The application of machine vision technology in the software design of parameter measurement system requires image processing. In order to narrow the gap between the parameter measurement results and the real parameter results, the manufacturing error is taken into account in the image processing process, so the Hough transform method is selected to process the image. Hough transform uses the duality of the points and lines in the image to recognize the special geometry of the cylindrical helical gear from the overall image, and synthesizes the points in the parameter space from the given curves in the original image space. The biggest advantage of this method is that it can transform the search and detection of non parallel and asymmetric lines and points in the original image into the problem of finding peaks in the parameter space. The Hough transformation process is shown in the figure.
In figure a, a point (Xi, Yi) is randomly taken in the XY plane, and the straight line L passing through the point can be expressed as:
When a and B change, the shape and position of the straight line l also change. If the linear expression is changed to the form of Fig. 2B, the corresponding linear equation with (Xi, Yi) as the variable can be obtained in the parameter space according to the values of a and B. In Fig. b, the intersection coordinates of the straight line passing through the point (XJ, YJ) and the straight line L are (a ‘, B’), and a ‘and B’ are the slope and intercept of the straight line equation respectively. The above transformation can map the points on the image plane to the parameter plane and convert them into lines. Finally, the statistical principle is used to compensate for the errors.
Assuming that there are two straight lines in the image plane acquired by the system, then they are finally displayed as two peak points in the parameter space. When the parameter coordinate system is derived, one image plane straight line can correspond to one of the peak points. In practice, the straight line Yi = Xia + B cannot represent a horizontal line parallel to the x-axis. In order to solve this problem, the parametric equation is converted into the linear equation described by the above formula, which can correspond to the nodes in the image plane and the corresponding curves in the parameter plane. Any line and ellipse in the graph space or a point on the ellipse are associated with the specific graph in the parameter space, and in the graph space, these nodes are associated with a certain equation. Therefore, all the graphs obtained after projection include these points in the parameter space, and these points appear on the same point in the parameter space in the form of aggregation. In the actual application of parameter measurement system, Hough transform method can detect the fixed target in the known cylindrical helical gear pattern, and its performance is relatively stable without external interference. For some minor manufacturing errors, defects or pollution, the image can be effectively compensated and accurate identification can be realized.