Waterway Facilities Extraction Based On Unmanned Aerial Vehicle Images

This stage is a critical period for China to build a comprehensive transportation system.Timely and accurate information on waterway facilities is of great significance for the inland waterway information construction and the green and efficient development of inland waterway navigation.Compared with the traditional way of inspecting waterway facilities,unmanned aerial vehicle(UAV)remote sensing technology effectively solves the problems of low efficiency,high cost,and long reaction time in traditional methods.However,existing UAV images monitoring methods using manual visual interpretation cannot be employed in real-time monitoring inland waterway facilities.Therefore,the research on the automatic extraction of inland waterway facilities has important practical application value.Taking inland river barrage facilities as the research object and using the UAV images,this paper completes the relevant researches on the automatic extraction of barrage.The main contents completed are as follows:(1)The UAV remote sensing equipment is used to obtain images from the experimental area,and then the corresponding preprocessing work is performed on the images.(2)For the characteristics that UAV images have high resolution and many different objects,the relevant algorithms of image segmentation and classification are under deep researched.Through comparative analysis,the method of simple linear iterative clustering(SLIC)algorithm combined with support vector machine(SVM)is used to perform water and land segmentation,and gets a more ideal result.(3)Due to the higher brightness of the water body around the barrage,this paper designs an algorithm based on the spatial relationship features of superpixels to extract barrage,and it is implemented by Matlab.The experimental results show that the completeness and accuracy of the proposed barrage extraction algorithm can respectively reach more than 90%and 87%,which has certain validity and applicability.