A NEW COMPUTER APPROACH TO MIXED FEATURE CLASSIFICATION FOR FORESTRY APPLICATION

摘要

Theory "and application of a new computer approach for mapping mixed forest features (i.e., types, classes) from computer classification maps are pre¬sented. Because standard statistical pattern recognition techniques fail to detect mixed features in U.S. Forest Service defined mixed stands, a new approach is proposed by which mixed features such as mixed softwood/ hardwood stands are treated as admixtures of softwood areas and hardwood areas. Large-area mixed features are identified and small-area features neglected when the nominal size of a mixed feature can be specified. At the same time, the computer program merges small isolated areas into the surrounding areas. This approach is accomplished by the iterative manipu¬lation of the postprocessing algorithm that eliminates small connected sets. For a forestry application, computer-classified Land Satellite multispectral scanner data of the Sam Houston National Forest were used to demonstrate the proposed approach. The' new technique was verified to be successful in cleaning the salt-and-pepper appearance of multiclass classification maps and in mapping those mixed softwood/hardwood areas that were admixtures of softwood areas and hardwood areas. However, for that particular application, the computer-mapped mixed areas matched very poorly with the ground truth because of inadequate resolution and inappro¬priate definition of mixed features for Land Satellite data analysis.