针对红外成像探测器单元响应不一致等因素导致的图像非均匀性条带问题, 提出了基于直方图加权和Savitzky-Golay拟合的非均匀性条带校正方法。首先, 计算每列的归一化直方图函数并将其作为权值对图像进行加权运算;然后, 利用Savitzky-Golay滤波器对加权后的图像列均值和列方差进行拟合, 并将结果带入校正公式, 通过可调参数的迭代完成校正。实验结果表明: 该方法能有效地去除非均匀性条带, 保留图像的光谱辐射信息和纹理细节, 各项评价指标均提高10%。

With factors such as the nonuniform response of infrared detector units, stripe nonuniformity in images is observed. To solve this problem, a novel correction method to use histogram weighting and Savitzky-Golay filter fitting(HWSG) together is developed. Firstly, the normalized histogram functions of each column are calculated and taken as weights of the stripe image. Secondly, the Savitzky-Golay filter is used to fit the weighted column mean values and variances. The fitting result is put into the correction formula, and then the corrected image is obtained after adjustable parameter iterations. The experimental results demonstrate that the proposed approach is efficient and retains the image spectrum and texture information excellently. A 10% improvement of each evaluation index has been achieved.