Design of Pothole Detector Using Gray Level Co-occurrence Matrix (GLCM) And Neural Network (NN). Issue 1 (June 2020)

• Record Type: Journal Article
• Authors: Sutrisno, Imam; Syauqi, Albiyan Wanda; Hasin, Muhammad Khoirul; Iskandar, ; Rahmat, Mohammad Basuki; Asmara, I Putu Sindhu; Suwondo, I'ie; Ardiansyah, ; Wiratno, Daviq; Setiawan, Edy
• Is Part Of: IOP conference series. Volume 874:Issue 1(2020)
• Publication Date: 2020-06
• Subjects: Materials science -- Periodicals; 620.1105
• DOI: 10.1088/1757-899X/874/1/012012
• Languages: English
• ISSNs: 1757-8981

Abstract

Abstract: Roads are land transportation infrastructure that covers all parts of the road. Roads with bad conditions will interfere with the achievement of activities to a destination. The situation also includes damage to the road surface in the form of holes. To overcome this, in this Final Project a hole detector was detected in the road using the Gray Level Co-occurrence Matrix (GLCM) and Neural Network (NN). The tool detects holes in the surface of the road using a camera by walking along the road being examined. The camera is used instead of the eye to detect road surface damage. The method used to detect holes is the GLCM. The GLCM method produces several features, namely entropy, contrast, energy, homogeneity, and correlation which will then be processed using a NN to produce a decision whether there is a hole or not. In addition to knowing where the location of the damage is equipped with GPS (Global Positioning System). The results of image feature extraction using the GLCM and road classification using NN can be used in the hole detection process. Testing is done using a car prototype that is monitored through the computer. The percentage of successful hole detection is 86.6% using 10 hidden. When a hole is detected the device manages to take a picture, then sends the hole coordinates to the server.