Light Field Target Recognition Based on Depth Network Neural Algorithm

Abstract

With the rapid development of modern information technology, target recognition plays an increasingly important role in various fields of society. Through continuous improvement of the algorithm, target recognition technology has been greatly developed. As a special representation of image, light field has been favored by many scholars. In this paper, an image recognition algorithm based on light field reconstruction and depth neural network is proposed. Based on the problems of existing recognition algorithms, the reconstructed light field is applied to image recognition as a target feature library. This method can solve the problem that it is difficult to identify because of different shooting angles. Simulation results show that the algorithm can improve the accuracy and recognition efficiency.