Department of Environmental Science, PSG College of Arts and Science, Coimbatore, Tamil Nadu, 641014, India.
Department of Environmental Studies, Kannur University, Mangattuparamba Campus, Mangattuparamba, Kerala, 670567, India.
Environ Sci Pollut Res Int. 2022 Dec;29(57):86349-86361. doi: 10.1007/s11356-022-18707-z. Epub 2022 Feb 4.
The present study undertakes to produce the land use/land cover map and to explore the change detection analysis of Noyyal watershed, Coimbatore, for a time period of 18 years. Based on the remote sensing and geographical information system for monitoring the temporal variations of land use/land cover, multi-temporal Landsat satellite 30-m spatial resolution images of Landsat 4/5 MSS and TM (1999), Landsat 7 ETM + (2008), and Landsat 8 Operational Land Imager (OLI) were obtained from the USGS website. The satellite images were geocoded into the universal transverse mercator (UTM) coordinate system zone 43 N. The unsupervised classification method was done by using an iterative self-organizing data analysis algorithm to compare the images and to classify the images into various land cover categories. Kappa statistics were used to assess the validation of the present study. The analysis suggests the total forest covered in 1999 was 22.69% and that of 2008 was 24.04% and reduced to 6.09%, in 2017. The agricultural land of 17.8% is reduced to 3.11% in 2008 and 0.86% in 2017. The settlements increased from 15.59 to 24.21% in 2008 and 27.14% in 2017. Increase in deforestation leads to increase in barren land. In 1999, the percentage of barren land was 17.2%; in 2008, it was 13.19%, and 50.93% in 2017. The overall accuracy estimation of the study is 73.19% and Kappa coefficient is 0.72. This study has proven a substantial strength of agreement for the map of 2017 from the result of validation rating criteria of Kappa statistics.
本研究旨在制作土地利用/土地覆盖图，并探索 18 年来 Coimbatore 的 Noyyal 流域的变化检测分析。基于遥感和地理信息系统监测土地利用/土地覆盖的时间变化，从美国地质调查局网站获取了多时期 Landsat 卫星 30 米空间分辨率的 Landsat 4/5 MSS 和 TM（1999 年）、Landsat 7 ETM+（2008 年）和 Landsat 8 运行陆地成像仪（OLI）卫星图像。将卫星图像地理编码到通用横轴墨卡托（UTM）坐标系第 43 区。使用迭代自组织数据分析算法进行无监督分类方法，以比较图像并将图像分类为各种土地覆盖类别。kappa 统计用于评估本研究的验证。分析表明，1999 年总森林覆盖率为 22.69%，2008 年为 24.04%，2017 年降至 6.09%。1999 年，荒地百分比为 17.2%；2008 年为 13.19%，2017 年为 50.93%。2008 年农业用地减少到 3.11%，2017 年减少到 0.86%。2008 年和 2017 年，定居点从 15.59%增加到 24.21%和 27.14%。森林砍伐的增加导致荒地的增加。1999 年，农业用地减少到 3.11%，2017 年减少到 0.86%。该研究的总体精度估计为 73.19%，kappa 系数为 0.72。本研究通过 Kappa 统计验证评级标准的结果，证明了 2017 年地图具有很强的一致性。
