Fractal generated turbulence is highly anisotropy and inhomogeneous, particularly, within its unique turbulence production regime. This study aims to develop an experimental-based detection method which allows the details of fluid-structure-interaction (FSI) via grid-net interplays to better express the strength of fractal grid induced flow dynamics. Hence, a novel approach is initiated to perform nets photogrammetry for 3D-reconstruction through an innovative means that permits the historically difficult task of reconstructing nets to be achieved. A keypoint extraction algorithm designed specifically for nets dubbed Neighborhood OCCupancy (NOCC), and a complementary stereo matching algorithm named Propagative Stereo Matching (PSM) are presented. The former identifies the center of intersections in the net by determining the percentage of an ellipse filled at each pixel. PSM works on the NOCC keypoints via matching pairs close to a known match and using the newly matched pairs as reference for subsequent matches in a way reminiscence to how humans would perform the stereo matching task manually. The effectiveness, accuracy and robustness of the detection are carefully validated using three different net geometries, which are individually subjected to three synthetic configurations. Consequently, the system repeatability, utilisation, stereo matching accuracy and reconstruction accuracy, of which, with the mean of all cases outstandingly deduced to 99.0%, 87.0%, 100% and 0.12mm, respectively. The presented method is then employed to elucidate the windward flow straightener influence on grid-net fluid flow interplays. Overall, the current framework may be an enabling factor in providing critical FSI-features to realize the mechanical strength of fractal generated turbulence.