Research On Image Pre-Processing Algorithms And Handware Design

Image sonar is widely utilized in marine exploration. Due to the complicated environment of ocean and the characteristic of acoustical image, the quality of acoustical image is lower than optical image. So image processing technique is important for the application of image sonar. The cubage of sonar system and the speed of image processing are concerned by the designer. Sonar system is considered as real-time system. Because image processing has high inherent parallelism, FPGA has very high performance in spite of working in lower operational frequency and consuming less hardware.This thesis built an architecture that consisted of receiving and storing control module, buffer cell, processing image data module based on FPGA. SDRAM controller and CYC9449 controller are included in this architecture. This thesis analyzes several image preprocessing algorithm which improve the quality of the acoustical image , including pseudo-color processing , histogram equalization, gray-scale morphology, gradient and laplacian edge detection, and basic filtering in frequency domain. The algorithms analyzed above are implemented base on FPGA, and the optimization design for hardware is discussed. Meanwhile the algorithms are implemented base on DSP6416 DSK. The advantages of FPGA are illustrated through comparing the application of DSP and FPGA.The algorithms are implemented and simulated on ISE 9.1 and ModelSim, because the system use Virtex4-XC4VLX15 which produced by Xilinx as image processor.