Nonuniformity Correction Method of Infrared Detector Based on Statistical Properties

The images captured by a plane-array type infrared detector, formed by integrating a large-scale probe array and readout circuits via interconnection, is distorted by two types of superimposed nonuniform noises caused by the probe and readout circuitry. These nonuniform noises arise from the detector's constituent materials, fabrication processes, and other defects. In addition, the output response of an infrared detector is limited by the dynamic range of the circuit, the low and high ends of the output signal exhibit strong nonlinearities. To overcome the impact of detector response nonlinearity and achieve simultaneous correction for two different types of nonuniformities in infrared images, in this study, we proposed a novel correction method. First, the pixel points of all images are reordered according to their pixel values to provide a uniform reorganisation data. Then, the linear correlated region based on the recombined data is extracted using a judgment algorithm that processes the linear correlations of multiple regions to simulate the detector's response curve and find the nonuniformity correction coefficient of each pixel. Experiments on simulated noisy and real images show that our method accurately eliminates nonuniform IR noise without causing blurring and provides optimal peak signal-to-noise ratio and structural similarity.