Fast Target Recognition Method Based on Multi-Scale Fusion and Deep Learning

Under the natural state of unrestricted conditions, accuracy effective recognition image target information captured by ordinary cameras is significantly reduced. At present, mainstream research methods for focus on processing images based algorithms with strong ability to describe features so as improve performance in cases complex noise interference. However, most cannot adapt various changes background when environment changes. Therefore, this article conducts studies fast method multi-scale fusion and deep learning. The optimized local binary pattern algorithm HOG are used extract features, dimension reduction extracted carried out generalized discriminant analysis algorithm, targets accomplished correlation analysis. experimental results verify effectiveness proposed algorithm.