Building Change Technology Research In The Verification Area Based On The UAV Image

The illegal land and building in the urban and rural areas is always a very difficult problem for government administration(hereinafter referred to as two-illegal).On the one hand,it need to timely discover the place and carry out the law in order to avoid the difficulty of tearing down after the building constructed;on the other hand,the evidence collection of illegal building and the fairness of the law carried out is an issue.Therefore,the rapid acquisition of building change has an important practical significance on the two-illegal governance.After the investigation of two-illegal construction activities,we conclude that we needs a short cycle and low cost on aquiring data.This paper uses the UAV fusion image as the data source to do a more in-depth study on the object-based building extraction and the building automatic change detection of two-stage data.And the application of this way was carried out in Miyi County,Panzhihua City,Sichuan Province.The specific study contents and main achievements are as follows:1.Analysis the characteristics of the aerial survey internal processing platform and select the INPHO software as the platform and perform data processing on the images collected by the UAV.Finally,Digital Orthophoto Map and image matching point cloud are obtained which achieves a high positional accuracy.2.For the shortcomings of the UAV image in lack of spectral information and the characteristics of the building with a certain height,the image point cloud is used to provide with height information.Based on the idea of multi-source data fusion,the normalized digital model generated by the image matching point cloud and the digital orthophoto image are bandfused and the UAV fusion image is used as the data source.3.According to the characteristics of the terrain features,the application of building object extraction technology based on SLICO segmentation and random forest classification is studied.The experimental results show that the correct rate and completeness rate of the area of Miyi County and Shuitang Village are respectively 85.1% and 85.8%;66.7%,84.4%.The method is higher than SLIC segmentation and multi-scale segmentation with the random forest classification.At the same time,the UAV fusion image with high information provided by the image matching point cloud is verified,which can improve the accuracy of building extraction.4.When the buildings changes between the two-phase data,the features will change on terrain and height.The technology of building change detection based on rule classification is proposed.This technology is used to perform the overlap segmentation on the building extraction results and describe the change classes.At last,automatic change identification is performed as unchanging building,new building,disappearing building,increasing building,and decreasing building.5.The application and the supervision mode: the ministry of housing and urban-rural development delimits the scope of supervision and determines the investigation cycle;the rapid investigation of illegal construction based on the study of this paper and the timely delivery of coordinate information;the law enforcement and the feedback of two-illegal confirmed;the area statistics of the illegal construction confirmed and the report finished;the ministry of housing and urban-rural development receives and confirms the report.Finally,the statistics results of the supervision cycle show that the supervision model based on this study is very effective.