The invention discloses an automobile tyre pattern recognition method based on machine vision, and the method specifically comprises the following steps: S1, obtaining an image of linear laser on a tyre pattern, and correcting the image; S2, removing the background and vignette of the corrected image; S3, solving a plurality of central coordinate points of a light bar on the processed image at step S2, carrying out the cubic spline fitting of the central coordinate points, and obtaining a cubic spline fitting formula; S4, carrying out the fitting of the central coordinate points of the light bar based on a polynomial method, and obtaining a polynomial fitting formula; S5, solving the difference between the cubic spline fitting formula and the polynomial fitting formula; S6, carrying out the convolution of the difference result at step S5, and solving the maximum values of a convolution curve; S7, filtering the maximum value less than a preset value, wherein the obtained maximum value is the depth value of the tyre pattern. The method can achieve the accurate and efficient recognition of the tyre pattern and the measurement of the depth of the tyre pattern, and is very high in precision and robustness.