Neural network-based multisensor image fusion algorithm

Abstract

Aiming at the problems of high mean square error and low fusion efficiency of existing fusion algorithms, a neural network-based multi-sensor image fusion algorithm is proposed. The fusion algorithm based on depth-separable convolution neural network (CNN) is determined by analyzing the quality evaluation and fusion methods of multi-sensor images, and summarizing the fusion rules. It is found that the integrity of image information acquisition is 97%, the mean square error is 4, and the fusion time is 2 s. Therefore, the algorithm has a good image fusion effect.