Study On Push-Broom Imaging System In TIR Of Split-Window Channels In 8~9.3 μm

The land surface temperature is the result of energy exchange between the earth and the atmosphere for a long term as well as an important parameter for the status of the resource and environment on the earth. The land surface temperature can be obtained by setting two TIR(thermal infrared) detection channels in the atmospheric window of 8-14 μm and eliminating the influence of the atmospheric radiation according to their different atmospheric absorption, which is called “split-window” retrieval algorithm. The two channels have to own comparative detection sensitivities and spatial resolutions as well as obviously different atmospheric absorption. This algorithm has many advantages, such as few required data, simple calculation process and relatively high retrieval accuracy, which make it an ideal choice.In tradition, based on the optical mechanical scanning mode, the spaceborne TIR remote sensing instrumens chose the single-element detector, whose response waveband is mostly between 10-13 μm. So the surface temperature retrieval in the use of remote sensing images of split-window channels in 10-13 μm is always a very important research content. As the development technology of the long linear array infrared detector improves, the use of push-broom scanning mode has become a trend of spaceborne infrared remote sensing instrument. It has the advantages of longer residence time, higher spatial resolution and so on. The long linear array infrared detector uses liquid phase epitaxy technology. And at present, the detectors whose response waveband is between 8-10 μm have large advantages in performance and maturity, while their refrigeration requirements is relatively lower, compared with ones whose response waveband is between 10-13 μm. So there is a practical significance of the research and exploration of surface temperature retrieval in the use of split-window channels’ images in 8-10 μm.The dissertation focuses the research target at the TIR split-window imaging system, which is carried on the “Wide-waveband Imaging Spectrometer”, in the use of Hg Cd Te infrared linear array detectors and FPGA(Field-Programmable Gate Array). In the request of “high sensitivity” and “wide FOV(Field of View)”,an implementation scheme for the TIR split-window imaging system was used, which consists of push-broom scanning mode, shared optical system, FOV joint, adjustable integration time and multiply accumulate. The image system was developed. The system performances were analyzed in theory and tested by experiments. The wavebands of the split-window channels in this dissertation are 8.125-8.825 μm and 8.925-9.275 μm, and they are different with traditional ones which are in the range of 10-13 μm. Based on the above difference, many split-window algorithms have been researched and a related calculation formula is proposed with the accuracy verified, which can be refered to for the application of land surface temperature retrieval in the furture. A variety of means for verifying the feasibility and validity of split-window in the 8-9.3 μm have been used. And the image preprocessing methods are introduced, based on the optical and mechanical features of the imaging system, including image registration and image restoration, which are suitable for this system.Aiming at the technological development of the new generation of remote sensing imaging systems for China’s manned space project, the dissertation proposed a split-window calculation model in 8-9.3 μm and verified the accuracy. It has an outstanding significance for engineering applications and highly academic value. The main research contents and creative highlights of the dissertation can be summarized as follow:1) The sea surface temperature retrieval algorithm based on the split-window channels of 8.125-8.825 μm and 8.925-9.275 μm was researched for the first time in our country, with the related calculation model built. The selection of wavebands of split-window channels was expanded. The accuracy was verified by MODTRAN and better than 0.4 K.2) Three classical split-window retrieval algorithms in 10-13 μm were induced, and the transportability was proved. And the difference of split-window waveband setting of 8-9.3 μm and 10.25-11.65 μm were showed in the use of images of ASTER, which illustrated the validity of the waveband setting of the dissertation.3) Based on the optical mechanical implement scheme, the kinds of mismatch transformations were analyzed and the method and parameters for image registration were given. The effect of image degradation for temperature retrieval was analyzed though real remote sensing images. The retrieval accuracy for specific objects was improved by image restoration and radiant quantity modification.