Linear marks similarity matching algorithm based on laser detection signals

It´s usually not easy to compare linear toolmarks quickly and quantitatively with the current image processing and three-dimensional scanning methods. In order to deal with these kinds of problems, a similarity matching algorithm for laser detection signals of linear marks is proposed. In this method, locally weighted scatterplot smoothing algorithm is used to performs noise reduction on the surface signals of linear marks that are picked up by a laser displacement sensor; followed by the identification of the features of `thick and consistent´ trends in the signal data. Then the feature vectors are quantized by cosine vector curve fitting to individually calculate the spatial distance of the sample. Results from similarity comparison tests of actual cutting marks validate the accuracy and validity of proposed method.