To realize integration detection and construct stability maintaining benchmark for large apertures of segmented telescopes, we propose a benchmark construction method. In this study, we use local pupil projection to perform pupil alignment mapping. In addition, we construct a system confocal spatial benchmark using a microlens array. On the basis of annular whole-body control mode, a joint analysis method of confocal and curvature radius enables joint adjustment of the curvature radius and system alignment. Finally, the stripe envelope formed by white light interference is used for coarse common phases detection, and the channel spectral method is used to obtain precise connection between coarse and fine common phases. Additionally, the spatial confocal reference positioning exhibits an accuracy of less than 125 μm, and the common phase reference has a coverage range better than 0.5 μm within a 20-μm-range. Furthermore, the uncertainty of the spectral reference is less than 5%. We have effectively improved the accuracy of optical system in-situ measurement by achieving hierarchical and multimodal suppression of disturbances from different spatiotemporal features. We have shortened the length of the traceability chain and increased the efficiency and accuracy of detection by utilizing the new method of common reference in-situ measurement.