Research Of Two-channel USB3.0 Picture Transmission And GPU Acceleration

With the rapid development of technology, image acquisition, application delivery technology are more and more widely used, and it’s being in the direction of high-speed, high-definition and multi-sensor. The amount of image data is getting biger and biger, the general rate of transmission doesn’t meet the matching of data size which has something to do with the high-resolution, multi-sensor. Traditional image processing method with the CPU serial doesn’t do it well in the field of processing speed, it is difficult to achieve the goal of video processing in time in the field of system with higher computational complexity.In this paper, we discuss the CMOS image transmission system based on dual-channel USB3.0, and we use the latest CUDA language to solve real-time image processing with the parallel computing GPU. Using FPGA as the driven hardwere, high-speed USB3.0 channel will collect and send data to PC in two high-resolution CMOS simultaneous transmission, and it can ensure the transmission speed and image integrity. We also discuss the real-time speed of the Cvvimage embedded MFC and OpenCV 2 self windows in this paper. Then we use GPU acceleration to complete real-time processing and display diagram, it saves much CPU space for image transmission and display CPU space.So it achieve the target of high-definition images, high-speed transmission and processing. Then we fixed the ration,at the same time we compare the effects of image processing by filtering between the differences of serial GPU technology and parallel GPU technology.At last we do the test for the system in which we analyse the transmissioning, displaying and processing speed of this system.This design centers around the principle of transmission system,image acquisition and displaying,as well as the design of software in the field of image acquisition about GPU accelaration.It reaches the target of real-time image transmission and image processing, and it laies the foundation of the application in the fields of target capturing and real-time monitoring.