Infrared imaging technology is widely used in many departments. Above all it is used in military field. Imaging system of infrared focal plane arrays (IRFPA) is future development direction of infrared imaging system (IRIS). It is very important to develop little mobile and high clearance imaging system of IRFPA. This project is supported by NSFC (NO. 60077025).At first, domestic and foreign development situation of infrared imaging is expounded in the paper, and development trend of IRIS is analyzed. At last importance of self-developing IRIS is stressed.In infrared imaging field it is valuable research to obtain real-time infrared image. IRFPA is widely used in IRIS, but the nonuniformity from IRFPA greatly limits the infrared image quality. In order to improve quality of infrared image the nonuniformity must be corrected. Before studying on nonuniformity correction ways the hardware real-time system is developed. The real-time system is made up of PC and DSP card. PC is used for host, while DSP card is auxiliary part. PC is mainly used for image display and system synchronic control. DSP card is used for video signal sampling and image data processing. To increase imaging system real-time speed, a new data joint transfer method is given out. Not only does realization of the new method raise the efficiency of transferring data, but also make the best of finite memory. In the real-time imaging system, for applying new method the data transfer neck is solved, while larger numbers of image data can be stored in internal memory of DSP, and this speed up accessing image data. So data joint transfer greatly improves data transferring and processing rate.Real-time nonuniformity corrections and image enhancements are studied on basis of hardware platform. In this part main work includes two points real-time correction, many points real-time correction, many points real-time compress correction, histogram equalization processing and so on. Based on thoroughly studying ordinary corrections many points real-time compress correction way is proposed. At first many points calibration data must be compressed before many points compress correction. After data compression the complicated arithmetic is simplified and the computation speed from DSP is evidently increased. The many points real-time compress correction is the fittest to real-time image processing among the three methods. It has advantages both from two points correction and from many points correction, because the two points correction has less data and easy real-time correction but the many points correction has clearer image. After nonuniformity correction of IRFPA, image enhancement processing is implemented, for infrared image has lower contrast it cannot meet user's need only by correction, besides there are other reasons that the whole gray range from 0 to 255 can not be covered, simultaneously black and white dots are not corrected. While the image enhancement processing can eliminate the above inadequacies. In a word the clear infrared image can be acquired after correction and enhancement processing.According to market demand for little mobile and low-price IRIS, an embedded image processing system is developed. The embedded system includes image signal pretreatment circuit, high-speed A/D sampling circuit, programmable logic circuit, high performance signal processing circuit and image displaying device. The embedded system has the characteristics from universal image processing platform because of the following three reasons: Firstly high speed A/D converter is used in the system and it is fit for the different wavelength solid image sensors. Secondly the pretreatment circuits are designed for general circuit. Thirdly the flexible programmable logic device is able to control different pixel sensor sampling. Now the embedded system has been successfully developed and clear image has been obtained. Two creative researches are achieved. Firstly creatively put forward data joint transferring way by DMA. Not only does it increase data transfer... |