Anomaly detection system in the automatic placement of composites during the manufacturing of structural elements
A system for inspecting structural elements during their manufacturing capable of detecting anomalies in the automatic placement of composite materials through computer vision, comprising an image capture module that can be integrated into an ATL machine extracting data from the images, from which an artificial vision module obtains information on the anomalies detected with computer vision, information that a human-machine interface module automatically translates into a language understandable to humans. By avoiding the need to manually enter data for the inspection, human errors are avoided, and the information obtained for each inspection can be automatically saved for record storage. The system is designed to inspect large structural element compounds, such as aircraft wings.
1. A system to detect anomalies in an automatic placement of composite materials during a manufacturing of structural elements, comprising the following modules:
an image capture module that can be integrated into an automatic placement machine for composite materials and comprising at least one camera configured to capture images of a surface to be inspected during the manufacture of a structural element comprising the surface, the image capture module being configured to extract data from the images captured by the at least one camera;
an artificial vision module configured to receive the data extracted by the image capture module and to obtain, by means of a computer vision algorithm using the extracted data, information on anomalies detected on the surface; and
a human-machine interface module configured to receive the information of detected anomalies obtained by the artificial vision module and automatically translate the information received into a language understandable to humans;
wherein the artificial vision module is further configured to
identify, using a labeling algorithm for identifying cluster areas, all the areas that the computer vision algorithm detects as anomalies,
calculate a porosity parameter from a total area value of each cluster area and a Euler number, and
compare a given minimum size of areas to be inspected together with the calculated porosity parameter, rule out false positives in the detection of anomalies by the computer vision algorithm.
2. The system according to claim 1 , wherein the camera is incorporated in a moving part of the automatic placement machine for composite materials.
3. The system according to claim 2 , wherein the camera is incorporated into a head for automatic placement for composite materials configured to move in a plane above the surface to be inspected.
4. The system according to claim 1 , wherein the image capture module is further configured to:
obtain, through a control unit associated with the automatic placement machine for composite materials, a selected numerical control program that is configured to perform displacements for the automatic placement machine for composite materials during the manufacturing of an area of the surface;
locate the area of the surface to be inspected by measuring positions of the machine in the displacement carried out according to the selected numerical control program;
apply a rotational-translation transformation to the measured positions to obtain a camera position from which to execute an image capture routine over the located area of the surface.
5. The system according to claim 1 , wherein the image capture module is further configured to store the data extracted from the images captured by the camera and send the extracted data, stored or in real time, to the artificial vision module.
6. The system according to claim 1 , wherein the artificial vision module is configured to activate an automatic inspection execution after the automatic placement of composite materials on the surface and even during the manufacturing of the structural element.
7. The system according to claim 1 , wherein the artificial vision module is further configured to, based on the data extracted by the image capture module, identify using a machine learning algorithm to classify the structural element to be inspected and obtain information regarding the inspection associated with the identified structural element.
8. The system according to claim 1 , wherein the artificial vision module is further configured to perform image preprocessing comprising cleaning and cutting the images captured by the image capture module and image correction comprising converting the preprocessed images into binary images that the computer vision algorithm uses as input images.
9. The system according to claim 1 , wherein the computer vision algorithm is configured to compare input images with reference images that are images without anomalies.
10. The system according to claim 1 , wherein the human-machine interface module is configured to translate the information on detected anomalies obtained by the artificial vision module into a color code associated with types of anomalies.
11. The system according to claim 1 , wherein the human-machine interface module further comprises a storage sub-module configured to generate a record of all the information obtained by the artificial vision module in each automatic inspection carried out.
12. The system according to claim 1 , wherein the detected anomaly information comprises an indication of lack of material, excess of material, or misplaced material on the surface.