The factors affecting the image quality of refrigerating medium wave infrared (MWIR) focal plane detectors include non-uniformity of response, drift characteristics of response and blind pixel. So signal processing in real time such as non-uniformity correction, drift compensation and blind pixel substitution is required in infrared focal plane imaging system. At first, a refrigerating MWIR image system with high frame frequency, low noise and 640×512 pixels is designed. And the dynamic range is calibrated to implement the NETD≤ 30 mK in full dynamic range. And then, according to the response characteristics of the pixels in the focal plane, the non-uniformity and blind pixel correction method suitable for the infrared imaging system are researched. A non-uniformity correction, blind pixel detection and substitution algorithm based on radiation calibration and scene fusion is put forward. At last, the infrared image data is collected in field experiment. Experimental results show that the method has a good image correction effect, easy to be implemented, and it has strong environmental adaptability.