Applying Convolutional Neural Network for Object Detection on FT-Matrix 7002 DSP
Qiang Zhang, Xiao Hu, Xi Tian, Siyu Lei · 2019 IEEE International Conference on Signal, Information and Data Processing (ICSIDP) · 2019
Object detection, as one of the most important components in image recognition, has always been the focus and difficulty in the field of computer vision research. It has a very important use value and is now widely used in various fields. In the military field, object detection plays an important role in unmanned combat, battlefield detection, and defensive alerting. The traditional object detection mainly adopts the method of manually extracting features, and its algorithm has strong limitations and limited space for improvement. With the rapid development of artificial intelligence theory and deep learning technology in recent years, new algorithms and ideas have been introduced into the research of object detection, which has made breakthrough progress. Based on the latest deep learning theory, this paper develops a mobile military object detection system based on convolutional neural network, and deploys and optimizes it on the high-performance, low-power FT -Matrix7002 DSP embedded platform.