3D & AI: The Perfect Match for Defect Inspections!

Abstract

This application note describes the use of Photoneo Phoxi 3D Scanner and Pekat Vision software for applications that require defect detections.

Problem Statement

Have you ever come across or left unsatisfactory reviews on online platforms because of the condition of the box the items arrived in? As most people know, first impressions last, so when receiving a damaged package, people may avoid that particular store the next time they buy items online. For larger items or big-ticket purchases such as televisions and desktop monitors, customers would often look at the quality of the outer cardboard box to ensure that their goods are safe inside. If the package is damaged, the product inside may or may not be damaged too, depending on the quality of the packaging.

 

Scan of 4 boxes with different heights using a 2D camera

Figure 1: Scan of 4 boxes with different heights using a 2D camera

This is why many factories and companies would use a 2D camera for inspection works. 2D cameras are able to scan for defects, but it may not be able to scan larger boxes, or boxes that are further away at high accuracy. For example, the working distance between the taller box on the bottom left and the camera is about 49cm, which is much shorter than the working distance between the shorter box on the top right, 53cm. This causes the defects on the top right not to be as visible as the ones on the bottom left box. 

Users would also need to control the lighting in order to see the defects on the boxes. This would mean they would need to get larger lighting options for those applications, which would be tedious for larger shipping boxes.

Hence, what tools can we use for this application?

Setup image with Photoneo Phoxi 3D Scanner S and 4 boxes of different sizes

Figure 2: Setup image with Photoneo Phoxi 3D Scanner S and 4 boxes of different sizes

Solution

For this application, the SODA VISION Tech Team used a Photoneo Phoxi 3D scanner to scan images of the boxes, and Pekat Vision Software to process the images for defects on the boxes.

 

3D Depth map of 4 boxes using Photoneo Phoxi 3D Scanner S

Figure 3: 3D Depth map of 4 boxes using Photoneo Phoxi 3D Scanner S

Photoneo’s Phoxi 3D scanner comes in five different sizes, ranging from XS to XL, and has a scanning range from 161 millimetres up to 3.7 meters. This allows users to scan items as small as a PCB or connector pins, or as large as euro pallets, all with high accuracy.

We used the Photoneo Phoxi 3D scanner S to capture four boxes with different heights and volumes in this application. From the scans, the 3D camera produces a depth map which captures imperfections, and even a dent that the 2D camera could not detect. Moreover, users do not have to worry about getting larger illumination options for larger objects, or worry about environment lighting during applications.

 

Figure 4: Comparison between 2D and 3D scans of 4 boxes with defects.
Top image: 2D image scans of boxes using an area scan camera
Bottom image: Resulting image of Photoneo Phoxi 3D Scanner and Pekat Vision Software to produce a 3D Detection Scan.

Furthermore, with the help of Pekat Vision Software, users can easily detect flaws on the boxes and highlight them in red. The software effectively finds anomalies, detects and classifies defects, and checks surfaces on materials and objects with the help of AI and focused-learning algorithms. Pekat Vision’s Focused-learning is a proprietary algorithm that combines multiple approaches since it does not require many images for AI training. Moreover, customers can enjoy its fast-processing speed and short training time needed for defect inspections with the software. 

This application setup is also able to ignore printings on boxes and focus on the defects, which 2D cameras may not be able to do. This setup is suitable for Coplanarity Inspection, Tyre OCR readings, and food inspections.

 

Reach out to us!

Find out how you can utilise Photoneo Phoxi 3D Scanner and Pekat Vision’s software in your applications today!