Estimation of forest stand volume using textural indices of aerial images

Abstract

Because of several radiometric errors, precise estimation of forest stand volume based on spectral indices is not achievable. In contrast to spectral indices, textural indices are more consistent dealing with these errors. In this research, estimation of forest stand volume based on textural indices was studied. For this aim, 150 plots were collected using systematic random design. Green, red and near infra red bands were used. Textural indices included second moment, contrast and homogeneity extract by different window size. Appropriate band, index and window size were chosen by stepwise regression. Based on this analysis, near infra red band, homogeneity index and 31×31 pixel window size were selected. RMSE and bias of estimation was 43 and 2 percent, respectively. Although, estimation accuracy of forest stand volume by textual indices was suitable for mapping purposes, however, its application in forestry operations needs more researches.

Keywords

• Forest stand volume
• aerial images
• textural indices
• stepwise regression