Hydrological Video Monitoring System Research And Realization Based On FPGA

My country has a vast territory and countless rivers and streams.Correspondingly,the distribution of rainfall is not well-distributed.Floods occur frequently during the summer flood season,which not only causes serious economic losses to the country but also makes it difficult for people to live in the vicinity of the river.In order to minimize the economic losses caused by natural disasters,timely warning and water dispatch are reliable methods,all of which are inseparable from hydrological monitoring information.Hydrological monitoring is a comprehensive subject that uses scientific methods to monitor,measure,analyze and forecast the temporal and spatial distribution and variation of water in nature.Measuring water level and water velocity is its main goal.With the rapid development of artificial intelligence in recent years,machine vision as a branch has also made rapid changes.Machine vision systems are also increasingly used in production and daily life.Utilizing the high-speed processing and good portability of field programmable gate arrays(FPGA),a convenient,accurate and stable remote hydrological monitoring system is designed and the image processing technology is used to measure the water level in order to give full play to the advantages of machine vision.In the system design,the camera adopts Omni Vision's OV5640 chip,the core processor uses ALTERA's Cyclone IV series and also includes the D/A converter ADV7123 chip and the SDRAM chip of HY57V2562 GTR.The system work process is as follows: using the camera to carry on the video image capture to the hydrological ruler,the data are put into SDRAM and dealt with by algorithm modules in FPGA,then they are sent to the video D/A conversion unit,the video monitoring picture will be showed through VGA interface.The realization of image processing algorithms and character recognition is the core part of the whole thesis,including the grayscale,edge detection,image enhancement and character recognition.In the part of image enhancement,the multi-scale Retinex algorithm is improved,scale factor ? is corrected by nonlinear "S" function,then the brightness of the image is distributed more evenly.The sub-modules of system and some algorithms are written by using Verilog HDL,programming and simulation are carried out in the Quartus ? and Modelsim.Finally,the experimental platform is set up to verify and analyse the results.The results show that the hydrological monitoring system can carry out real-time monitoring and edge detection of hydrological feet to meet the hydrological and water level monitoring needs of small and medium-sized rivers in China.It is particularly suitable for water environment which are remote and complicate,and provides powerful guarantee to monitor and utilize water resources.