Monogenic Filtering Based Automatic Defect Detection from a Single Fringe Pattern

We propose a fringe analysis algorithm for the automatic detection of defects from a single fringe pattern (FP). Typically, the surface defects exhibit high fringe density areas in the FP. Consequently, high fringe density regions can be utilized as a signature for detecting and locating the surface defects. An algorithm based on monogenic filtering of the FP is proposed for an efficient computation of the fringe density. The defect related high density fringe areas are segmented from the defect-free region based on a threshold derived from the fringe density histogram. The algorithm is found to be noise robust and it does not require any pre-processing of the FP. A single FP based analysis approach is suitable in identifying defects using interferometric systems in an industrial environment. Simulation and experimental results are provided to demonstrate the feasibility of the proposed algorithm.