Research On Application Of Wavelet Theory To Data Processing Of UAV Remote Sensing Temperature Measurement

The UAV industry developed so rapidly during recent years that it can be applied to measurement,military and even family entertainment.At present,China implements UAV remote sensing temperature measurement for the first time around Hongyanhe Nuclear Power Station,where the experiment adopts UAV carrying Infrared thermal imager to scan and measure sea-surface temperature;it is a new trial in UAV measurement field.Based on the experimental data of remote sensing temperature measurement in Hongyanhe Nuclear Power Station,this paper mainly researches the solutions to problems of image stitching and temperature fitting precision in experimental data post-processing.On the foundation of wavelet transform theory,the feasible method suitable for data processing of UAV remote sensing temperature measurement is proposed,this properly solves the problem that the precision cannot meet requirement during measurement.The main research contents of this paper include:(1)Firstly,this paper briefly introduces the experimental background of this UAV remote sensing temperature measurement,the technological difficulties encountered,domestic and overseas research status,and method as well as significance of applying wavelet transform theory.(2)Secondly,the flow of image stitching integration technology in UAV remote sensing temperature measurement is introduced;it emphatically elaborates that,when infrared wide-angle lens triggers barrel distortion in practical application,the mathematical method is used for rectifying the deviation so as to meet the demands for later image stitching and integration.(3)Thirdly,it compares and analyzes the application effect of common stitching algorithm in this UAV remote sensing temperature measurement;including Harris corner detection algorithm and SIFT feature point algorithm.Meanwhile,according to the temperature deviation issue encountered during application,the method of introducing wavelet transform theory into image processing is proposed to eliminate error.(4)Finally,aimed at the temperature error of UAV measurement caused by environment and other factors,wavelet neural network is applied to temperature fitting,which is compared with traditional BP neural network method.The results show that,wavelet neural network can guarantee to control the error of temperature fitting within 0.40 C and it is superior to BP neural network.The research on above links in UAV remote sensing temperature measurement indicates that,wavelet transform theory can be applied to the data post-processing of UAV remote sensing temperature measurement,which enables the acquired data result to be more accurate and greatly improves measurement precision degree.