新型光电混合相关器的图像识别研究

研究并设计了一种基于＂FPGA＋DSP＋ARM＂架构的新型光电混合相关器。该系统采用DSP与FPGA完成目标图像的采集、预处理以及畸变不变处理,采用光学处理模块实现联合图像的傅立叶变换,得到联合功率频谱,最后,S3C2440完成对该频谱的采集、振幅调制滤波、傅立叶逆变换以及图像识别。大量的实验表明：该光电混合相关器实现了图像识别的智能化、实时化及小型化,具有较强的实用价值。

Research on Image Recognition of a Novel Optical-Electrical Hybrid Correlator

Based on ＇FPGA ＋DSP＋ARM＇ frame, a novel optical-electrical hybrid correlator is proposed and designed. In this system, TMS320C6416 and FPGA are used to do image collection, pre-processing, and distortion-invariant processing, and optical processing module is developed to realize joint image FFT and to obtain joint power spectrum; lastly S3C2440 is adopted to complete spectrum collection, amplitude-modulated filter, IFFT, and image recognition. The large number of experiments show that this optical-electrical hybrid correlator realizes the intelligency, real time, and miniaturization of image recognition, and it is of re＇eat annlicable values.