Automated in-line object inspection
A system and method for performing real-time quality inspection of objects is disclosed. The system and method include a transport to move objects being inspected, allowing the inspection to be performed in-line. At least one optical acquisition unit is provided that captured optical images of the objects being inspected. The captured optical images are matched to CAD models of objects, and the matched CAD model is extracted. A laser with an illumination light beam has a wavelength in the violet or ultraviolet range then conducts scans of the objects, which are formed into three-dimensional point clouds. The point clouds are compared to the extracted CAD models for each object, where CTF are compared to user input or CAD model information and the object is determined to be acceptable or defective based on the extent of deviation between the point cloud and the CAD model.
1. A method for performing in-line inspection of objects comprising:
positioning an object to be inspected on a transport;
moving the object along a transport path using the transport;
capturing an optical image of the object with an optical acquisition unit;
capturing point cloud data representing the object with a laser module;
selecting a CAD model for an object matching the object being inspected based on the captured optical image;
converting the CAD model to a CAD mesh, and extracting planes from the CAD mesh to generate CAD plane meshes;
extracting critical to function parameters from the CAD plane meshes, and converting the CAD plane meshes to CAD point clouds for comparison to the captured point cloud data; and
comparing the CAD point clouds to the captured point cloud data.
2. The method of claim 1 , further comprising identifying the object being inspected, and its position and orientation on the transport, based on the captured optical image.
3. The method of claim 1 , further comprising generating a heat map indicating differences between the CAD model and the captured point cloud data, and determining whether the heat map levels fall with predetermined tolerance levels for the object being inspected.
4. The method of claim 2 , wherein identifying the object being inspected includes generating bounding boxes on areas of interest in the captured optical image, and performing an object recognition process including pattern matching to CAD models stored in a database.
5. The method of claim 4 , further comprising performing translation and rotation of the CAD models in order to align the CAD data with the captured point cloud data.
6. A computer-implemented method for inspecting an object, the method comprising:
receiving a target three dimensional point cloud representing a rigid object obtained from a scan of an object with a laser module;
loading a projection library and a characteristic vector set of a reference 3D mesh model of the rigid object;
calculating a two dimensional projection of the target point cloud and generating a corresponding characteristic vector;
performing a coarse two dimensional registration of the target point cloud by identifying a rest position and rotation angle of the rigid object matching the target point cloud;
generating a corresponding initial geometric transformation between the target point cloud and a reference mesh model based on the two dimensional registration;
applying the initial geometric transformation to the target point cloud;
performing a fine three dimensional registration of the transformed target point cloud with the same reference mesh model;
generating a final geometric transformation between the target point cloud and the reference mesh model resulting from the three dimensional registration;
applying the final geometric transformation to the target point cloud;
measuring differences between the transformed target point cloud and the same reference mesh model, and recording the measured differences.
7. The computer-implemented method of claim 6 , wherein generating a projection library comprises computing a convex hull and center of mass of the reference mesh model.
8. The computer-implemented method of claim 6 , wherein the three dimensional registration comprises performing an iterative closest point calculation for determining the correspondence between target point cloud and reference mesh model.
9. The computer-implemented method of claim 6 , wherein the initial geometric transformation comprises a rotation and translation matrix.
10. The computer-implemented method of claim 6 , wherein the final geometric transformation comprises an affine transformation.
11. The computer-implemented method of claim 6 , wherein the reference mesh model and target point cloud are mapped or down-sampled to two dimensional images.
12. The computer-implemented method of claim 6 , further comprising performing a matching criterion to compute the difference between the transformed target point cloud and the reference mesh model.
13. The computer-implemented method of claim 6 , wherein the two dimensional registration comprises performing a principal component analysis of different views of the reference mesh model.
14. The computer-implemented method of claim 13 , wherein the principal component analysis is performed at various angles of different rest positions of the reference mesh model.