC technology, image preprocessing functions can be integrated into CMOS sensors, which demonstrates the advantages of CMOS image sensors.diode

The preprocessing of images mainly includes a series of processes such as defect correction, removal of FPN noise, color difference, image sharpening difference, aperture correction, gamma correction, etc.diode

The above image preprocessing process can be achieved through digital image processing algorithms, and the hardware platform can be an image processing circuit integrated in SoC, ASIC image processing chip, or a general-purpose DSP chip. The first step is to eliminate defects in the image. If there is a defect in a pixel that causes its output level to be clamped at a high level (black dot) or a low level (white dot), it needs to be compensated for through image processing. Usually, the average value of pixels of the same color around it is used to replace the output value of that pixel.diode

Usually, there are differences in column analog-to-digital converters for different columns, which leads to the generation of fixed pattern noise (FPN). The data in Black Lines in Figure 4 is used to eliminate FPN. The coprocessor will utilize this portion of data to achieve the goal of eliminating FPN.diode