Segmentasi Api dan Asap Pada Kebakaran Dengan Metode K-Means Clustering

Abstract

Fire is a frequent disaster, especially in tropical climates. Factors causing fires that often occur are intentional and human negligence. When fires occur, fires often get out of control and spread in the direction of global warming, climate change, landslides, and floods, causing harm to the environment. Because it is difficult to analyze the size of the fire when a fire occurs, therefore with the development of technology and sophistication, it is now possible to know the position of the fire and the thickness of the smoke in a fire using digital images. Image is another term for image as a multimedia component that plays a very important role as a form of visual information. This study aims to segment the fire image by separating 2 images, namely fire and smoke to determine the position of the fire and the thickness of the smoke in the fire. The fire and smoke segmentation research uses the K-means clustering method using the Matlab application. The stages carried out in this research include the Image Input Process, Converted Original Image to Binary, Threshold Conversion Segmentation, Image Segmentation Process. The image segmentation process for the fire and smoke fire images succeeded in separating the fire and smoke objects with a value of T3 = 61. The smoke and fire objects were in the matrix coordinates [50:100,300:200].