Glue Path Detection

Project Overview

Project Background

Accurately measure the step difference of pins to promptly identify and eliminate defective products, thereby reducing the defect rate.

Camera Selection

LMI Laser Line Scan Camera Gocator 2530

Detection Requirements

Height Accuracy ≤ 0.01mm
Width Accuracy ≤ 0.02mm
Measurement Cycle ≤ 4s

Solution

AI-Vision loads a glue-free template, performs template matching after image capture to extract the glue path part, and uses the glue path detection algorithm to directly detect the height and width values of the glue strip and determine whether the glue path is broken.

Design Concept

Execution Effect Display

• Project Result Display:

  
  

Project Process

1. Initialization

Select the Load Point Cloud tool to acquire the point cloud.

2. Preprocessing

1. Select the Add 3D Square Probe tool, frame the protruding part to obtain the center point coordinates and output them.
2. Select the 3D Position Adjustment Tool to adjust the x and y positions of the point cloud according to the output center point coordinates.
3. Select the Create ROI tool to frame the base area for plane fitting.
4. Select the 3D Plane tool, bind the ROI fitted plane output by the created ROI area, and adjust the point cloud to the zero plane.
5. Select the 3D Clipping tool to clip the point cloud near the glue strip.
6. Select the 3D Downsampling tool to downsample the point cloud and reduce the data volume of the point cloud.
7. Select the 3D Point Cloud Ordering tool to order the point cloud; the reason for ordering is that the Glue Path Detection - Trajectory operator requires ordered point cloud input.

3. Glue Path Detection

1. Select the Point Cloud to Depth Map tool to convert the ordered point cloud into a depth map.
2. Select the Glue Path Detection - Trajectory tool, select the glue path trajectory on the depth map, detect the width and height of the glue strip as well as whether it is broken, and output the results.