宽带多光谱成像系统CCD响应的线性化校正方法

Linear Correction Method of CCD Response for Wide-Band Multi-Spectral Imaging Systems

对多光谱成像系统CCD响应特性进行了分析,并提出使用物理阶调模型作为新的线性化校正方法。在分析了数码相机CCD响应值的线性化校正原理和方法的基础上,通过校正精度和光谱重构效果两个方面,将其与传统的多项式模型和指数模型等方法进行了比较分析。实验结果表明,经过线性化校正的各通道的线性相关性进一步增强,其中GOG模型各通道的相关系数值均接近0.996,具有最好的线性校正效果;且GOG模型校正过的RMSE和ΔE值分别降低了37%和61%。因此,采用GOG模型,可以较好地实现CCD响应值的线性化校正,提高多光谱图像获取中光照不均匀性的校正精度和效率。

The CCD response characteristic of the multi-spectral imaging system was analyzed, and the physical model was proposed as a new method of linearization. Based on the analysis of the linear correction principle and method of the digital camera CCD response value, it was compared with the traditional model such as polynomial model and the exponential model in two aspects of the correction accuracy and spectral reconstruction effect. Experimental results show that the linear correlation of each channel after linear correction was further enhanced, and in the GOG model the channel correlation coefficient values was close to 0.996 with the optimal linear correction effect; The RMSE and ΔE values which were corrected by the GOG model were reduced by 37% and 61% respectively. Therefore, the GOG model could be used to achieve the linear correction of CCD response value and improve the accuracy and efficiency of illumination nonuniformity correction in the process of multispectral image data acquisition.