High-resolution remote sensing image road extraction method based on visual saliency detection
A remote sensing image and road extraction technology, which is applied in image enhancement, image analysis, image data processing, etc., can solve the problems of saliency detection process load and insufficient accuracy of road extraction, so as to save running time, improve accuracy, and improve detection algorithm simple effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0066] like figure 1 As shown, the road extraction method of high-resolution remote sensing images based on visual saliency detection includes the following steps:
[0067] S1. Obtain the remote sensing image of the road to be extracted and preprocess it to obtain the image block of the remote sensing image;
[0068] S2. Carry out image saliency detection based on boundary probability on the image block of remote sensing image, and generate a saliency map based on boundary weight;
[0069] S3. Carrying out saliency detection based on background weighted contrast ratio on the image block of remote sensing image, and generating a saliency map based on background connectivity;
[0070] S4. Fusing the saliency map based on the boundary weight and the saliency map based on the background connectivity to obtain the fused saliency map;
[0071] S5. Perform binarization processing on the fused saliency map, and use the binarized image as a mask to extract salient road areas to reali...
Embodiment 2
[0073] In the above-mentioned embodiment 1, when performing salient target detection on remote sensing images, the shadows of ground objects will usually seriously affect the effect of the image saliency detection algorithm. To improve the quality of the remote sensing image and the effect of the saliency detection algorithm, in the embodiment of the present invention, the global adaptive shadow detection algorithm is used to detect and remove the shadow of the remote sensing image. Therefore, step S1 in embodiment 1 is specifically:
[0074] S11, performing denoising and color space conversion processing on the remote sensing image in sequence;
[0075] Specifically, in the case of figure 2 When processing the remote sensing image shown, the guided filter is used to filter the full-frame image, and then the image is transformed from the RGB color image to the CIE Lab color space;
[0076] S12. Perform shadow detection on the remote sensing image after the color space conver...
Embodiment 3
[0083] In the above-mentioned embodiment 1, an important feature of the road in the remote sensing image is the edge feature with clear outline. In order to improve the detection accuracy of the salient area, the embodiment of the present invention uses the salient detection method based on the boundary weight to obtain the corresponding A saliency map based on boundary weights is used as a saliency priori map; therefore, step S2 in Embodiment 1 is specifically:
[0084] S21. The convex hull detection algorithm based on Harris points is used to detect the convex hull of the image block of the remote sensing image, and determine the salient points;
[0085] S22. According to the relationship between the superpixel corresponding to the salient point and the detected convex hull, calculate the corresponding boundary probability mean value;
[0086] S23. Perform image boundary saliency detection on the image block of the remote sensing image based on the calculated mean value of t...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More