As of now, I’ve successfully implemented 3 different systems involving up to 3 simple usb 2.0 webcams. The nature of each application is to receive a trigger from an embedded controller(via serial), take the picture and process the image using a windows computer, visual studios, and the pre compiled opencv 3.0 library, and then send back a pass or fail to the same controller.
The most recent system involves a bit more complex set up. I am using 2 imaging source cameras (model # DFK 42BUC03) that have a usb 2.0 interface as well as 2 usb 2.0 webcams from Pluggable technologies. I am working on a Dell optiplex 3020 series with windows 8, 2 pci usb cards that I’ve added on to it, as well as a serial port pci card (for board communication). The system usually works as intended, with very few false negatives. Recently, the system has been running into what I believe is a bandwidth issue. The machine will run for some time, and then on the next trigger, the image will not update for one of the cameras. I’ve implemented a “retrigger” option in the code and when retriggering, one of the cameras will update it’s image, the other will remain exactly the same almost as if the camera somehow disconnected, froze, or just stopped working (i.e. I put my hand in front of the camera and retrigger the system and the image still stays the same). The two imaging source cameras are 1/3” sensors with 1280 x 960 pixels and each of them is plugged into a separate pci card plugged into the motherboard. On Friday, 11/6/15, I updated the code to not use the VideoCapture class from opencv and instead use the camera’s native API to capture images and store into a Mat object. The issue occurs much less frequently now, but it still has occurred at least once since then with the imaging source cameras, and a separate time where the 2 usb cameras would not update.
I was wondering what else you would need to know to help resolve this issue and possible solutions to try.