Structural Feature Extraction from Regular-Shaped Rigid Objects(1) - Based on Gradient Direction Constraint Randomized Hough Transform

摘要

The structural feature extraction from regular-shaped rigid objects is studied. First, some specific measures for facilitating the conventional Randomized Hough Transform (abbr. RHT) in our application are introduced. Then, a gradient direction constraint RHT is proposed for improving the poor performance of the conventional RHT that too much false response happens in noisy real images. The basic idea of the new method is that the edge points of the true straight edges have more consistent distribution of the gradient direction than the noise has. Last, experiments including subjective observations and quantitative statistics validate that the new approach has much more robust performance than the conventional RHT.