Hello,
for a design, I am developing I need to see if an object( in my case, a bottle) is present on a shelf. There are 3 rows and 12 columns for each shelf, which is about 60 cm x 150 cm. I can use further than 1 cam to see the objects.
I never used any type of image recognition, so I would like to understand if you suppose it’s a good idea- in your opinion- to use a cam to capture the bottle on the shelf and also make an image recognition.
What do you think about it? Could it be possible? What instruments and libraries should I use?
Thanks!
The problem will be simpler if you can use a single camera above the shelf, and good contrast colors. also, you're just counting circular patches of not-shelf- color or bottle-cap- color. OpenCV might be oversupply for that, but probably has the exact law already written. Still, counting gets harderRead more
The problem will be simpler if you can use a single camera above the shelf, and good contrast colors. also, you’re just counting circular patches of not-shelf- color or bottle-cap- color. OpenCV might be oversupply for that, but probably has the exact law already written.
See lessStill, counting gets harder, If you must look edge- on the shelf. But just detecting something is on the shelf( not counting) might be as simple as painting the wall behind the shelf a contrasting color and counting patches of not- background- color with further than N pixels.
Good lighting will be important.