Camera calibration is an essential part of 3D sensing technology based on structured light, which links the intrinsic and extrinsic parameters of the measurement system with the 3D coordinates. The camera calibration approach based on feature points retrieved from gray information can be easily affected by image noise, contrast, and other factors. Thus, this study proposes an active camera calibration approach based on absolute phase target (APT), in which absolute phase is determined using the temporal phase unwrapping algorithm. The point where phase value is the integer multiples of 4π is selected as the feature point. The local window least-square linear-fitting approach is used to determine the precise subpixel coordinates of every feature point, then the one-to-one correspondence between the image and world coordinates of feature points is established. In simulation experiments, according to the absolute error change of the camera parameters under different degrees of Gaussian noise and blur, we concluded that the proposed APT calibration method is more robust to noise and blur than the traditional checkerboard calibration and dot calibration approach, and calibration accuracy is relatively high. The validity and feasibility of the proposed APT approach are further demonstrated by the real comparison experiment, which reveal that the calibration accuracy of the APT approach is higher and the reprojection error can be reduced by 58.68% under out-of-focus conditions (the target is outside the working distance of the camera).