Need advice on Quality IP Cameras for a project

I have been tasked with setting up a system that records the packing progress of parts to customers. If the customer comes back later and says the part was not included in the packages, we can go back and verify that the part was included or not. Due to the nature of the parts being packed and the customers involved, the retention time of the footage has been requested to be at 2 years.

In planning stages Ive decided that a server rack with Synology Rackstation RS3618xs NAS with all required rack, expansion, drives etc for a total storage of 288TB to record from 12 Cameras, 8 Hours a day on Motion detect triggering.

Now here is where I need help…
The bosses want to be able to video a 4ft by 8ft table where the parts are tagged, bagged and laid out for the work order. In the video footage we want to be able to zoom in on the parts tags to find the part or parts in question the day they were prepped for shipping. My issues is…I cant seem to be able to find cameras that can view a 2in x 4in mailing label with the part information CLEARLY from 5 to 8 feet directly overhead. The Max I can go is 8ft up from the top of the table. Each Table will have its own fixed camera directly above.

Im testing a couple Amcrest Cameras…One is 5MP HD and one is 4K UHD. The 5MP camera zoomed in can’t read the label, lighting is all washed out and when I do get the lighting right the text is too blurry… The 4K is better but not much better than the 5MP. The 4K camera had washed out colors and I was able to read the text on the label but it too was blurry after some work with camera settings. Bother cameras exhibited an intense amount of video noise as well. I was very disappointed in the performance of these cameras.

With that said, does anyone know of cameras ideal for use in this type of application? I need something with superior clarity in video quality in resolution ranging from HD to 4K.
Camera requirements: Motorized Zoom and focus control, ONVIF protocol for use with Synology hardware. IP Communication with POE, basic motion detect done on the camera preferred to keep load off the Synology NAS. Cost is not a issue but the cheaper the better $999 for an axis is steep but if we have to use those, so be it. I just can’t seem to find any reviews or image samples of some of the better cameras out there.

To summarize the set up:
Table is 4ft by 8ft, camera mounted 8ft up and need to be able to zoom in and read a 2in by 4in label with the part information. 16-18pt bold fonts are used. Label contains Part#, Description, QTY, Bag#, OP#, W/O# 000000000

The cameras Im testing have very good images when looking out into a room or large area…but zooming in on near targets from 8ft up just doesn’t work very well. The company is ready to invest in this project, but I need to find the right cameras that will work the first time. Any and all suggestions about good cameras to use based on experience or knowledge of Optics, FOV and Sensor sizes are greatly appreciated.

Edit: Typo/Grammatical errors

This may or may not be a helpful comment, just thinking out loud…

4K is a resolution of 3840x2160
8 foot is 2440mm, 4 is (not suprisingly) 1220mm.
That gives 1.5 pixels per mm and 1.7 pixels per mm

Maybe that works better in DPI…
8 foot is 96 inches so 40 DPI…

If you print a label with 18pt bold at 40 dpi, can you read it?

1 Like

Or print the label as normal then scan it at 40 dpi might be a better way of thinking about it.

We had a client with similar requirements like 10 years ago. There layout was a little different but essential we put up Big Brothers to watch the rooms, and Little Brothers on the individual workspaces to track the the handling.

This was very helpful in fact. You steered me in the direction to research the math involved in selecting the proper hardware and made me realize the real underlying problem.
Thank you!

Now, going to go do some more researching and looking…

You are going to need to redesign your product tags, and probably change the lighting. Gray tags and big bold writing will help (maybe). As well as going to need some really high resolution to be able to read the tags. You would be better off with a camera to watch the parts get put in the box after they go across a barcode scanner to scan the product number. You can watch the part go into the box, and you can check the scan log to make sure it was the correct part. You could also print out (at the table) the scanned packing list, as well as save a copy to storage.

If you require cameras to read the tags, get black and white cameras, the resolution can be better, the dynamic range can be better, and the B/W files should be smaller (per resolution) because you don’t have the color information. There are some really good machine vision cameras out there that are used for high speed inventory control and quality control, but be prepared to pay for these as they as low volume so high price.

