Research On "L" Shape Pixel Infrared Super-resolution Technology Based On MTF Evaluation Method

How to improve the resolution of the imaging system has been a hot research field of aerial remote sensing. The resolution of an imaging system is limited by both optics and detectors. When the aperture and the Optical Transfer Function of the system is fixed, the smaller the detector pixel the higher the system resolution. However, other factors such as physical structure, fabrication difficulty, detector sensitivity and Signal-to-Noise Ratio constrain the smallest size of the CCD detector, and more so for infrared systems with longer wavelength.In terms of the problem how to further improve the detector resolution and make the whole system resolution not being limited by it. We proposed a L-shape super resolution method by reshaping the detector and using the existing oversampling technique at the same time. This method increases both the sampling rate and absolute cut-off frequency of the detector effectively, so that it could enhance the imaging quality of an aerial camera. It provides a theoretical basis and technical reserves for the independent development of high-resolution infrared camera program in China.During the research, we have done some meaningful analysis and got some theoretical and experimental results:1. Through large amounts of data research on the existing super-resolution technology both home and abroad, we divided it into three categories ----icro-scanning technology, sub-pixel imaging technology and hybrid computing techniques.2. Three are three factors affecting the imaging system resolution in theory. They are optical cutoff frequency, sampling Nyquist frequency detectors and detector absolute cutoff frequency. Through conducting a detailed analysis and discussion of relations among the three factors, it was concluded that despite the existing super-resolution technology can effectively reduce the detector spectral aliasing phenomenon, there are still some limitations.3. According to the special structure of TDI CCD, we explain the reason why the resolution in the along scan direction is lower than that in the cross scan direction and model the innate smear in the along scan direction. Influence on the MTF is also considered.4. Regarding the limitations of the existing super-resolution technology, we proposed L-shape element focal plane coding method. MTF calculations of the L-shape detector were derived in detail. Compared to the conventional rectangular detector, this method could improve the MTF cutoff frequency by2times.5. Image simulations were performed to validate this new technology. Also a MWIR (Middle Wavelength Infrared) experiment was conducted. The system includes a a6000-mm f/10collimator and a50-mm IR lens. We use46um×46μm infrared detector which is100%fill factor and being reshaped to photo the resolution target; The result shows the resolution has been improved at least by1.78times.