A Kind of Multi-Channel Filtering Based Wavelet Packet Remote Sensing Image Fusion

2012 ◽  
Vol 500 ◽  
pp. 748-753
Author(s):  
Su Rong Dong ◽  
Hai Ying Zhou

A kind of multi-scaled texture feature based wavelet packet fusion approach be presented, which applies the principle of multi-channel filtering fusion to take wavelet fusion on the image with aplenty scaled texture feature and the multispectral image, meanwhile, for obtaining diversified scaled texture feature information and high resolution image, a regional characteristics based self-adapted wavelet packet fusion criteria be proposed. The experiment result and statistics analysis shows that the improved wavelet packet algorithm get the relatively improvement in the aspects of information amount and the clarity comparing with the traditional wavelet packet fusion.

2013 ◽  
Vol 405-408 ◽  
pp. 3001-3006 ◽  
Author(s):  
Shuang Ting Wang ◽  
Chun Lai Wang ◽  
Wei Bing Du ◽  
Le Le Tong ◽  
Fei Wang

Pepper and Salt" phenomenon and misclassification phenomenon are serious and the accuracy is low based on pixel classification, when only use a single remote sensing image. In this article, joint LiDAR data and high resolution image together based on feature per-parcel classification,and in the image segmentation stage, texture feature is introduced, these can full use of spectral informationtexture feature and elevation information in classification, to solve same object with different spectra and same spectrum with different objects. Compared with the classification based on pixel, only use a single remote sensing image, the method based on feature per-parcel with spectrumtexture and elevation information achieved a high accuracy,96.94%, improved the classification result.


Author(s):  
C. Pohl ◽  
Y. Zeng

During the last decades number and quality of available remote sensing satellite sensors for Earth observation has grown significantly. The amount of available multi-sensor images along with their increased spatial and spectral resolution provides new challenges to Earth scientists. With a Fusion Approach Selection Tool (FAST) the remote sensing community would obtain access to an optimized and improved image processing technology. Remote sensing image fusion is a mean to produce images containing information that is not inherent in the single image alone. In the meantime the user has access to sophisticated commercialized image fusion techniques plus the option to tune the parameters of each individual technique to match the anticipated application. This leaves the operator with an uncountable number of options to combine remote sensing images, not talking about the selection of the appropriate images, resolution and bands. Image fusion can be a machine and time-consuming endeavour. In addition it requires knowledge about remote sensing, image fusion, digital image processing and the application. FAST shall provide the user with a quick overview of processing flows to choose from to reach the target. FAST will ask for available images, application parameters and desired information to process this input to come out with a workflow to quickly obtain the best results. It will optimize data and image fusion techniques. It provides an overview on the possible results from which the user can choose the best. FAST will enable even inexperienced users to use advanced processing methods to maximize the benefit of multi-sensor image exploitation.


Sign in / Sign up

Export Citation Format

Share Document