The Research Of Auto-Extraction Drainage System Network Based On Remote Sensing Image

The formation of drainage system is affected by geological structure and natural environment, the shapes and distributions of drainage system are determined by different geological structures. So the research on drainage system can help to know the geological structure and the lithology of mother rock. Remote sensing image has the characteristics of low cost, high overlay, fast update and good up-to-the-minute merit, so the research of drainage system basing on image has more practical significance.This paper adopts the way of extraction which drainage system is extracted by extracting water rivers automatically and extracting dry rivers emi-automatically, eventually, drainage system is extracted successfully.A new arithmetic is brought forward, which named combined condition of sensitive factors, during the course of water rivers extraction. Water rivers are extracted by combining several sensitive factors conditionally, which are created according to object characteristics. Experiments prove that this method can distinguish similar objects, realize water rivers extraction.During the course of dry rivers extraction, another new arithmetic is brought forward which named seed pixel diffused arithmetic. This arithmetic realize the extraction of dry rivers semi-automatically by taking the seed pixel as a center, compartmentalizing the space of image as eight subdomains, selecting seed pixel manually.In the course of standardization and resume of drainage system, Mathematical Morphology is adopted in the process of image processing, including thinning image, eliminating speckle noise, checking marginal, conditional inflation, resuming drainage system and this methods achieve satisfied effect.Basing on the researches and analyses aboved, a module of drainage system extraction is programmed. This module can achieve the operations of the extraction, standardization and resume of drainage system etc.