ANN-based visibility prediction for camera placement in vision metrology

Simulated based network design is a powerful technique in vision metrology systems (VMS) to automate camera placement process. Relatively expensive and time-consuming process of building 3D simulated CAD model is the main drawback of this technique which causes to do not be practically attended by users. In this paper, a new visibility prediction analysis using artificial neural network (ANN) is proposed. This flexible and reliable method can be installed in intelligent cameras to predict the visibility of weak points during imaging in order to suggest the best position and situation of next camera station. Our primary tests on a complex network demonstrated high capability of the proposed method.