Correction method for stripe nonuniformity in infrared images based on Savitzky-Golay weighted fitting algorithm
-
摘要: 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%。
-
关键词:
- 非均匀性条带校正 /
- 红外成像探测器 /
- 直方图加权 /
- Savitzky-Golay滤波器拟合
Abstract: 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. -
图 1 不同拟合方法处理图像列均值数据
Figure 1. Comparisons of mean value fitting results by different methods
图 2 不同拟合方法处理图像列方差数据
Figure 2. Comparisons of variance fitting results by different methods
图 5 不同方法处理图像后图像列均值分布
Figure 5. Mean value comparisons of the original image and corrected images
表 1 不同处理方法客观评价指标数据
Table 1. Comparisons of image evaluation criteria by different algorithms
方法 PSNR IQ UIQI ρ 带非均匀性条带图片 25.307 3 0 0.881 9 67.159 1 矩匹配法 25.856 5 1.820 9 0.902 8 71.268 6 傅里叶低通滤波法 25.362 5 2.392 9 0.912 4 47.561 2 直方图匹配法 26.792 6 2.207 3 0.962 9 65.268 1 本文方法 29.570 5 4.809 7 0.977 3 54.295 1 
下载: 导出CSV
-
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] -
下载:
图(7) / 表(1)
计量
- 文章访问数: 1912
- HTML全文浏览量: 401
- PDF下载量: 864
- 被引次数: 0
