Automated Inspection of Defects in Industrial Application of IIOT and AI
This automated inspection of tiles using AI consists of following parts: - Tiles detection and classify them as defected and perfect - Sorting of defected tiles so that only perfect tiles may follow further process. - Visualize the whole report for each targeted parameter in order
2025-06-28 16:25:14 - Adil Khan
Automated Inspection of Defects in Industrial Application of IIOT and AI
Project Area of Specialization Artificial IntelligenceProject SummaryThis automated inspection of tiles using AI consists of following parts:
- Tiles detection and classify them as defected and perfect
- Sorting of defected tiles so that only perfect tiles may follow further process.
- Visualize the whole report for each targeted parameter in order to take immediate actions.
We are heading towards our fourth industrial revolution where efficiency, mass scale production and less time consuming methods would be the key features the whole industry would rely on. To approach the needs of this revolution we are trying to make a system that with machine efficiency and without making human beings distress, can automatically detect the defected parts of the tiles and classify them as an unallowable to be processed further.
The tile would move on the conveyer belt and as the camera approaches, it will start its work due to AI enabled technique embedded in it, and would classify defected and perfect tiles. Then a microprocessor would signal the sorting mechanism to sort out the defected tiles and pass only the perfect tiles out of the system for further process this will maintain the work efficiency and status of the industry. Now that we have smooth and reliable process, we move towards the report part in order to have the insights so that profit/loss could be calculated and preferable actions could be taken. This will be done with visuals by applying a visualization tools to the targeted parameter, this will give a graphical and instant look to the concerned persons who are holding the authority, this will help them in having a summary report of whole process.
We started to develop this project so that following objectives could be achieved:
- An efficient and reliable system to be developed which can detect defects in tiles
- To make sure no any defected tile could be passed, sorting mechanism is built that will sort out the defected tiles and only perfect tiles will be allowed for further process.
- Visualization system will be developed that will tell a summary of defected, perfect and profit/loss in visuals to the managers and supervisors which will be easier for them to have a look at summary report and could take immediate actions if desires conditions are not being met.
- To reduce the costs and expenses and increase the work efficiency of industry.
- We start working on project first by brainstorming the idea. Assessed the idea that whether or not it's applicable or not and then our approach to meet the requirements.
- Then comes the implementation, we design a system with multiple methods and approaches and finalized the one that was efficient.
- We start to implement on local level to test the system in order to make sure it should meet all requirements.
- Analyze data and observe performance so that a finalized approach can be taken further to the mass level providers.
- Meanwhile, calculate budget in order to see where we would be standing after completing project and estimation is necessary so that budgeted approach should be the first priority.
- Also, risk assessment and losses calculation is also a part. This will facilitate in having the insights of project period which will be helpful in taking actions related to needs and requirements.
It's much difficult for a human to stand hours and hours for one particular task with focus throughout the time so that no that much defected tiles should be passed which counts to many. And in this era of machinery and robotics, we are well aware of the power and consistency that machines can give us, so why not try to reduce pressure on humans and give them some glimpses of calmness?.
By having the idea in our mind we start to work on alternatives and come to the benefits of AI and computer vision which are listed below :
- AI enabled our system can detect much defected tiles than humans without being overwhelmed.
- It will reduce the time consume by a human for one particular task and that time can be used somewhere as well.
- We humans feel less stressed when working in a pair or group and having a machine in partnering would definitely make workers feel relaxed.
Technical details of Final deliverables of our project are listed :
- An efficient and reliable system will be delivered which could detect the defects in tiles. This will be done by making the camera capable of having AI embedded in it. We would apply computer vision with AI techniques using deep learning libraries and the famous computer vision library, OpenCV.
After having multiple approaches we finalized an approach which was more efficient than others. That algorithm works on edges and pixel values to do the task, as our task was supervised learning based so when desired conditions were seemed to be met the AI enabled camera with that algorithm would signal the microprocessor for further process.
- secondly, a sorting mechanism will be embedded with it which would be able to sort out approx all the tiles that are considered defected by AI camera. After detection the microprocessor, with some specific commands so that only defected tiles could be separated out, will signal the mechanism which consist of solenoid rod (or maybe something else depends on efficiency) to separate out those defected tiles. Now that we have all the perfect tiles moving on our conveyer for further processes.
- further, visualization system will be made that could show the managers and supervisors for record of every loss/profit things. This is necessary as a company should have its insights in order to take further decisions in future.
Lastly, but not least, the one of the most important parts is our conveyer belt. We will also make a conveyer by ourself for a local purpose to test our system as a prototype, after that it can be implemented in industries without thinking about losses that could be faced.
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 71660 | |||
| Raspberry pi 4 4GB | Equipment | 1 | 24500 | 24500 |
| Raspberry pi power adapter | Equipment | 1 | 350 | 350 |
| SD card class (10) 64 GB | Equipment | 1 | 1220 | 1220 |
| SD card reader | Equipment | 1 | 80 | 80 |
| Ethernet cable | Equipment | 1 | 150 | 150 |
| VGA cable | Equipment | 1 | 200 | 200 |
| Node MCU | Equipment | 1 | 800 | 800 |
| Data cable | Equipment | 1 | 130 | 130 |
| Proximity Sensor | Equipment | 1 | 950 | 950 |
| Gear motor | Equipment | 1 | 6800 | 6800 |
| male /female sockets | Equipment | 2 | 100 | 200 |
| camera | Equipment | 1 | 3500 | 3500 |
| 1.5 inchs metal pipe 20 feet long for conveyor | Equipment | 2 | 2700 | 5400 |
| motor Belt 6 feet long | Equipment | 2 | 600 | 1200 |
| metal pully for conveyor | Equipment | 4 | 400 | 1600 |
| Ball Bearing for rollers and motor | Equipment | 5 | 370 | 1850 |
| Coupling for pully/ rollers | Equipment | 4 | 310 | 1240 |
| coupling of motor | Equipment | 1 | 800 | 800 |
| Belt guard of conveyor | Equipment | 1 | 8500 | 8500 |
| power supply of motor | Equipment | 1 | 3700 | 3700 |
| color spray | Miscellaneous | 1 | 190 | 190 |
| labour cost for welding and cutting of metal pipes | Miscellaneous | 1 | 6500 | 6500 |
| fuel cost during visit of industry for formation of data | Miscellaneous | 12 | 150 | 1800 |