UAV Image Acquisition And Its Object-Oriented Land Use Classification Application

At present,high-resolution UAV remote sensing images have been applied in many fields,and they are also widely used in the rapid dynamic monitoring of mining areas.Utilizing the low-altitude photogrammetry technology of the drone to monitor the terrain of the mining area,the required working time is short and the precision is high[1].In this paper,UAV low-altitude photogrammetry is used to collect the original image,and the original image is pre-processed.The generated UAV orthophoto image is extracted,and the area of each area in 2017 is counted.Landsat8 multi-spectrum is also used.Remote sensing data using the normalized water body index(NDWI)to extract the waters in the subsidence area,Landsat8 multi-spectral remote sensing data is used to extract the water area of the subsidence area in the past five years.According to the analysis of statistical data,the water area of the subsidence area has increased from 2013,with an annual growth rate of 1.07 from 2013 to 2014;an annual growth rate of 0.06 from 2014 to 2015;and an annual growth rate of 1.13 from 2015 to 2016;The annual growth rate from 2016 to 2017 is 0.15;the annual growth rate from 2015 to 2016 is the largest,and the water area in the subsidence area has increased by 1.9521 km2.Thereby,the dynamic monitoring of the area change of the subsided water area in the mining area is realized.UAV images have more abundant feature information,such as color,texture,geometry,etc.,but the number of UAV image bands is small.Therefore,this paper applies object-oriented classification[2],and uses human-machine interactive interpretation for the orthophoto generated by UAV preprocessing[3].Through the effective use of the classification method,the orthophotos of the UAV are quantitatively classified and obtain high interpretation accuracy[4].In the experiment,a mining area in Huainan City was selected as the research area.The SKYLAND-DF150 fixed-wing UAV was used to obtain single image data,and the obtained single image data was preprocessed by Wuhan Aerospace Vision Software to generate the orthophoto image DOM of the mining area.And digital elevation models.Continue to use Ecognition,ArcGIS and other software to classify images.And use MATLAB software to analyze and sort the image data.The data obtained by the analysis is used as the basis for image segmentation and classification.The main research results are as follows:1)Apply SKYLAND-DF150 fixed-wing UAV to obtain single image data of the mining area,use EPT and Agisoft Photoscan software to preprocess the original single image data,and the aerial triangulation encryption control point works in HAT software[5],And correcting the image of the processed center projection to an orthographic projection image.2)For the characteristics of object homogeneity and object-to-object difference after remote sensing image segmentation,use eCognition software to segment the orthophotos processed above,and import the segmented vector data into ArcGIS,combined with segmentation quality evaluation formula,and use MATLAB data.The processing software analyzes the data and seeks appropriate segmentation scales for different features.3)Use eCognition software to select representative samples of certain types of objects for the segmented objects,import the selected sample vector data into ArcGIS,generate the open file form of MATLAB software,and use the feature library optimization formula to calculate the suitable Characteristics and thresholds.4)Using the above-mentioned preferred features,the first stage of the UAV image is classified,and the second stage of supervised classification is adopted for the object categories that are prone to confusion and difficult to separate.5)Using the normalized water body index method to extract the water area from the multi-spectral image data,The spectral remote sensing data extracted the water area of the subsidence area in the past five years,and calculated the annual growth rate to realize the dynamic monitoring of the change of the water area in the mining area.Figure[36]table[22]reference[62]...