Analisis Model Pengukuran Tinggi Permukaan Air Dengan Metode Canny Edge Detection dan Image Contouring Sebagai Indikator Peringatan Dini Bencana Banjir

Abstract

Received Apr 14, 2021 Revised Agus 1, 2021 Accepted Sep 06, 2021 Flood disaster remains a natural phenomenon that often occurs in Indonesia, especially in the Wisma Tajur Housing Complex area, Tangerang City which causes property losses including the safety of the souls of the affected community. The difficulty experienced so far is how to measure the water level to obtain alert status information as an indicator of flood warning. As a solution in overcoming these problems, this research proposes a method based on digital image processing with canny edge detection algorithms and image contouring to measure river water levels. Canny edge detection and image contouring were chosen due to their accuracy in detecting the edges of the image and the ease of the computation process. The steps taken in this research are to conduct a simulation experiment of measuring the water level using a water container that can describe the situation in the river, then doing field testing. Canny edge detection produces an outline that can then be detected by the contour, then water level measurements can be made on the bounding rectangle that is formed and changes dynamically with fluctuations in water level. The contribution of this research is the use of black measuring lines that are processed using thresholding techniques to facilitate the process of measuring water level using a combination of canny edge detection and image contouring techniques as well as adding attributes/features using threshold, MinVal, and MaxVal values on the canny edge. Sampling testing produces an accuracy of 99.96%, prototype testing produces 100% accuracy, and direct testing produces an accuracy of 99.85%.