Change detection method for remote sensing images based on multi-resolution nmf and treelet fusion
A remote sensing image and change detection technology, applied in the field of image processing, can solve problems such as being easily affected by noise, missed detection or many false alarms, and cannot handle nonlinear related data well, so as to achieve reasonable data interpretation and improve accuracy , the effect of suppressing adverse effects
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Publication Date
- 2014-10-01
- Estimated Expiration
- Not applicable · inactive patent
Smart Images
Figure 1 Figure 2 Figure 3
Abstract
Description
technical field
[0001] The invention belongs to the technical field of image processing, relates to remote sensing image change detection, mainly relates to multi-resolution non-negative matrix factorization (Non-negative Matrix Factorization, NMF) and Treelet fusion remote sensing image change detection, specifically a multi-resolution NMF and A method for change detection in remote sensing images based on Treelet fusion. It is used to detect the changing regions of remote sensing images. Background technique
[0002] Remote sensing image change detection can obtain the dynamic development and change information of ground objects by analyzing the differences in spectral characteristics or spatial structure characteristics of remote sensing images in different time periods in the same area. The change detection of remote sensing images has been widely used, such as resource monitoring of land, forest, grassland, etc., crop yield estimation and pest monitoring, map data corr...
Examples
Embodiment 1
[0062] The present invention is a remote sensing image change detection method that combines multi-resolution NMF and Treelet. The experimental simulation environment is: MATLAB R2010a, CPU Inter Core2 1.80GHz, memory 2G, Windows XP Professional, refer to figure 1 , the remote sensing image change detection method includes the following steps:
[0063] (1) Take two registered remote sensing images of size P×Q acquired at different times in the same area, and the phase 1 image is X 1 ={X 1 (p,q), 1≤p≤P, 1≤q ≤Q}, see figure 2 (a) and image 3 (a), phase 2 image is X 2 ={X 2 (p,q), 1≤p≤P, 1≤q≤Q}, see figure 2 (b) and image 3 (b), where X 1 (p,q) and X 2 (p, q) are the gray value of the pixel at the spatial position (p, q) of the phase 1 image and the phase 2 image, respectively. Subtract the gray value of the pixel corresponding to the spatial position of the input two-temporal remote sensing image to obtain the absolute value, and obtain a difference image X’ D (X' D...
Embodiment 2
[0098] The remote sensing image change detection method of multi-resolution NMF and Treelet fusion is the same as embodiment 1, wherein in step (3), the calculation for threshold is:
[0099] First, estimate the filtered difference image X D noise standard deviation
[0100] σ X D = Median ( | X D | ) / 0.6745 - - - ( 8 )
[0101] Among them, Median(·) represents the operation of taking the median value.
[0102] Select threshold If the gray value of the image is less than the threshold T x , then set the value to 0, if the gray value of the image is greater than the threshold T x , the value is retained. Obtain an image Y in which the gray value of the background pixel is 0 and t...
Embodiment 3
[0111] The remote sensing image change detection method of multi-resolution NMF and Treelet fusion is the same as embodiment 1-2, wherein the value of the maximum number of iterations e in step 2b) is 50, and the value of stop precision ε is 10 -4 , the value of the current iteration number b is 50, in this case, the calculation speed is faster, and the detection accuracy is slightly lower.
[0112] In addition, the threshold T in step (5) R A value of 20 can also get better change detection results.