利用CCD中的部分坏点作为温度敏感器，提取CCD温度估计值，建立温度估计值与暗电流的函数模型，并结合均值滤波的方法，在真实场景图像中提取CCD温度估计值，实现了面阵CCD的暗电流估计。在方法分析的基础上，以真实的暗电流数据为基准参考，对暗电流的估计结果进行了比对实验。实验结果表明，在不同的积分时间及大动态范围的成像条件下暗电流的估计结果十分准确，偏差小于0.4%，并且具有一定的抗噪性。该方法利用场景本身的特征信息对拍摄时刻的暗电流进行估计，不需要额外采集暗电流数据，节省了图像采集时间，十分适合于积分时间较长的高光谱成像或天文观测领域，采用无温度控制的低成本CCD成像探测系统进行实时图像采集。

Bad pixels in the area CCD detector are used as temperature sensor, to estimate the CCD temperature, and to establish the function model between the CCD temperature and the dark current. In order to make the result insensitive to noise; mean filtering method is used to estimate the CCD temperature in the image of the real scene. After extraction of the CCD temperature of the real scene, the dark current of area can be estimated. Based on analysis of the method, using the real dark current as standard, the result of the dark current estimation has been validated. The experiment shows that the estimation of the dark current is very exact in both different integration time and large dynamic range conditions with the rate of error under 0.4%. Furthermore, it is not sensitive to the noise within the images. The method uses the character of the image to estimate dark current, so it is not necessary to collect the dark-current data, and the image acquisition time can be saved. Therefore, this method is very suitable to the low-cost CCD imaging system without temperature control for real-time image acquisition in hyper-spectral imaging or astronomical observation fields.