为解决FY-3C/MERSI反射波段非线性响应导致传统两点定标不适用问题，提出了在轨宽动态综合辐射定标方法。融合多种不依赖于地面同步测量的替代方案，利用各方案定标样本的多等级反射率特性，实现遥感器宽动态范围的辐射定标。每种方案的定标样本采用分段平均方法进行等权重融合，最终定标系数基于融合后样本通过加权拟合回归获得。非线性特征采用发射前室外定标实验分析的二次项系数进行修正，在轨计算定标斜率和截距。分析和验证结果表明：各替代方案样本分布具有很好一致性，融合样本相关性可达0.99；综合定标结果在反射率大于10%的中高端目标回归残差小于1%，高端定标结果与基于深对流云目标的交叉定标(DCC)检验样本的相对差异小于1.5%，敦煌同步观测实验检验结果表明中端精度优于3%；由于MERSI对低亮度辐射响应能力显著降低，发射前室外定标实验获得的非线性系数不能适用，定标结果在低目标反射率出现高估现象。该成果可应用于所有FY-3C/MERSI反射通道，明显降低两点法定标的不确定性以及非线性的影响。

Because of the nonlinear response of FY-3C MERSI reflective bands , traditional calibration based on two-points is inapplicable. An integrated method for on-obit wide dynamic vicarious calibration is developed to resolve this problem. By combining various vicarious calibration methods which are independent on the ground synchronous measurement, and using the multi-scale reflectance characteristics of every program’s calibration sample, the integrated method can achieve wide dynamic radiometric calibration. Calibration samples from each program are integrated equally and weightedly by subsection-average method, and calibration coefficients are calculated based on integrated samples using weighted fitting regression. The prelaunch quadratic term of calibration experiment analysis is used to perform nonlinear correction, calibration slope and intercept are evaluated on-orbit. Anlysis and verification results show that sample distribution of each vicarious program shows well consistency and the correlation of integrated samples can reach 0.99. Regression residual of comprehensive calibration results is less than 1% for the medium to high dynamic with reflectance greater than 10%. The relative difference between the high-end calibration results with cross calivration based on deep convective clouds (DCC) test sample is less than 1.5%and Dunhuang simultaneous observation test results show that the medium accuracy is better than 3%. Because of the responsiveness to low-light radiance of MERSI shows large degradation, nonlinear coefficients acquired form prelaunch outside calibration experiment are inapplicable, and calibration results show over estimated at the low target reflectance. The results can be applied to all reflection channels of FY-3C MERSI, which can reduce the uncertainty of the two-points calibration and the influence of nonlinearity.