Estimation of canopy closure based on improved dimidiate pixel model

ObjectiveThe canopy closure is an important factor in forest resource investigation. It can not only reflect the closure degree of forest canopy, the degree of forest using space, but also indicate the density of the forest. Remote sensing provides an unprecedented opportunity for accurate estimation of canopy closure at regional and global scales, making it possible to monitor and map large area canopy closure. This study aims to estimate canopy closure of Larix gmelini on regional scale through the ZY-3.

MethodIn this study, Larix gmelini in Mengjiagang of Huanan County, Jiamusi City of Heilongjiang Province, northeastern China was taken as the research object. Firstly, dimidiate pixel model was used to estimate vegetation coverage. By exploring the relationship between vegetation coverage and canopy closure, the dimidiate pixel model was improved. The normalized difference vegetation index (NDVI) was calculated by calculating the cumulative frequency as model parameters. The optimal model was obtained by comparing the model fitting effect of different parameters, and the canopy closure was estimated by the optimal model.

ResultThe fitting results showed that the fitting effect of 2% cumulative frequency was the best, with R² was 0.838 and RMSE was 0.045. Finally, the model was used to obtain the distribution map of the canopy closure of Larix gmelini in Mengjiagang.

ConclusionThe results show that improved dimidiate pixel model can get accurate estimation of canopy closure. This study provides a more effective way to estimate the canopy closure of Larix gmelini in the north of China, and also provides a reference for the investigation of forest resources and parameters.