Research And Application Of Segmentation Algorithm For Overlapping Rock Art

Rock paintings are pictures drawn by human ancestors and carved on rocks.The information contained in rock painting images is an important window for scholars to study socio-economic and ideological activities of ancient human beings.Due to longterm natural erosion and man-made damage,petroglyphs are digitized.Protection has become a must-have option for cultural relic protection departments.Using image and image processing technology to process image segmentation of rock art images is a research focus of digital protection of rock paintings.At the present stage,image segmentation technology has been widely used in the segmentation of medical images and remote sensing images and has achieved a good segmentation effect.However,image segmentation algorithms have only a few applications in rock art.In the rock paintings discovered so far,there are many overlapping rock painting images.In this type of rock painting,a plurality of individual rock painting images coexist in a rock painting.The contents of the rock painting images obstruct each other,and the separation boundaries between the individual rock art images are not obvious.At present,there is no relevant segmentation algorithm for the segmentation of overlapping rock paintings,and most of them are still done by manual segmentation method.Such a segmentation method has the disadvantage of low segmentation efficiency and can no longer meet the requirements of the existing segmented rock image segmentation research.Based on the research of image segmentation algorithm,this paper takes the Jinsha River overlapping rock paintings as the object,and applies the commonly used segmentation algorithms to the Jinsha River overlapping rock paintings.After analyzing the segmentation result maps of these algorithms,it is found that only a single image segmentation algorithm is used.The image segmentation of the Jinsha River overlapped rock paintings has the disadvantages of low accuracy,high error rate,and incomplete segmentation images.After comparing and analyzing the result graphs of commonly used segmentation algorithms,it is found that the SVM algorithm has the advantage of high image segmentation efficiency.The SRG algorithm has the characteristics of complete image segmentation.For the image features of overlapping rock paintings,the SVM and SRG fusion segmentation algorithm is proposed.The fusion algorithm is applied to the Jinsha River overlap rock paintings,and the overlapping rock art is image segmented in the HSV color space to obtain a segmentation result map of the fusion algorithm.The result graph of the fusion algorithm and the previous use of a single SVM and SRG algorithm The comparison of the segmentation results shows that the fusion algorithm has higher segmentation accuracy and lower segmentation error rate than a single SVM and SRG algorithm.