Automatic sorting and boxing equipment of component assembly line based on vision

[0019] The present invention will be further described below in conjunction with the drawings and embodiments.

[0020] This embodiment discloses a vision-based automatic sorting and packing equipment for parts assembly line, which is a non-contact image processing method based on vision, such as figure 1 As shown, it includes randomly placed parts on the conveyor belt 2, the conveyor belt 2, the proximity switch 3, the area scan camera 4, the ring LED light source 5, the industrial computer for image processing 6, the robot controller 7, and six axis The robot arm 8, the suction cup 9 and the material box 10 used for packaging different models, the area scan camera 4 and the LED light source 5 are installed above the conveyor belt 2, and the proximity switch 3 is located within the camera's field of view.

[0021] Such as figure 2 As shown, the specific operation process is: First, when a new type of part 1 is added to the conveyor belt 2, the part 1 needs to be placed offline in the camera's field of view, and the area scan camera 4 is triggered to collect zero by controlling the LED light source 5 to light up. The image of component 1 is processed by the industrial computer 6 to extract the edge contour coordinates, and its polar coordinate vector and dimensionality reduction feature vector are calculated, and the polar coordinate vector and dimensionality reduction feature vector of the new model component 1 are stored in The database is used for online recognition and positioning, and then teach the robot to pick up the parts 1 to the designated bin 10. After adding the new model part 1 offline, during the online work process, a proximity switch 3 needs to be installed on the conveyor belt 2 to detect whether part 1 passes by. When the proximity switch 3 detects the part 1 and controls the LED light source 5 to turn on To trigger the area scan camera 4 to collect images, use the same image processing method to obtain the contour polar coordinate vector and the dimension reduction feature vector of the current conveyor belt 2, and then compare it with the polar coordinate vector and the dimension reduction feature vector in the database to determine The current part 1 model, and its translation and rotation relative to the offline teaching position, and then send the translation and rotation to the robot controller 7, and control the suction cup 9 of the end joint of the robot arm 8 to pick up the part 1 and place it on the specified material Box 10.

[0022] image 3 It is the comparison result of the polar coordinate vector of the current component edge contour and the polar coordinate vector of the teaching component.

[0023] Figure 4 It is the distribution of the polar coordinate vectors of the edge contours of the 3 types of products in the application example after being reduced to three dimensions. The calculation method of component contour polar coordinate vector dimensionality reduction is as follows:

[0024] Let (x, y) represent the two-dimensional coordinates of the image, (Cx, Cy) represent the center coordinates of component 1 obtained by image processing, and take (Cx, Cy) as the coordinate center, and set the contour coordinates of component 1 (x i ,y i ) Converted to polar coordinates (θ i,i r), where i=1,...,N, N represents the number of contour points extracted. The array {(θ i ,r i ),i=1,...,N} perform linear interpolation to obtain the r value at the fixed θ array to form the polar coordinate vector R.

[0025] Place a single component 1 at any position and angle within the camera's field of view, collect ten groups of images and process them to obtain a vector group R i ,i=1,...,10, calculate the vector group R through the kernel function k(x,y) i ,i=1,...,10 kernel matrix K, calculate the eigenvalue of matrix K λ=[λ 1 ,...,Λ 10 ], take the eigenvector corresponding to the first three largest eigenvalues ​​as the base vector.

[0026] Through the present invention, the relative movement position and rotation angle of the online component 1 and the teaching component 1 can be obtained, so as to control the robot to accurately suck the specified position of the component.

[0027] With regard to the limitation of the protection scope of the present invention, those skilled in the art should understand that on the basis of the technical solutions of the present invention, various modifications or variations that can be made by those skilled in the art without creative work are still within the protection scope of the present invention. Within.