Temple Rock Damage Detection System in Digital Image at Borobudur Conservation Center

Borobudur is a temple where the building from andesite which is in an open space so that the temple stone will be susceptible to various problems that can cause stones to be damage and weathering. In this research, a system can be made that can study the types of damage to objects through the image file being tested, the segmentation method used K-means clustering, the method of texture feature extraction is Gray Level Co-occurrence Matrix (GLCM) and the classification method used K-Nearest Neighbor (KNN). In this test used 70 types of rocks that have causes of damage (alveol, microorganisms, salting) which are divided into two data, training data and testing data with a composition of 44 training data and 26 testing data. Then the testing of the test image dataset was 26 images consisting of 8 alveol images, 11 microorganisms images, and 7 salting images. The highest level of accuracy was obtained at 57.69% using the GLCM degree parameter θ = 45° and k = 7.