Comparison of land use/land cover change of fused image and multispectral image of landsat mission: a case study of Rajshahi, Bangladesh

To monitor the impacts on environment due to climate change and anthropogenic influences, it is important to have precise and accurate information regarding the land use/land cover (LULC) change. The current study has been carried out to understand the LULC pattern and changing scenario using remote sensing and GIS techniques. The data were used in two ways, i.e. first the multispectral image was used to classify by maximum likelihood (ML) classification system on both 1999 and 2016 images separately, and then compare the change history, and second, the multispectral image and panchromatic band was fused using Gram-Schmidt (GM) fusion technique to enhance the image quality. Then, the fused image of 1999 and 2016 was classified separately using the ML process. Several statistical analyses on fused image were carried to quantify the image quality. The classified images were compared to get the most accurate result. Four land use classes were detected, namely cultivated land, vegetation, built up area, and water body. During the classification, the vegetation, built up and cultivated land classes were biased and slightly deviated from the accuracy. The maximum difference between the classified panchromatic and multispectral images in 1999 for cultivated land was 12.2 km2, and the minimum difference was found for the water body, i.e. 0.8 km2. The similarly in 2016 presented the maximum difference for vegetation equal to 7.3 km2 and minimum equal to 4.4 km2 for the water body. Accuracy result shows that the fused image is better for classified proposes for analysis and decision making.