Object Oriented Classification for Mapping Mixed and Pure Forest Stands Using Very-High Resolution Imagery

The importance of mixed forests is increasingly recognized on a scientific level, due to their greater productivity and efficiency in resource use, compared pure stands. However, reliable quantification the actual spatial extent stands fine scale still lacking. Indeed, classification mapping populations, especially with semi-automatic procedures, has been challenging issue up date. main objective this study evaluate potential Object-Based Image Analysis (OBIA) Very-High-Resolution imagery (VHR) detect map broadleaves coniferous trees Minimum Mapping Unit (MMU) 500 m2. This evaluates segmentation-based paired non-parametric method K- nearest-neighbors (K-NN), trained dataset independent from validation one. forest area mapped as canopies amounts 11%, an overall accuracy being equal 85% K 0.78. Better levels user producer accuracies (85–93%) are reached conifer broadleaved dominated findings demonstrate that very high resolution images (0.20 m resolutions) can be reliably used fine-grained pattern rare forests, thus supporting monitoring management resources also scales.