摘要

The linear Radon transformation is applied to extract inflection points for online test system under the noise conditions. The linear Radon transformation has a strong ability of anti-noise and anti-interference by fitting the online test curve in several parts, which makes it easy to handle consecutive inflection points. We applied the linear Radon transformation to the separation test system to solve the separating clearance of automatic brake adjuster. The experimental results show that the feature point extraction error of the gradient maximum optimal method is approximately equal to +/- 0.100, while the feature point extraction error of linear Radon transformation method can reach to +/- 0.010, which has a lower error than the former one. In addition, the linear Radon transformation is robust.