Research on Auto-initial-value LSB-Snake Model to Extract Roads Semi-automatically
Xiaokun Zhu · Beijing Surveying and Mapping · 2008
LSB-Snake model is an effective method to extract linear object semi-automatically,but it needs manual input of road characters for extraction,and it is not robust while the initial seed points are not dense enough.These reduce the working efficiency of LSB-Snake model.This paper puts forward an auto-initial-value LSB-Snake model,which uses self-adapt template matching method to provide the road characters for LSB-Snake model,and adds the seed points based on the initial points at the same time automatically.According to experiments,it is indicated that: Given the same amount of initial seed points,the method presented is more robust than LSB-Snake model;it needn't manual input of the road character,the auto-initial-value LSB-Snake model is more automatic than LSB-Snake model;The auto-initial-value LSB-Snake model can overcome the shades or shelter of land objects such as building and trees,and it is more powerful in anti-jamming than LSB-Snake model.