Research And Implementation Of Image Preprocessing Based On FPGA

Since the beginning of the twenty-first Century,the image processing technology has entered a stage of rapid development,and many industries have begun to pay attention to the application of image processing technology.At present,the implementation scheme of the most of the image processing technology is based on universal computer or DSP,and with the increasing demand of volume,power consumption and real-time of product,the product developed in accordance with the above methods already cannot keep up with market demand.This paper firstly introduces all kinds of hardware platform of image processing technology,and analyzes their advantages and disadvantages,and generally describes the present situation and future development direction of image preprocessing technology.Then,explain the hardware structure,software development environment and design process of FPGA,and then analysis of the structure of real-time image preprocessing algorithm,in combination with the features like flexible interface of FPGA,strong parallel computing ability,strong expansibility,establishing a low-cost implementation scheme of image preprocessing that use the FPGA as the core processor.The design is mainly divided into four parts: image acquisition,image storage,image preprocessing and image display.The image acquisition function is performed by a CMOS camera module,the module can transfer the image data of analog signals into digital image date of different formats,setting the working pattern of camera through the camera module SCCB interface.Since the amount of image data obtained from the camera module is very large,the FPGA internal storage is not enough to save so much data,and need the external SDRAM storage as the buffer and storage part of the data.The initialization of the SDRAM and other necessary operations is implemented by the SDRAM controller.This design has implemented a variety of image preprocessing algorithms including format conversion,mean filtering,median filtering,edge detection and morphological image processing.Finally,convert the image data into analog signal through DAC,and then input to the VGA display to observe the effect of image processing in real time.In the process of design and development,the author adopt pipeline structure and parallel operation to realize each function,and the FPGA parallel processing ability is fully utilized to ensure this design with the ability of real-time processing.This design has two key points need to be elaborated,one is FPGA hardware interface design and the other one is FPGA implementation of image preprocessing algorithm.The high-speed communication with external SDRAM is the difficult point of hardware interface design.Only by ensuring the correctness of read and write operation of SDRAM then ensure the correctness of data.The difficulties of FPGA implementation of the algorithm is the segmentation and reorganization of the algorithm,and realize the parallel computation and pipeline structure of the image processing algorithm,and increase the speed of image processing.The hardware circuit of this design has been debugged,and the output results verified that the image preprocessing that based on FPGA is feasible.