The 3D Digital Image Correlation (DIC) method [[1], [2], [3], [4], [5], [6], [7], [8]] can perform full-field static and dynamic deformation measurements of structures and is widely applied in fields such as civil engineering, transportation, and aerospace. However, this method has limitations in certain scenarios. For example, when the target structure is large or has a high curvature, the spatial resolution of a single DIC system with only two cameras may be insufficient. To overcome these challenges, researchers have explored multi-camera DIC systems (using three or more cameras) [[9], [10], [11], [12], [13], [14]] to improve the spatial resolution for both static and dynamic deformation measurements.

Compared with the two-camera Digital Image Correlation (DIC) method, the main challenge in multi-camera DIC is to unify the coordinate systems among different subsystems. Currently, there are mainly two types of solutions. The first approach utilizes the 3D information of calibration plates or feature points in the overlapping regions to calculate the extrinsic parameter matrix between the main cameras of two subsystems, thus unifying the measurement data of multiple subsystems. Malesa et al. [15] calibrated the relative positions and poses of multiple cameras using a checkerboard calibration plate, achieving the stitching of measurement data of a multi-camera system and successfully expanding the measurement field of view. Shao et al. [16] proposed a multi-camera calibration method based on speckles, which reduces the reliance on external calibration plates and effectively solves the problem that traditional planar calibration methods require large-sized calibration plates for large fields of view. Poozesh et al. [17] calculated the coordinate transformation matrix through marker points in the overlapping region, realizing a dynamic spatial data stitching method based on a multi-camera system and successfully applying this method to the dynamic deformation measurement of wind turbine blades. Ge et al. [18] proposed a calibration method for a ring multi-camera system based on a transparent calibration plate, significantly improving the calibration accuracy and measurement capabilities of the ring multi-camera system and enabling panoramic dynamic and static measurements.

The other approach is to achieve full-field extrinsic parameter calibration of multi-cameras through some external auxiliary means or devices. Tribou et al. [19] proposed a new parameterization method to solve the multi-camera calibration problem for non-overlapping fields of view, effectively avoiding the limitations of requiring overlapping fields of view or known markers in traditional methods. Malowany et al. [20] significantly improved the extrinsic parameter calibration accuracy of multi-cameras by combining the data of stereo DIC systems and laser trackers. Bartsch et al. [21] proposed a multi-camera system calibration method based on visual ray calibration, avoiding the complexity of separately calibrating each camera and calculating its relative position and pose in traditional methods. Wei et al. [22] proposed a multi-camera calibration method based on speckle images and close-range photogrammetry technology, which improves the calibration efficiency of multi-cameras for large fields of view.

In addition, Koo et al. [23] proposed a two-step optimization method for the extrinsic parameter calibration of multi-camera systems, addressing the potential instability issues in existing methods and improving the calibration robustness and accuracy. Rameau et al. [24] proposed a multi-camera calibration toolbox named MC-Calib, making important contributions to solving the calibration problems of complex synchronous multi-camera systems, especially in terms of versatility, robustness, and automation. He et al. [25] proposed a speckle-based compensation method, effectively reducing the systematic errors caused by camera motion, especially in non-laboratory environments.

Although existing multi-camera DIC systems have made significant progress in expanding the measurement field of view and improving environmental adaptability, they still face some challenges in practical applications. For example, traditional calibration methods often rely on large overlapping regions or external auxiliary devices, which not only increases the experimental complexity but also may lead to a waste of camera resolution. The existing non-overlapping field-of-view multi-camera systems can ensure a sufficient measurement field of view but suffer from poor measurement accuracy or complex calibration processes. Moreover, there are relatively few research achievements on multi-camera DIC technology for hollow structures or targets with complex geometries. To address these problems, this paper proposes a multi-camera DIC method based on multi-pose speckle compensation calibration, aiming to reduce the reliance on overlapping regions and improve the adaptability and accuracy of the system in measuring complex structures. The main research objectives of this paper are as follows:1)

Develop a speckle compensation calibration method to minimize overlap requirements and enhance accuracy for hollow structures.

2)

Validate the method's robustness in static/dynamic measurements and demonstrate its superiority in spatial resolution and practicality for industrial applications.