Dot-array structure light based photogrammetry of flexible thin-film surface morphology using machine vision

Aiming at the morphology measurement problem of flexible thin-film surface, a visual measurement method is proposed, which uses dot array structure light as labels to mark the measured surface to achieve the photogrammetry measurement of the flexible thin-film surface morphology without obvious visual characteristics. The propose method can quickly finish photogrammetry task and does not require expensive equipment and complex operation. In the imaging process, the marked dots have small imaging error. Especially, the highly accurate dot center can be obtained with elliptic fitting method. A dot array can mark hundreds of mark points on the target surface at a time, the fast one shoot imaging is achieved without scanning process and the measurement efficiency is high. Aiming at the feature extraction and matching issues in the measurement process, the image target region extraction method based on area criterion and the identification and matching algorithm depending on array order constraint are proposed, which have small calculation burden and fast execution speed. Using the proposed method, a 1000 mm×1200 mm thin-film surface was measured and reconstructed; the experiment results show that the surface morphology error PV value is less than 0.668 mm and the RMS value is less than 0.33 mm within 2200 mm working distance.