We propose an optical measurement technique devoted to on-line inspection of large-scale and thin-wall objects. The proposed approach is based on a measurement network that consists of three-node optical sensors. With this specifically designed measurement network system, we can achieve the inspection tasks such as multi-view range-data acquisition, registration, and integration while it is also possible to make a comparison of reconstructed model with the computer-aided-design (CAD) model to precisely determine the error distribution on each external and internal facet walls. An on-line calibration scheme for multiple-node sensors based on phase-mapping combined with photogrammetric technique is also presented and verified by experiments. Furthermore, the registration of multi-view range-data can be achieved using the calibrated parameters and iterative closest points (ICP) algorithm. Moreover, we also employ ICP algorithm for the alignment of reconstructed model with measured data with the nominal CAD model, and then the dimension computation and error comparison can be easily acquired. Both the theoretical analysis and the experimental results show the effectiveness of the proposed approach.