Motion detect is often flawed, you would be better off just recording the entire day to make sure you are getting all the video required.

A modification of procedure could also work, have the cameras only cover the box, and make the packers show the product tag to the camera as each item goes into the box. Modified tags will still help.

And why the two year time frame? Do you box and store product that long before shipping? I’d rescan everything before it went out the door in this case (yes, stacking wood twice).

On the other end, if the customer doesn’t check the box as soon as it arrives, it is on them if something turns up missing two years later. I’d give them like 60 days maximum unless some kind of contract says that up to two years may pass before opening the box. And I’d get a new contract as soon as current expires. Sounds like a products sold to government thing.

Could this be what some like to call a “technical solution to a management problem?”
Is there really such a widespread issue with mistakes in packing and/or customer dishonesty that the cost of such a reactive measure is justified over a more proactive solution, like improving the packing process or not doing further business with customers who make repeated claims?

Just food for thought. Not sure if OP is a third-party or internal IT, but if it were me I’d ask myself if I might be getting tangled up in a bit of a mess. Not trying to be negative. Just a warning.

1 Like

Everyone in this thread have been putting forward some good points and questions that our management and the IT staff have discussed many times. Yes, I am the IT staff with the company on this project, not a 3rd party IT business. The following is a description of the situation we are dealing with, we are a Manufacturing Co. that deal with fortune 500 companies…

I suppose it’s more of a “Cover our company’s you know what” as we deal with this multi-million dollar project for this global entity. Also the product does sit in its crate for up to a year and a half on the other side of the planet, so sending a replacement part isn’t easy or quick. Bags go into boxes, boxes into crates, and crates go into shipping containers the size of a semi truck. So a complete receiving process of unpacking, counting, and repacking on their end would cause more issues than management can justify. We are just needing to prove we did it right the first time so if there is an issue we can avoid any penalty.

Sounds like a government contract or a company that contracts with a government.

I’d go with cameras focused on the boxes and a scanner to scan tags going into boxes. Then maybe a scanner going into crates. If money was available, UPS and Fedex have built really good system to handle things like this in an automated process of sorting and tracking objects.

I’d again look into machine vision applications and see what you can get that can automatically read and notate the parts going into the boxes and the boxes going into crates. I’m guessing crates may be sub assemblies for repair, manufacturing, or upgrades on a “thing”.

I might suggest paying the packers a higher wage to make sure things are correct, but I don’ think that works in today’s workforce.

Having thought about this more and reading what others have said…

Track the parts and potentially record video earlier in the process, bag / box them with an RFID tag then video each item being added to the larger pack with an obvious visual indicator that the part was scanned (green light when the reader is triggered) along with an audit log of the tag number that was read with a time stamp and an ID for the production line / packing table.

It depends on what you or your customers consider to be proof; an audit log of the RFID tags that were added or an audit log + video showing all the items were scanned as they went in or do you need to be able to see the part number on the pack.

It sounds like there are better technical solutions than trying to get video of the packing process but if they are not acceptable to management or customers then it doesn’t matter how technically good they are.

If you go RFID tags, can’t you scan an entire box and pull out all the IDs at one time? I haven’t really played with this so I’m just guessing at this point. And if you go with this type of system, then you need to use the cheaper tags, the $1 each label type.

Also if you used RFID and you can mass scan the entire box/crate, then the receiver could also scan each crate before it get’s stacked to make sure the parts all arrive. A tag without a part attached would be a clear enough condition that you needed to send another one.

Again, I don’t know if you can mass scan these tags, but I think you can. Some small racing organizations are using this type of tag to scan vehicles as the go across the finish line of other timing points. Just simple adhesive backed tags, one on the vehicle and one on the helmet (small motorcycle racing). Might be worth testing and considering the expense to install this, I’d like to think that companies would be willing to bring a system to your warehouse and give you a proper demo (if they said it would work).

Sounds like a really nice project for you to get your teeth into @TFanch

You may not want a nice project but it sounds like you have one!