The significance of this step is to determine the sign of R and B. First, the connected domain with too small area in the canny edge image of the ROI image is removed, and then the edge image is traversed to find out the number of structural element 1 and structural element 2 respectively. According to the number of structural elements, it is determined whether the helical gear is left-handed or right-handed. Some of the structural elements used by the author can be expressed as: to measure the inclination angle of the workpiece to match the image coordinate system with the world coordinate system, use the shaft to fix the helical gear, and obtain a more accurate inclination angle through the image of the shaft. The ROI image of the shaft is extracted from the image. After edge extraction, the analytical formula of the inclination angle of the shaft and the center line of the helical gear is obtained by least square fitting. Combined with the coordinates of the ROI area, the position of the origin of the image coordinate system is obtained.
Where: structural element 1 – + 45 ° slope structural element; Structural element 2 – 45 ° slope structural element; Structural element 5 – horizontal structural element. For left-hand helical gears, the slope of the side image texture is mostly negative, while for right-hand helical gears, it is opposite. This algorithm is greatly affected by noise, so it needs to process the image Gaussian filter in advance. The algorithm is implemented in 781 × The running time on the ROI image of 1558 was 27.3 Ms.
In several helical gear edge diagrams, the number of structural elements is shown in the table.
| Structural element code | 45 ° 1.5 m26z left handed | 45 ° 1.5 m 26z right handed (rusty) | 19 ° 1.5 m 22z left rotation (overexposed) | Spur gear with inclination of 5 ° |
| 2 | 18 088 | 11 946 | 2 637 | 872 |
| 1 | 4 941 | 28 861 | 1 569 | 917 |
| Structural element code | 19 ° 2 m30z left handed | 19 ° 1.5 m30z left rotation | 19 ° 1.25 m 25Z left rotation (dark light) | Spur gear with inclination of – 1 ° |
| 2 | 3 988 | 1 004 | 1 809 | 3 242 |
| 1 | 2 270 | 443 | 1 161 | 934 |
It can be seen from the table that the number of the two structural elements of helical gears with different rotation directions is very different, and the accuracy of the algorithm will not be affected under extreme conditions such as 19 ° to 45 ° helix angle, rust and low illumination. Because spur gears are affected by irregular specular reflection, this method cannot be used to determine whether they are spur gears.
The above experimental data prove that this algorithm is feasible.