Research On The Design Methods Of Spectral Imaging Systems Based On The Image Chain Approach

Spectral imaging is a cutting-edge technique which combines spectroscopy andimaging technology to detect the world. It simultaneously captures both spatial andspectral information from the objects of interest. Since its birth dated back to the1980’s, imaging spectroscopy has found successful applications in many fields, e.g.military, agriculture, geology, environmental sciences, biology and etc. This thesisfocuses on the research of spectral imaging system design and global optimization. Adesign method based on the image chain analysis is presented. In this design strategy,each link in the image chain of spectral imaging process is mathematically modeledfor the image simulation and analysis. During the R&D phase of spectral imagingsystems, the simulation and analysis results are made use of to help support the design.The main research aspects in this thesis include:1. Summarizing the state-of-the-art researches in spectral imaging and imagechain analysis and discussing the prospect of applying the image chain approach tothe design of spectral imaging systems.2. Introducing basic theories in spectral imaging and analyzing the currentspectral imaging technologies and their pros and cons.3. Modeling the image chain of spectral imaging mathematically in itsapplication in remote sensing.4. Designing a software package called ISAPET (Image Simulation, Analysisand Performance Evaluation Tool) in MATLAB based on the mathematical models ofspectral imaging systems. Its architecture and function are introduced in detail. Forseveral kinds of spectral imaging systems, simulation examples are presented toillustrate the implementation of ISAPET’s basic functions.5. Introducing the applications of ISAPET in spectral imaging system design. 6. Illustrating the design methods of spectral imaging systems based on theimage chain approach in detail by taking a filter-array-type airborne imagingspectrometer for example. A prototype was built up with the very design result.Experiments were carried out to compare the actual results and the simulation results.It would be concluded that by applying the image simulation and analysismethods in spectral imaging system design, global optimization can be achieved toassist in designing the whole system, shortening the design cycle and improving thedesign efficiency. With the rising requirements for higher spatial and spectralresolutions in spectral imaging applications, the design strategy based on image chainapproach will play more and more important roles in spectral imaging system design.