Real-time defect identification of products on a conveyor belt

[@pysource-com]

🔥Learn how to build your own AI vision solutions: pysource.com/community

[@romuloprieto2529]

My group is currently considering fault detection using CV, and to keep track of the process by sending the data to a web page, using bottles sounds like a great option, thanks for the ideas!

[@kafkaesk3449]

This is amazing. Thanks

[@baselbasel3320]

Thank you for your amazing tutorials

[@geeklypuff]

This is a great channel!

[@mohammadaliavazpour3268]

I appreciate this buddy ❤

[@umairyounas793]

Thank you for this brief video. Can you please share the detail and specs of the high speed camera used in this project and how many bottles scanned / captured in one secon. Did you just measure the sides length of the label or you also performed the text recognition or color detection etc as well ?

[@erikdahl95]

Thank you for a great tutorial! Do you need a set of images with defects to train your models? How many examples of every defect do you approximately need to build a robust model in your opinion?

[@pysource-com]

Hi Erik, yes you need a set of images with defects to train the model. There is not a specific number, it depends on the complexity of the things you want to detect, on how different the defects are, the variety of environmental changes and so on. In general I would start with a small sample of 1000 and increase it if necessary.

[@sharanyar1693]

Hi, can you please share your thoughts on identifying drift in image dataset? Eg - changes in lighting, focus, angle and other useful properties between 2 image datasets. Thanks.

[@jorgebarge2956]

Thank you for your amazing tutotrial video .Wath software do you use to do this

[@brasilpi3664]

I love Opencv

[@HA-mr5wn]

I am looking for blister inspection in pharma industry. The feature we looking to inspect like 1. Empty blister 2. Double tablet in one cavity 3.wrong or mixed tablet 4.broken tablet 5.to identify and distinguish between same colour of capsule and its blister backgrounds.

[@agustinpicabea9045]

Hi, what camaras do u use?? Thanksss

[@christinapmalliou2449]

Is this included on your RCNN or your Computer Vision courses?

[@JonathanRamirez-by5kc]

Hello, Which videos from your uploads can I watch in order to build something like this. I have a project for school and I would like to implement an automated QA control. These things are not taught at my college. Your help will be greatly appreciated.

[@jacamilelouwiei.3075]

i hope next video is about region of interest in live capture

[@joaobarreira5221]

Hi, I am developing a system that detects if the object is damaged. This detection is done in an uncontrolled environment, by which I mean under different weather conditions, scales and rotations. As we have little data, and we probably won't be able to obtain more, I'm thinking about using "feature matching" since it's likely that Yolo and similar techniques won't be robust. Have you tried using "feature matching"?

[@pysource-com]

nope, it would be very unreliable that way. Data is very important for the accuracy, so I would focus on that aspect first

[@joaobarreira5221]

@@pysource-com Hi, Maybe I wasn't precise about the amount of my data. I have a lot of data on the undamaged object and I can acquire more. What I'm missing are more images of the damaged object and I can't acquire more. The dataset is very unbalanced. I think I have enough data to train a "feature matching" model and I can combine it with other models, such as Object detection + "zernike moments", ... I don't think I have enough data to create a robust model with yolo or for simple classification. Do you know of other techniques that can result in robust models?

[@clydearellano4739]

Good day to you sir, i have a capstone project that detects egg size, the problem is that its android base so i have to build an opencv sdk with contrib modules. The problem is that my laptop cannot perform to build the sdk. Can you help me with this?

[@AbdurRahman-x9i1l]

Think you for this nice tutorial . Can you share me your dataset?

[@jihadabdurrahman8142]

Sangat membantu luarbiasa, saya memiliki tugas akhir yang menjadi syarat kelulusan tingkat diploma III elektronika industri, dan saya sangat membutuhkan ilmu ini karna sesuai dengan penelitian dan alat yang sedang saya buat, jika berkenan bisakah anda membuatkan video setiap metode deteksi yang digunakan dalam sistem ini, terimakasih banyak sudah sangan membantu

[@hushamibadi3480]

Hi dear, I've been watching ur tutorials on object detection .. however I'm working on such project where need to send opencv output video to web browser. Knowing that i do not want to use flask or any network related tool. I need simply the video to appear on my local host webpage . I tried using some thing like ffmpeg but couldn't link it with opencv output. I hope u can help me in that . Best wishes

[@hemachandhers]

i want the source video of bottle without detection can you provide me the link please

[@UmapathyVanmathi]

Hai where to learn these concept but I have shortage of money

[@yusufismailalrazaz6781]

Thank you, sir, for the excellent video. May I know how much is the cost required to develop this program?

[@pysource-com]

Hi, The cost depends on many factors (complexity of production line, type of products, device constraint, environment and so on. To ask for a quote for commercial solutions you can contact us here: pysource.com/services/

[@hhhh-vc1zt]

Thanks very much.I have a question -- How to detect the Water level?

[@pysource-com]

The water level could be a simple classification task comparing images with the right amount of water and the one that have lower/higher level than they should have and calssify them as wrong. Or if it's necessary to get exactly the line, it would be a segmentation task where the deep learning model will draw the line.

[@syedjameelahmed19]

@@pysource-com I dont know why your videos are underrated, I have really found your videos to be helpful. If you dont mind I have one request Please can you make a video to measure exact water level like if the water is 100 ml then we use CV and Deeplearning to measure it to be 100ml.

[@montsersabri5624]

Thank you for the inf🌹 Can i get code source and dataset

[@swapnilgaware590]

where can i get this dataset

[@bobbyraduloff]

евала баце

[@FarahDarwish-ui6ho]

where can i find this video and dataset ?

[@ahmedsobeeh6999]

yes where?

[@xuantai2233]

i want buy code

[@HungNguyen-hd3mb]

Can you send me the code?

[@guzzler6584]

Hi