Vector distance algorithm for optimal segmentation scale selection of object-oriented remote sensing image classification

Aiming at the optimal segmentation scale selection for object-oriented remote sensing image classification, thesis brought forward a new method named vector distance index. Research verified the validity and applicability of this method by carrying out experiment at wetland area with China Brazil Earth resource satellite image. Both two experiment showed that it could realize the optimal segmentation scale selection for object-oriented remote sensing image classification. Based on the basic theory of vector distance method, aiming at ldquosubmergencerdquo and ldquofragmentationrdquo phenomenon, research further brought forward a scale index, which could reflect the segmentation scale status for given object type, and provided a quantitative tool for assessing conflict degree between the two situations.