A Machine Vision based Industrial Inspection System for Bottle defects
Since most of the industries in Pakistan heavily rely on human-based Pre-Shipment Inspection (PSI) or Final Inspection of the product, which has high chances of human-error. Developed countries, such as US, UK, China and others use automated inspection of the products using Machine Vision systems, w
2025-06-28 16:30:05 - Adil Khan
A Machine Vision based Industrial Inspection System for Bottle defects
Project Area of Specialization Artificial IntelligenceProject SummarySince most of the industries in Pakistan heavily rely on human-based Pre-Shipment Inspection (PSI) or Final Inspection of the product, which has high chances of human-error. Developed countries, such as US, UK, China and others use automated inspection of the products using Machine Vision systems, which includes high speed smart cameras, high speed robotics systems. As the automation of industrial inspection is one of the fastest growing technologies in the world, industries in Pakistan are still using manual methods to inspect the quality of the finished products. By automating the inspection process, time will be saved and more products will be yielded with minimum possible chances of errors.
This project is basically a part of Industrial Automation process of a leading consumer product industry of Pakistan. To automate the process of Quality assurance of their manufactured product, especially one of their bottle filled product, we are designing a system using principles of Machine Vision to detects the defects which usually occur after manufacturing of their product, which includes Artwork alignment, Batch reading, Dent in the bottle shape and seal gap of the bottle cap. After gathering the images of the bottle on conveyer belt using high speed camera, we will use image processing techniques and computer vision principles, to compare the obtained results with the required detection of defects according to their TAMU (Target, Acceptable, Marginal, and Unacceptable) specifications. The unacceptable bottles will be then discarded with immediate effect.
Project Objectivesi) To grab frames of the bottles in quality assurance process which includes batch number, bottle shape defects (dents), artwork misalignment on bottles and seal gap on the bottle cap by using high speed cameras.
ii) To implement machine vision and image processing algorithms to process the images grabbed from the cameras.
iii) Using a workstation to make decision on the basis of comparison of bottle images with the given TAMU document.
It is not an easy task as the bottles placed on conveyor belts are moving with fast speed, thus making it harder in getting correct bottle images and detecting the above-mentioned defects. So accuracy of algorithms and optimization of the solution are one of the important factors.
Project Implementation MethodThe implementation method is smoothly paced taking initiative with first grabbing the images of bottles from the conveyor belt using one or more high speed cameras. The grabbed images will be then passed from the image processing techniques (which includes image segmentation, edge detection, character detection). We have to detect four important things from the bottles which are, Alignment Artwork, Dent, Batch reading, Seal gap. Using Image segmentation algorithms, the edges of the bottle will be detected and then it will be analyzed with the given TAMU document. Alignment Artwork and Batch reading will be done by using AI algorithms (mainly Computer Vision algorithms) to identify the obtained result and then compared with the given TAMU document. Recursive testing will be done to maximize the accuracy of the software model. After achieving maximum optimization, hardware interfacing will be started which includes high speed cameras and a peripheral component (such as micro controllers) with the conveyor belt and the system will be deployed on a work station (such as GPU). The peripheral component will be designed with the logic to make decision in real time to reject the product on conveyor belt or to pass it on.
Benefits of the ProjectSince this project is basically for automation of a leading consumer product industry of Pakistan, they will be the direct beneficiary of this project. Other than them, the scalability and flexibility of the project allows it to be implemented and optimized for different products for different inudstries.
Automation of one of the time consuming process of manufacturin of a product, which is Qaulity Inspection, is the main targeted benefit to be achieved by this project. Time will be saved and more and more product yielding will be done, with quick testing of the product by our developed solution which will be quite faster and will be more accurate than human inspection.
Technical Details of Final DeliverableA completely deployable solution, basically a Hardware/Software integrated system, having a high speed industrial camera and high performance work station with an experimental chamber. The technical chamber will contain a fixed camera with proper illumination using LED lights, and a fixed suface on which the testing object(product) will be placed and fixed. The surface will then rotate with several angles(like 45 degree) per image capture. In this way, 4-6 pictures of the object will be clicked and then send to repective modules of our project.
- The backside image will be feed to the batch number reading and verification module.
- The frontside image which contains the artwork will be feed to the artwork alignment module.
- All of the images will be feed to the dent detection and cap defects detection module.
The UI/UX design of the project will then show the results after all modules will finish their execution.
Final Deliverable of the Project HW/SW integrated systemCore Industry ManufacturingOther Industries Others Core Technology Artificial Intelligence(AI)Other Technologies OthersSustainable Development Goals Industry, Innovation and InfrastructureRequired Resources| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 79000 | |||
| HD Camera | Equipment | 1 | 35000 | 35000 |
| Industrial LED lights | Equipment | 4 | 250 | 1000 |
| Microcontroller | Equipment | 1 | 1000 | 1000 |
| Experimental Chamber | Equipment | 1 | 15000 | 15000 |
| Mechanical Motor | Equipment | 2 | 2000 | 4000 |
| Bottle Holder | Equipment | 1 | 2500 | 2500 |
| Camera holder | Equipment | 1 | 1500 | 1500 |
| Bar code Detector | Equipment | 1 | 10000 | 10000 |
| Stationary and Printing overheads | Miscellaneous | 10 | 900 | 9000 |