The Study Of Compressive Sensing Based Multi-spectral Imaging And Multi-focus Fusion Methods

With the development of sensor technology, the types of visual sensor are increasing all the way. Using a variety of visual sensors with complementary characteristics to work together can effectively improve the reliability and stability of the system. Therefore, the study of how to improve the single image sensor to obtain more accurate visual information and how to efficiently use the complementary information with multiple sensors is of great theoretical significance and has practical application value.(1) Taking advantage of the natural signal recovery ability of compressive sensing theory, we transfer the Bayer pattern color image restoration problem into a sparse signal recovery one by making the assumption of the image sensor color filter array which is equivalent to sensing matrix in the compressive sensing theory. Furthermore, through improving the color filter array of Bayer model, the single sensor achieves synchronization imaging of the RGB true color image and the corresponding near infrared image, effectively improving its ability of multi-spectral imaging. The simulation results validate the proposed method.(2) Since the occurrence of the compressive imaging technologies, the image fusion based on compressive imaging draws more and more attentions from scholars domestic and overseas which, as a result, various algorithms have been proposed. However, a better method to the measurement of how active of the compressive sampling data is needed so as to make greater progress in fusion methods. Firstly, given the constant property of discrete cosine transform coefficients after compressive sensing, we calculate the approximate discrete cosine transform coefficients by sparse representation. Secondly, clarity level of compressive imaging measurement is calculated by using the approximate coefficient. And then effective fusion of compressive imaging measurement is realized. The simulation results verify the effectiveness of the proposed algorithm, compared with the traditional fusion methods; the proposed fusion algorithm has obvious advantages on both subjective evaluation and objective evaluation.