Research On Gear Detection And Measurement System Based On Machine Vision

Gear products play an important role in many fields,so it is necessary to detect the products strictly to ensure the quality.Manual detection has many disadvantages,such as more errors,slow speed and poor real-time data storage,so manual detection is not suitable for the real-time online detection in production process.In the HDevelop development environment of machine vision software HALCON,this paper designs a kind of gear automatic detection and measurement system.The system includes image acquiring,image processing,image recognizing and so on.The main contents are as follows:1.Placing the gear which is to be detected on the double tracks annular conveyor belt.When the gear is moved to the position of the photoelectric sensor in the black box,its image is acquired and sent into the computer for processing.2.The machine vision software HALCON is used to perform the preprocessing of the image that to be detected.Firstly,convert the color gear image into gray image by image transformation.Then,the anisotropic diffusion filtering is selected to deal with the gray images.The process has the advantage of removing the image noise while preserving the edge of the image.3.In the process of gear detection,the black box in the machine vision system can reduce the impact of the external environment on the system,so the fastest threshold segmentation algorithm is chosen for smoothing the image.Then use morphological processing and subtraction processing to obtain the number of gear teeth and the area of a single gear.And then exclude substandard products.4.After selecting the region of interest,Canny and bilinear interpolation algorithm is used to detect the sub-pixel edge of the gear.In the process of recognizing different types of gear,a method of using template matching and image pyramid search is proposed.The shape of the center hole of the sub pixel is used as the matching template.And the template supports anisotropic scaling.After the template matching,in order to display the matching results,the matching image is processed by an affine transformation.Experiments show that the method can identify different types of gears quickly and accurately.5.After obtaining the sub-pixel edge,the Green theorem is used to obtain the area and the center of the gear.Then use the least squares method based on Tukey to fit the circular curve,and get the radius of each circle.Next,use one-dimensional arc measurement to get the gear parameters,such as tooth thickness,tooth width and tooth pitch.Finally,we can complete the measurement work with the system calibration.