Semantic Network-Based Impervious Surface Extraction Method for Rural-Urban Fringe From High Spatial Resolution Remote Sensing Images

Impervious surfaces, as a key indicator of urban spatial environmental factors, have great significance in exploring the distribution law and pattern rural-urban fringe areas. To handle increasingly rich feature information complicated structure high resolution remote sensing images (HSRRSIs), semantic network model-guided extraction method for HSRRSI impervious surfaces fringes is proposed. The proposed mainly includes three parts: First, construction model ground covers dimensionality reduction its features. Second, optimization multi-scale segmentation algorithm based on estimation scale parameter 2 fitness function. Third, proposal ReliefF selection spectral, texture, geometry features to reduce data redundancy HSRRSIs. Finally, with Geoeye-1 image Zhanggong District source, CART, RF, SVM classifiers are used extract two different areas (named Q1 Q2), comprises edge densely distributed industrial plants, Q2 pronounced transition from rural Results show that highest surface accuracy classifier obtained when at 210 215. producer overall (94.27%, 86.41%) (94.46%, 89.47%), respectively.