Mapping Burnt Areas Using the Semi-Automatic Object-Based Image Analysis Method

Abstract

Forest and land fires in Indonesia take place almost every year, particularly in the dry season and in Sumatra and Kalimantan. Such fires damage the ecosystem, and lower the quality of life of the community, especially in health, social and economic terms. To establish the location of forest and land fires, it is necessary to identify and analyse burnt areas. Information on these is necessary to determine the environmental damage caused, the impact on the environment, the carbon emissions produced, and the rehabilitation process needed. Identification methods of burnt land was made both visually and digitally by utilising satellite remote sensing data technology. Such data were chosen because they can identify objects quickly and precisely. Landsat 8 image data have many advantages: they can be easily obtained, the archives are long and they are visible to thermal wavelengths. By using a combination of visible, infrared and thermal channels through the semi-automatic object-based image analysis (OBIA) approach, the study aims to identify burnt areas in the geographical area of Indonesia. The research concludes that the semi-automatic OBIA approach based on the red, infrared and thermal spectral bands is a reliable and fast method for identifying burnt areas in regions of Sumatra and Kalimantan.

Opportunities and Challenges of Remote Sensing, Geospatial Data, and Machine Learning in Obtaining Accessibility and Location Information for Sustainable Development in Indonesia

“…Remote sensing data has become widely used along with machine learning as a new means of data processing and analysis. One of the remote sensing data is satellite imagery which provides spatial information on surface reflection daily and is publicly available which has proven to be applicable and helped in all kinds of fields such as crop monitoring [2], [3], or disaster relief [4]- [6]. This advancement in technology has also provided us with other forms of data other than traditional tables such as road network maps or public facilities geotagging.…”

Machine Learning Applied to Sentinel-2 and Landsat-8 Multispectral and Medium-Resolution Satellite Imagery for the Detection of Rice Production Areas in Nganjuk, East Java, Indonesia

“…18 No. 1 June 2021 & Maurel, 2017); poverty prediction (Jean, Burke, Xie, Davis, Lobell, & Ermon, 2016); burned area mapping (Fitriana, Suwarsono, Kusratmoko, & Supriatna, 2020); and crop classification (Kussul, Lemoine, Gallego, Skakun, Lavreniuk, & Shelestov, 2016).…”