The Design And Implementation Based On Sopc Video Processing System

Many different types of video signal interfaces have emerged with the increase of newvideo equipments. These differences have lead to complex designs when one wants totreat different video signals. This thesis implements a video conversion processingsystem (named VSYS) based on the SOPC technology. It can transform a variety ofanalog TV signals into multi-resolution analog VGA video and LVDS video, therebysolving the video compatibility problem. Moreover, based on the VSYS system, thisthesis also introduces a low-cost way to display overlay strokes video and how to do aneffective real-time detection of aircraft targets. My research work is summarized asfollows:First, VSYS achieves a multi-mode video formats transform. VSYS’s input are:analog composite video signal, S-video signals, the color component video signals, andLVDS signals. ADI’s ADV7401was used to sample multi-format videos, then TaiwanQuintiles CH7034B was used to change resolutions, finally after D/A conversion, theexpected standard VGA video output was obtained.This thesis then develops a low-cost stroke signal with a VGA video overlaydisplay scheme. Stroke signal is usually used as a video’s auxiliary information.Traditionally displaying video and stroke signals usually need professional’s head-updisplay equipment. While this system uses the FPGA to process the sampled strokesignal and digital video, the joint video signal can be displayed on a standard VGAmonitor.Additionally, using our simple communication command format with the AlteraFPGA and SOPC technology, the Nios II soft-core, we can perform a flexible control onthe VSYS system. The Nios II programmable configurable features make post-modifydebugging in the system easy, which can significantly speed up development. The CPUpart is shared on an FPGA chip so there is no need to add another master chip, therebyreducing the complexity and cost of hardware implementation.Using the traditional way of detection of aircraft targets in analog video, the falsealarm rate is high and difficult to meet the real-time requirements. This thesis achieves a new aircraft target detection system based on P-Viola and Jones face detection method.We modified the method by collecting a large number of different aircraft targettest images, and with the help of the OpenCV computer vision libraries we trained foraircraft target classifier. From the simulation results, it can be seen that our new aircraftdetection system is more effective compared to older methods.