| China has a long history of civilization,from Yuhang Liangzhu civilization of Zhejiang to the generation of a new era of socialist characteristics under the leadership of the Communist Party of China now,in China,this vast and vast land space-time changes undergone five thousand years,saved up countless Cultural treasures China has a long history of civilization,from Zhejiang Yuhang Liangzhu civilization to the generation of a new era of socialist characteristics under the leadership of the Communist Party of China now,This vast and vast land space-time changes undergone five thousand years,saved up countless Cultural treasures in China.Nowadays,with the rapid development of science and technology,the continuation and protection of the Chinese national context has increasingly entered the life and vision of the general public.With the attention of people to the cultural heritage,cultural relics The higher the level,the more emphasis on cultural relics protection,cultural relics research,cultural relics exhibitions,cultural relics related social education,cultural relics digitalization,cultural communication,and cultural and creative products development.The research on the extraction of cultural relics is the basis of cultural relics protection and cultural relics research.The digitization of Chinese cultural relics in cultural relics protection projects is the prerequisite for the extraction of cultural relics.At present,for the extraction of cultural relics,the manual extraction method based on digital information of cultural relics is still generally used.The problems are as follows:(1)If there is no on-site lighting environment adjustment parameters and no color card for subsequent color correction when acquiring image data,the disease cannot be ensured.The correctness of the texture information;(2)The disease image data obtained in many cases during the disease investigation process is not an orthophoto.The application of such images for disease investigation,the reliability of the investigation results is not guaranteed;(3)the disease extraction is mostly The application software manually extracts the diseased area,and the workload is heavy and inefficient.Therefore,the automatic extraction of cultural relics from undistorted images is the focus of scholars at home and abroad.At present,the research on automatic extraction of regions mainly includes methods such as edge extraction and image segmentation.Among them,the edge extraction algorithm mainly includes differential edge detection,Reborts operator,Sobel operator,Prewitt operator,Kirsch operator,Canny operator,etc.The main problems are:(1)extracted edge data caused by the complexity of cultural relics It is difficult to judge whether it is a disease edge;(2)it is difficult to count the size of the disease itself and the length of the disease boundary.Traditional image segmentation methods such as multi-threshold segmentation,cluster-based segmentation,region-based segmentation,and the like.The main problems are:(1)multi-threshold segmentation method may not be able to obtain thresholds on images with flat histograms or peaks;(2)cluster-based segmentation method needs to first determine the number of clusters,the result of clustering The image of the initial clustering center is larger;(3)the region-based segmentation method is easy to produce over-segmented combined region merging can get better results,generally this method is applied in the field of medical image analysis.Based on the above problems,this paper proposes an automatic extraction method based on orthophotos for superpixel segmentation diseases.The main research contents include:(1)Orthographic image generation method,mainly based on point cloud orthophoto generation method and all-digital orthophoto generation method;(2)Research on super pixel segmentation algorithm,mainly including several super Comparison of pixel segmentation algorithms and optimization for simple linear iterative superpixel segmentation method(SLIC);(3)information transfer based clustering method(AP),applying AP clustering method to extract diseased regions after superpixel segmentation(4)The automatic extraction system of artifacts based on orthoimages is constructed.In the aspect of orthophoto generation,there are two ways to generate different orthophotos according to the data source.Method 1: construct a depth image model by point cloud and image registration,and then generate an orthophoto by digital differential correction;Second: the image data is generated by the empty three solution to generate a dense point cloud,and then the digital model is used to correct the depth model generated by the dense point cloud to construct an orthophoto.In the aspect of disease orthophoto segmentation,this paper chooses SLIC superpixel segmentation method by comparing several superpixel segmentation algorithms.By analyzing the limitations of the existing SLIC superpixel segmentation algorithm,the algorithm is optimized and improved.In order to improve the fit of SIIC segmentation results at the edge,the segmentation algorithm is based on the number of initial cluster centers and compactness.The improvement is mainly as follows: the iterative calculation of the compactness in the SLIC algorithm is carried out through the pre-given initial clustering center,and the edge recall rate is used to improve the super-pixel initial cluster center compensation and adjust the boundary to make the super The boundaries of the pixel clusters gradually approach to true,and the conditions are finally satisfied by iteration.In the aspect of super pixel segmentation clustering,this paper adopts the AP clustering method to make the over-segmented interest regions re-cluster according to the given conditions,so as to obtain the disease area.The main contents include: taking the number of super-pixels as the data set,constructing a similar metric matrix according to the Euclidean distance of the color space for each super-pixel block,and iteratively calculating according to certain clustering rules,and finally obtaining the clustering result.In order to realize the automatic extraction of cultural relics,this paper constructs an automatic disease extraction system,which includes functions such as image enhancement,template matching,super pixel segmentation,disease clustering and statistics.In this paper,the information of disease is extracted by orthophoto data,which involves the method of generating orthophotos.After optimizing the SLIC,it can better imagesegmentation of cultural relics,and the edge fit is higher.The method based on information transfer clustering clusters the superpixels of the over-divided diseased area so that the area size and the boundary length of the diseased area can be obtained from the result.On this basis,this paper develops a system for the extraction and research of cultural relics,which can meet the daily research work of cultural relics researchers,and lays a foundation for the continuous monitoring of cultural relics in the future. |
PDF Full Text Request