The Implement Of Melon Image Procession Based On FPGA

Melon is a common kind fruit, the appearance detection and quality-rating are usually done by hand. Because of the operator's mental state, the operating time and other factors, the classifications by hand is inefficient.Replacing manual completion of grading with computer vision processing technology can improve the classification speed.During the preprocessing of melon image, there are large amount of low-level image data, and the process of relatively operation is very simple. Software algorithms using the traditional hard-line real-time processing to meet the requirements, which contain a large number of pipeline and parallel units FPGA platform is ideal for image preprocessing.In this paper, it study gray image of static molen and design the median filter, order filtering, mathematical morphology, convolution filtering, Gaussian filtering and Sobel operator, Prewitt operator and Laplace operator with hardware description language on FPGA platform.With improving the fast mediam filter algorithm can icrease the speed processed by mediam filter . In the design process, with optimizing the structure and improving algorithm, activly and effectivly mine algorithm parallelism and optimize the pipeline structure ,improve the processing speeds of top-level filter moduleAlgorithms of design are logic synthesised with the development environment of ISE8.2,the experiments results achieve to preprocess the melon image successfully and show that the FPGA hardware processing of image data can not only get very good treatment effect, but can achieve higher operating frequency , so image processing speed is much higher than the image processing software algorithms, it can meet the requirements of real-time image processing.