Improvement of Forest Canopy Density Model Using Remote Sensing Data Integration

Abstract

Forest Canopy Density Mapper is a method based on spectral indexes integration in forest canopy density classification. In this paper, a data integration procedure is used to improve the result. In this respect, SFIM method and spectral response algorithm is utilized without a bad effect on the spectral and radiometric properties of bands. In the following, Landsat images of Hyrcanian forests in the north of Iran were used to implement the conventional and improved methods. Also, the ground measurements including grass-land, thin forest, semi-dense forest and forest is utilized for evaluation. The result shows that the forest canopy density model is inefficient in the thin and semi-dense forests. Alternatively, the results in the dense forest and grass land is reliable. Additionally, the improvement of the proposed method in these two areas is clearly seen. It seems that a high resolution image should be used to improve the accuracy of the forest density classification in the semi-dense and thin forests.

Keywords

• spectral response
• data integration
• classification accuracy
• forest density classes
• forest canopy density mapper