Research On Gully Erosion Parameters Extraction Based On Multi-Source Date Fusion

In recent years, with the development of spatial information and measurement technology, especially these high image scale, stereo photogrammetry, radar interferometry measurements emerge, RTK-GPS measurements, three-dimensional laser scanning and other new technology, which for gully erosion study provides massive multi source data, broaden the ways and means of gully erosion studies.However, these data were applied to gully erosion alone and lacked of real integration,which with the application of a variety of data only exists in the independent cross-validation. Therefore, the first problem which need to address was seamlessly integrate data from multiple sources to construct gully parameter extraction method.Taking the west coast watershed of Anli Reservoir in Qixia City Shandong province as study area, we discussed the fusion technology of multi-source remote sensing data and established the method of extract gully parameter, which was used to broaden gully erosion research methodsand to improve the accuracy of research gully. The results showed that:(1) Compare the three fusion method of panchromatic and multi-spectral image:Hsv transform image had the highest brightness, which led to the serious loss of spectrum and poor visual effect; Pca image had a rich information with the worst clarity, a greater degree of spectral loss and the poorest visual effect; Gram-Schmidt spectral sharpening image had the best visual effect, a good ability to maintain spectrum, a rich information and a good clarity, which was the best fusion method for panchromatic and multi-spectral image of ZY3.(2) The gully extracted basing on the Gram-Schmidt spectral sharpening method,hsv transformation and pca transformation fused image were more accurate than the gullies extracted by original multispectral image in number and area; Based on RTK-GPS field measurement gully to test the gullies parameters extracted by image fusion, which showed that the gully parameter extracted by Gram-Schmidt spectral sharpening image were closer to measured gully parameters, had the highest accuracy.Gram-Schmidt spectral sharpening method was a suitable method for combining the pan and multispectral image and was the most suitable for gully extraction with thebest accuracy.(3) Using SRM multiplication fusion, Gram-Schmidt spectral sharpening method and HSV transform to fuse the multi-spectral images and DEM, which showed the three methods can remove the false topography effectively:multiplying the images by SRM can contain a high-resolution of the original image, but poor three-dimensional effect;the remote sensing image distortion based on hsv method was smaller than GS spectral sharpening method with abundant information, good clarity and three-dimensional effect; the GS spectral sharpening method had the better three-dimensional effect,but had the poor clarity and information. Comprehensive the the correction effect of false topographic perception, the degree of the remote sensing image distortion, information and the degree of clarity, hsv method were Suitable for rectify the false topographic perception of ZY3 remote sensing images.(4) Based on HSV transform to fuse multi-spectral images and DEM, which enhanced and enriched the topography information in the remote sensing image was constructed. Based on the fusion image and the method which combined the shoulder-line and the thalweg of gullies threshold processing of the slope and visual interpretation t, 76 gullies were extracted in the study area, with a total area of 1.13km2,the erosion density was 5.2km/km2,which indicated the study area was extremely seriously eroded; Based on the gullies extracted by DOM image, the identification accuracy of the shoulder-line and the thalweg between the interpretation gullies by fusion image and standard gullies were all 91%, which were significantly higher than extracted by original multispectral image. The gullies by fusion image had the higher accuracy than original multispectral image. Fusing the multi-spectral images and DEM date can improve the interpretation accuracy of gully.