Three-dimensional Hessian-based laser stripe centerline extraction method using spatiotemporal volume construction

This is the authors' abstract. We don't add key points for this paper.

Precise centerline detection of the projected laser stripe is crucial for the overall accuracy of a linear laser measurement system. Yet real-world factors such as stray light interference and reflectance non-uniformity often disrupt conventional methods. To overcome these limitations, we introduce a spatiotemporal three-dimensional Hessian-based approach. Our method "lifts" the detection problem into a three-dimensional space-time volume by stacking a short sequence of consecutive stripe images along the temporal axis. By incorporating the temporal dimension, the proposed algorithm effectively enhances spatiotemporal coherence: the extraction accuracy of genuine stripes is improved, while transient noise and flicker varying over time are suppressed. Experimental results further verify the effectiveness and reliability of the proposed method.