LULC change detection analysis of Chamarajanagar district, Karnataka state, India using CNN-based deep learning method

• Published in: Advances in space research 2024-12, Vol.74 (12), p.6384-6408
• Main Authors: Mahendra, H.N., Pushpalatha, V., Mallikarjunaswamy, S., Rama Subramoniam, S., Sunil Rao, Arjun, Sharmila, N.
• Format: Article
• Language: English
• Subjects: Change detection; Convolutional neural network; Deep learning; Geographic information system; Land use land cover; Remote sensing
• ISSN: 0273-1177
• DOI: 10.1016/j.asr.2024.07.066

•The objective of the work is to develop CNN-based DL model for LULC classification.•An assessment of LULC is carried out for the classified maps of 2011 and 2021.•The change detection analysis of Chamarajanagar district is performed for one decade.•The proposed method is outperformed, with an accuracy of 95.27 % and 94.57 %. The change detection analysis of land use land cover (LULC) is an important task in several fields and applications such as environmental monitoring, urban planning, disaster management, and climate change studies. This study focuses on the use of remote sensing (RS) and geographic information systems (GIS) to identify the changes in Chamarajanagar district, which is located in Karnataka state, South India. This paper mainly focuses on the classification and change detection analysis of LULC in 2011 and 2021 using linear imaging self-scanning sensor-III (LISS-III) satellite images. Traditional methods for LULC classification involve manual interpretation of satellite images, which provides lower accuracy. Therefore, this paper proposed the Convolutional Neural Network (CNN)-based deep learning classification method for LULC classification. The main objective of the research work is to perform an accurate change detection of the Chamarajanagar district using the classified maps of the years 2011 and 2021. The proposed classification method is outperformed, with a classification accuracy of 95.27 % and 94.57 % for LISS-III satellite imagery of the years 2011 and 2021 respectively. Further, change detection analysis has been carried out using classified maps and results show a decline of 3.23 sq. km, 22.7 sq. km, and 3.83 sq. km in the areas covered by vegetation, agricultural land, and forest area, respectively. In other classes, such as built-up, water bodies, and barren land, an increase in land cover was observed by 5.59 sq. km, 1.99 sq. km, and 20.92 sq. km, respectively.