Sea Surface Wind Field Inversion Algorithm From Marine Radar Image

Sea surface wind vector information is an important parameter of the sea environmentand it is an important driving force factor of the ocean movement, so the remote sensing ofwind vector measurement has been receiving increasing attention of scholars from variouscountries. Marine radar is used for modern ships sailing and provides the reference parameters,and this can makes the route planning, weather measurements more convenience and better.Using marine radar image for the wind vector inversion is a major content of the researchabroad and domestic in recent years, it has become possible to measure the wind farms on thebasis of the results of inversion of waves, currents and other information.Early in the last century, inversion algorithm of the wind vector information by usingradar image has a primary development, but using gradient algorithm and neural networkmodel to study the wind direction and speed in respective of the wind vector is a new methodin domestic. Therefore, based on the analysis of the results in domestic and abroad about theocean telemetry of marine radar images, it is designed in theory using gradient and neuralnetwork and tested in experimental in order to prove the feasibility of the proposed algorithm.The main research work are as follows:1. Briefly narrate the important concepts of ocean remote sensing measurement systemand analyse the basic working principle of the system, and on this basis make a briefintroduction about classic wind vector inversion algorithm of ocean remote sensing systemthat includes the mature development of the optical flow algorithm and the inversionalgorithm based on the gradient and BP neural network which this article mentioned.2. Summarize the the subject technical means for the research direction of the subject.Image processing technology and the gradient algorithm is used for wind direction inversionincluding image filtering, smoothing, sharpening and other operations; BP neural networkmodel is used for wind speed inversion whose main content is training and testing.3. Introduce the important characteristics of wind vector and the basic characteristics formarine radar images. Illustrates the advantages of the gradient and BP neural networkalgorithm compared with the previous algorithm in theory, and study the algorithm wholedesign options. Make an overview of the overall process of wind gradient algorithm inversionand detailed calculation description of the gradient algorithm; Establish and design BP neuralnetwork model based on actual collected measurement data, and allocate measured data fortraining and testing. 4. Outline the experimental system including the location and the measured compositionas well as all the major parameters of the devices in the system and other information. Processthe data provided by the system for a certain basic degree. Using the measured data tosimulate the wind field based on gradient and BP neural network inversion algorithm in orderto prove the feasibility of the algorithm.5. Summary of the simulation results of the algorithm and formulate the future directionof further study.