An algorithm for detecting tool damages in end milling is presented. T
he surface texture is recorded by taking gray level images with a vide
o camera. These images are transformed to the Fourier domain where a f
eature sensitive to cutting edge breakage is extracted. A parametric s
ignal model for the groove texture is given. It models the essential s
tatistical and geometrical texture properties and leads to an analytic
al expression of the power spectral density of the texture generating
stochastic process. Using this, the spectral feature for tool damage d
etection can be defined specifically. The suitability of the method is
demonstrated with experimental results.