A Robust Feature Detection Method for an Infrared Single-Shot Structured Light System
Chu Shi, Jianyang Feng, Suming Tang, Zhan Ping Song · 2018
This paper introduces a robust feature detection method based on an infrared structured light system (SLS). Subject to the absorption of infrared light by human face, the projected grid-pattern is difficult to be extracted directly. In the proposed image preprocessing stage, two nonlinear filters are firstly applied and their subtraction is calculated. Then, a local adaptive threshold filtering procedure is implemented to boost the contrast around grid-lines and to suppress the image background. Based on the enhanced structured light image, a cross filter is applied to extract candidate grid-points. Finally, a symmetry-based feature detector is introduced to localize the grid-points precisely. Real human face is used in the experiments. And results show that, most of the grid-points can be accurately detected and 3D model of human face can be successfully retrieved via single projection of an infrared grid-line structured light pattern.