Method of Extracting Irrigation and Drainage Ditches Based on High Remote Sensing Image

Farmland water conservancy is an important content of land consolidation projects which scientific and effective supervision can ensure projects carried out smoothly.According to distribution and geometric characteristics of irrigation and drainage ditches in high resolution remote sensing image，an approach to extract irrigation and drainage ditches from high resolution remote sensing image was presented.Spectral range of irrigation，drainage ditches and restrain the noise by binaryzation and median filter were analyzed.Rest of ditches which covered by the border trees were extracted by Canny algorithm.Taking Dalate′s land remediation project area of Inner Mongolia as an example，taking World View-2 as the data source，system was carried out to verify the proposed method.Result showed that the extraction accuracy was 91.46%，which met the demand of land consolidation projects.This method solved the problem that current main algorithms could not extract micro bending and tree-coved ditches accurately.It also provided an effective technical supporting farmland grading and farmland linear-object facilities monitoring.