Because of the impact of ambient illum ination,there may exist large deviation in calculating micro-image definitio n according to the traditional definition estimation methods,when the microscope lens group is used to capture micro-image to the same target.The impact may lead to that the assessed micro-image′s definitions fluctuate fiercely.In order to eliminate the effect of ambient illumination,an i llumination separated model and a target object location model are proposed based on the dark channel prior.Furth ermore,an illumination separated model is adopted and a new micro-image definition assessment method is given,wh ich is robust to ambient illumination.Transmission attenuation of the light is first obtained by computi ng dark channel of micro-image, target areas of micro-image are located and the illumination separated image is obtained by analyzing the transmission attenuation,and finally the micro-image definition is more effici ently assessed.Experiment is performed to verify the efficiency of the proposed method.The experimental results show that the proposed method can efficiently suppress the impact of ambient illumination in the proces s of computing micro-image definition,and as a result,the maximal normalized variance is suppressed with 85% with the proposed method. Moreover,the proposed method can be used in the in-focus auto-detecting of mi croscopy lens group.