Minimum resolvable temperature difference objective testing method of thermal imager

The invention provides a minimum resolvable temperature difference (MRTD) testing method of a thermal imager. The method is a thermal imager's MRTD objective testing method based on image processing and adaptive neuro-fuzzy inference system technology. By the method, testing speed is raised, and subjective factor interference of traditional testing of a thermal imager's MRTD parameter is minimized. According to the method for real-time objective testing of the thermal imager's MRTD parameter, an adaptive neuro-fuzzy inference system generates a corresponding classifier model according to learning sample training, and the classifier model predicts the thermal imager's MRTD parameter. In comparison with a traditional method, the method brought forward in the invention can be adopted to rapidly and effectively test the MRTD parameter of a thermal infrared imager. Meanwhile, manual repeated operations and subjective factor interference in a traditional testing method are avoided. Thus, MRTD parameter of the thermal infrared imager can be tested rapidly and effectively, and mutability of test results is reduced.