Wavelet domain image fusion algorithms with applications
Date of Award
Doctor of Philosophy (PhD)
Electrical Engineering and Computer Science
Pramod K. Varshney
Wireless, Multisensor images, Weapons detection, Wavelet, Image fusion
Electrical and Computer Engineering
A single imaging sensor, depending on its features and/or viewing position, may not provide all the information necessary for a particular application. If multiple imaging sensors with different properties are available, then the image information obtained from each sensor can be combined to provide a higher quality image. Consequently, image fusion techniques have emerged for combining multisensor images to form a single composite image. This final composite image may have more complete information content or better perceptual quality than the individual source images. The image fusion algorithms discussed in this dissertation utilize multiresolution image decomposition and reconstruction to isolate the image details at different scales. Image fusion rules are applied to the source transforms to calculate the coefficients of the fused transform. The composite fused image is then reconstructed by performing the inverse pyramidal transform on the composite transform.
The objective of this research is to develop multiresolution image fusion algorithms for different applications using wavelet decomposition and wavelet domain fusion rules. Specifically, this dissertation introduces the use of image fusion for two different applications. The first application is the Concealed Weapons Detection problem where the objective is to utilize different types of imaging sensors to determine if a weapon is hidden underneath a person's clothing. The second application is error correction/concealment in wireless image communications for both compressed and uncompressed image data. A wireless channel is subject to channel impairments such as multipath and fading which results in images of poor quality. The objective of the image transmission problem is to use image fusion techniques to provide an image of good perceptual quality at the receiver. Diversity is used to obtain multiple image bit streams at the receiver and diversity combining is performed in the wavelet domain to combat the wireless channel errors. Results indicate that this diversity combining method enhances performance when used in addition to error control coding and interleaving.
Surface provides description only. Full text is available to ProQuest subscribers. Ask your Librarian for assistance.
Ramac, Liane Carmela, "Wavelet domain image fusion algorithms with applications" (2000). Electrical Engineering and Computer Science - Dissertations. Paper 172.