Infrared And CCD Image Fusion Research And Hardware Design Based On FPGA

For the same scenes or targets, we use more than or equal to two image sensors to collect image information, which are different owing to the used sensors but there has some redundant information and also a complementary relationship among them. Importantly, the image fusion is made for them, the multiple sources image. It makes a full use for all those and aims to get the more accurate, more comprehensive, more reliable image description.For problems that the texture details have lost in visible light image while the resolution for targets is low in infrared image, which all occur after the traditional fusion methods, this paper puts forward a fusion method which makes a combination between Non-subsampled Contourlet Transform(NSCT) and Wavelet Transform(WT). Firstly, it makes a contrast enhancement for infrared image. Secondly, it does NSCT decomposition to visible light image and the strengthened infrared image, also uses wavelet decomposition for low frequency,which comes from NSCT decomposition. Then, for high frequency sub-bands after NSCT decomposition and low or high frequency sub-bands after wavelet decomposition, it applies different fusion rules. Finally, it will get the fused image by wavelet and NSCT reconsitution.The experiment shows that this method has a better fusion effect on the retaining of rich texture details in visible light image and highlighting for targets in infrared image.Based on the analysis for the theory of digital image fusion, this paper makes the FPGA design for image fusion by the using of lifting wavelet algorithm. Features such as fast and compatible of lifting wavelet algorithm, make it easy for hardware to implement. This design uses the Verilog hardware description language to make the RTL description for lifting wavelet module, which can be integrated completely. Also it does simulation and implementation to each module under the QuartusII 10.0 circumstances, which is offered by Altera company and is a software for FPGA. The results say that this design can achieve the image fusion better.