基于定标和图像处理的数字相机串扰图像修正方法

Method for Crosstalk Image Correction of Digital Camera Based on Calibration and Image Processing

针对多通道数字相机因信号串扰导致的输出图像重叠问题,基于自行研制的四通道CCD数字相机系统数据,对串扰系数进行了追踪测试,发现串扰系数与输入信号强度和系统读出速度有特定的规律。对于这类系统噪声类串扰,提出了一种基于定标和图像处理的串扰图像修正方法。该方法通过定标建立串扰比例系数与像素数字编码(Digital Number,DN)的关系,再由串扰通道反推被串扰通道的干扰量,通过还原算法进行系统差修正,最终消除串扰。测试结果表明,经过该方法修正后的图像,串扰程度可降至基底白噪声量级,图像质量得到大幅改善。该方法对成像系统中因硬件工艺缺陷造成的固定图案噪声的消除有显著效果,对系统噪声类干扰也有抑制作用,在实际应用中可以作为一般成像系统优化图像质量的补充。

For the problem of output image overlap caused by signal crosstalk in multi-channel digital cameras,we tracked the crosstalk coefficient using the data from the self-developed four-channel CCD digital camera system,and found a strong correlation between the crosstalk coefficient,the input signal intensity,and the system readout speed.For such system noise crosstalk,we propose a crosstalk image correction method based on calibration and image processing.This method establishes the relationship between the crosstalk proportion coefficient and Digital Number,which can be used to infer the amount of interference in the crosstalk channel and realize the crosstalk image recovery algorithm and the systematic error correction.The test results show that after the image restoration correction,the crosstalk degree is reduced to the level of base white noise,and the image quality is greatly improved.This method has a significant effect on the fixed pattern noise caused by the hardware process defects in the imaging system,and also suppresses the noise-type interference of the system,which can be used as a supplement to the image quality optimization of the general image system in practice.