In order to address the problems of online machining insitu product quality evaluation, an online detecting method taking digital image as the monitoring vehicle is put forward. 16 bits image is acquired by photographing for the metal milling surface. The brightness equation regularization problem considering the continuous restriction of the machined surface is obtained to recover the three dimension surface. The dual tree complex wavelet transform is also applied to denoise the reconstruction surfaces. A face milling case of aerial aluminum alloy 7075 is investigated to verify the effectiveness of the proposed methodology. The result of the proposed method is compared with that of offline roughness measuring instrument. It is demonstrated that the roughness indicators derived from the proposed method are consistent with those of the offline instrument. The comparison results indicates that the noncontact method is effective for the online metal milling condition monitoring and machining quality assessment.