Abstract

The canopy density of forests is highly infuenced by population pressure which cause forest fragmentations, deforestations, forest degradation and also land reclamation for infrastructure and agriculture. This study was envisaged with the objective of mapping the forest canopy density with two diferent methods by using Landsat 8 OLI dataset of the year 2016 after mapping the vegetation indices. One of the two methods is the semi-expert FCD mapper model, while the other model is newly developed by us and consists of eight vegetation indices. After running the models, the results of both the models were compared. It was found that for the semi-expert FCD model, the three classes viz. high canopy density, moderate canopy and low canopy covered an area of 81,615.51 ha (40%), 84,474.72 ha (41%) and 38,844.18 ha (19%), respectively. And for the modifed FCD model, the same classes covered 69,134.670 ha (37%), 84,062.250 ha (45%) and 32,529.150 ha (18%), respectively. It was observed that the diference between semi-expert FCD model and the modifed FCD model’s accuracy is about 1.75% and diference in Kappa statistics is 0.0362. Thus, the modifed model is more accurate than the semi-expert FCD model and gives us more detailed canopy density map than the semi-expert FCD map.