Multispectral Remote Sensing Imaging System Based On Cmos Image Sensor Research Several Key Problems

Multi-spectral imaging technology is a new generation of photoelectricity remote sensing and detection technology, which can capture and display accurate color information of surface features. Multi-spectral image contains more spectral information, and there is a consistency between image and spectral curve, so multi-spectral imaging technology can overcome the phenomenon of metamerism,and can capture the image information of target more accurately and more effectively, so in many fields,such as astronomy, military reconnaissance, environmental monitoring and so on, multi-spectral imaging technology has a very broad application prospects, and has become an subject paid attention extensively.This paper concludes the current situation of multi-spectral detection, the structure and working principle of CMOS image sensors, focus on several practical problems of multi-spectral imaging system based on CMOS image sensor, it is very important that we prove the remote sensing technology, the main research content and research achievements of this paper are summarized as follows:Firstly, research the structure, working principle, data reading way of CMOS image sensor and multi-spectral imaging theory, propose a conceptual design of multi-spectral imaging system of placing thin-film interference filter system in front of the prime lens, and devement a prototype device. The optical system includes a thin-film interference filter system in front of the prime lens, lens is connected with camera by the mechanical structure. The parameters of thin-film interference filters is reference the wave band and bandwidth of main earth observation satellites, select13reference band to customize the filters, and we can free combination according to different detection goals and requirements. I have developd the driver electronics and control unit, which could achieve the manual and automatic exposure control.Secondly, research the filtering algorithm of multi-spectral images, propose a joint evalution method of image based on SNR. First, filtering the image based on Matlab software, analyse the processing results, give the evaluation efficiency of each algorithm, then calculate the SNR of spectral image based on Visual C++, give a quantitative evaluation about spectral image, the result show that this CMOS imaging system has a better capacity of spectral imaging.Thirdly, research the influence of noise on CMOS image sensor, classify the noise according to sources, give the effective methods of noise suppression, provide a support for optimizing of imaging quality, promote the development and application of CMOS image sensor.Finally, based on the research of atmospheric radiative transfer model, build the multi-spectral imaging system model and transfer function of DN output, give a method about using MC method to simulate physical process of multi-spectral imaging, this research coule improve the detective precision and efficiency of multi-spectral, promote the development of detection technology.