Design Of Data Interface For VLSI Image Processing System Based On FPGA

In the field of image processing,the parallel capability of FPGAs is often used to achieve rapid processing of large-scale images.However,in the uploading of image data,if the peripheral interfaces on the FPGA,such as network ports and optical fibers,are used,special IP cores need to be introduced.Taking up a large amount of on-chip resources of FPGA,if using the upper computer to transmit to the FPGA through the USB interface,it is not convenient for on-site debugging and testing.Therefore,how to achieve rapid uploading of large-size images has become one of the problems to be solved in this type of project.This topic uses the U disk to support the characteristics of large-capacity storage,high-speed data transmission,proposed a method to reduce FPGA overhead using ARM + FPGA architecture to achieve rapid read U disk large-size image data.This solution greatly improves the overall portability,integration,and ease of use of the system,resulting in more stable operation.The main workflow of this interface is: First,the device mapping of FPGA under Linux is completed by ARM,and its drive and read/write interrupt control are realized.Then FPGA sends an interrupt request to ARM,ARM responds to the interrupt request,and reads from U disk.The image files to be processed are processed by slicing and transferring to the FPGA by means of asynchronous parallel technology.Finally,the FPGA-based receive buffer management system transfers the images correctly to the subsequent image processing unit,thereby realizing the cross-platform automatic transmission function of large-size images.After testing,the interface can correctly transmit the image to the image processing unit within the time required by the system.Compared with the traditional MCU+FLASH implementation scheme,the transmission speed is increased by more than twice,reaching 2 Mbyte/s.