We would like to integrate a few IP cameras on our production line to record events related to faults. Ideally when a fault occurs we would save a video stream 15 seconds before and after the trigger. I realize the recording of events prior to the trigger is an issue which is why I am posting this question here.
I have been looking at the Axis line of cameras but nothing has been purchased yet. For storgage I have been looking at Synology, Buffalo and QNAP. I haven’t used any of them so feedback about your experiences are welcome.
If you are buying multiple camera’s my advice would be buy one and make sure it works the way you want it to and has the resolution and clarity required. You mention Axis cameras’ and while I have not used them they are mentioned in the manual for the IP camera component, and thus the connection string example is already done: docs.inductiveautomation.com/di … era+Viewer
With regard recording 15 seconds before your event, I am about to do a similar venture in a few months and these are my initial thoughts (but not yet coded/tested). Store a rolling one minute of video, into a database. There was an integrator’s webinar on extending ignition functionality using databases. Once one minute has elapsed, if no fault conditions go back to the start of the minute and record over it again. That way the file size will remain small. I am tracking discrete movements so will likely use JPEG stills triggered each time the IO asks the machine to do something. Upon a fault I will cycle through the stills frame by frame and hopefully pinpoint the issue.
Be aware that 15 frames a second is only capturing 1/2 the data of a process occurring 30 frames/second. To catch an event for an error I recommend 30 frames/second capture rate. Now the network will need to be able to handle a data flow rate of 3 megabytes a second. Implementing fiber optic is the best solution from the camera to the video recorder using a hard drive. Implementing a 1 minute window for recording a video stream, as suggested previously, would also work but understand that the hard drive magnetic head transferring the digitized information to the hard drive will be working alot and MTBF comes into play here.for maintenance. But this is what is needed to insure quality images for analysis latter on. Also the frame rate of the video will depend upon the action occurring. Starting at 1/250 of a second is a good place to start, but it may need to go up to 1/1,000 of a second or higher depending on the speed of the action trying to be analyzed. Another note to mention is that the faster the frame rate will be a stronger light source will be needed to illuminate the viewing area… This has worked for me in the past.