GPRS-based Remote Measuring And Forecasting System For The Ice Slush Image Of The Yellow Rive

For construction warning systems of ice disaster in the upper and middle reaches of the Yellow River, the research in this paper will provide the basis for the decision project of ice-jam flood forewarning and prevention of Yellow River, in which one gets real-time ice image of river channels remotely during the Yellow River ice flood season with especial technology and obtains important ice parameters as distribution density and velocity of flow by further image data processing.Especial technological scheme was proposed for this purpose. It is like this: fisstly, on-site ice image data of river channels are sampled by CCD video camera, the wireless remote transmission of which is achieved by GPRS image data transmission technology. In the center of image information processing, special software is used to deal with and calibrate these data to get real-time distribution density of ice and the velocity of ice flow.To realize the above scheme, theoretical and experimental studies are carried on from four aspects.(1) the feasible simulation experiments are performed based on the image processing toolbox (IPT) of Matlab:under laboratory conditions, IPT is used to treat the acquired ice image, which demonstrates the feasibility of the method to get parameters as distribution density and velocity.(2) Calibrating experiments with video camera:the accuracy of camero calibration influence the final results directly. In this paper, we employ single CCD camera fixed-position mesuring system for the front ice image sampling device instead of Pan/Tilt/Zoom(PTZ) control, which will decrease the workload of the earlier camera calibrating.(3) GPRS Communication Experiment based on Huawei GTM900C: Huawei kernel comunication module GTM00C is utilized to support a lot of data transferring experiments of GSM and GPRS.(4) The optimal effectiveness experiment works for image receiving and decoding on host computer and menthod to deal with ice.Based on the theoretical research and our experments, the author and his research team commonly designed the ice image forcasting and processing system based on GPRS for the Yellow River, which has been used to experiment for sections in the first inflexion hydrological station of the Yellow River in inner Mongol for three months(2009.12-2010.3). The complete observation of the whole process of closure and open of river channels produces amount of ice image data. The analysis of engineering test and data shows that our system perserves stable work and reliable service under serious field environment. With this system, we can transmit ice image remotely and wirelessly. The software of ice image processing will accurately obtain distribution density parameter and give better result.There are aslo some problems to be resolved in the new system, such as:(1) the bottleneck of GPRS transmission rate results to a bit nonfluency of Image Streaming.(2) due to the complexity of image matching algorithm, the velocity of ice flow cannot be ontained stablely.Some advices are given to solve above problems:to quicken image data transmission rata, enlarging the coverage of 3G network is expected. The curacy of ice image mathing, which influence the parameter acquisition of velocity of ice flow, may be improved by BP neural network. A simplified example was performed in this paper, which produced better result for the velocity of ice flow.