基于曲面拟合的红外图像盲元检测方法

Detection Method for Infrared-image Blind Pixels Based on Curved-surface Fitting

针对传统红外图像盲元检测方法采用固定检测阈值和不能兼顾全局非均匀性的问题,提出了一种基于曲面拟合的盲元检测方法。首先分析了有效像元和盲元在灰度分布三维图上的显著差别,进而提出整幅本底图像的曲面拟合方法,然后用拟合后的曲面值作为比较基准运用3σ原则进行盲元检测。使用拟合后的曲面值替代固定基准能够消除探测器响应非均匀性对盲元检测的影响,提高盲元检测的准确性和有效性。进一步开展了验证实验和比较实验,实验结果表明该方法比直接运用3σ原则法检测精度更高、更有效。

Aiming at solving the problem that the traditional blind-pixel detection methods for infrared images use the fixed detection threshold and could not give consideration to global non-uniformity, a blind-pixel detection method based on curved-surface fitting is proposed. Firstly, the analyzation of the difference between effective pixels and blind pixels was conducted in the 3D distribution of grayscale. Then the curved-surface fitting method based on the whole background image was put forward, and the fitted surface values could be used afterward as the benchmark to detect blind pixels with the 3σ principle. Using fitting values instead of a fixed benchmark can eliminate the influence that the non-uniformity noise of the detector brings to the blind-pixel detection, and improve the accuracy and effectiveness of blind-pixel detection. Then the verified experiment and comparison experiment were carried out, and the experimental result shows that the proposed method is more effective and precise as compared to directly using the 3σ principle.