Volume XLII-1/W1
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-1/W1, 387-392, 2017
https://doi.org/10.5194/isprs-archives-XLII-1-W1-387-2017
© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-1/W1, 387-392, 2017
https://doi.org/10.5194/isprs-archives-XLII-1-W1-387-2017
© Author(s) 2017. This work is distributed under
the Creative Commons Attribution 3.0 License.

  31 May 2017

31 May 2017

A DETECTION METHOD OF ARTIFICIAL AREA FROM HIGH RESOLUTION REMOTE SENSING IMAGES BASED ON MULTI SCALE AND MULTI FEATURE FUSION

P. Li1, X. Hu2, Y. Hu1,2, Y. Ding1, L. Wang1, and L. Li1 P. Li et al.
  • 1Chongqing Geomatics Center, Chongqing, China
  • 2School of Remote Sensing and Information Engineering, Wuhan, China

Keywords: Artificial Areas, Automatic Detection, Geometric Features, Visual Saliency Features, Multi Scale, Feature Fusion

Abstract. In order to solve the problem of automatic detection of artificial objects in high resolution remote sensing images, a method for detection of artificial areas in high resolution remote sensing images based on multi-scale and multi feature fusion is proposed. Firstly, the geometric features such as corner, straight line and right angle are extracted from the original resolution, and the pseudo corner points, pseudo linear features and pseudo orthogonal angles are filtered out by the self-constraint and mutual restraint between them. Then the radiation intensity map of the image with high geometric characteristics is obtained by the linear inverse distance weighted method. Secondly, the original image is reduced to multiple scales and the visual saliency image of each scale is obtained by adaptive weighting of the orthogonal saliency, the local brightness and contrast which are calculated at the corresponding scale. Then the final visual saliency image is obtained by fusing all scales’ visual saliency images. Thirdly, the visual saliency images of artificial areas based on multi scales and multi features are obtained by fusing the geometric feature energy intensity map and visual saliency image obtained in previous decision level. Finally, the artificial areas can be segmented based on the method called OTSU. Experiments show that the method in this paper not only can detect large artificial areas such as urban city, residential district, but also detect the single family house in the countryside correctly. The detection rate of artificial areas reached 92 %.