It’s not the computer that matters. The main thing is the program and it can be such that it is easy to set up.
But really no difference, use the program in play or not…sometimes one variant is better - sometimes another.
But not using programs when processing data is just silly.
Listen, I’m talking about computer vision. I have the same kind of computer vision as you, and I’m also working in this field. I use it to track certain elements on the roulette wheel and then predict which number the ball will land on.
Computer can’t tell the back spin in the ball, can it? A human can listen to the sound and notice the way the dealer spin the ball. This is one big hurdle for computer vision.
Hi, I recently went to a casino and recorded about 100 spins at 240fps for my own research. I’ve just seen your post, it’s very interesting, I shoot on my iphone and my videos don’t always show the whole roulette wheel, so if your suggestion is relevant, I can send you what I have as an example and then if anything, go again and take more pictures. It is also interesting about the wheels themselves, are any suitable (for example, I primarily look at the number, location and shape of the crystals, so that after the first impact on the crystal, the ball falls onto the wheel as much as possible without scattering) or are there any criteria?
If anything, you can write in Telegram: @joingop
P.S. Sorry if something is not clear, I do not speak English, I write through a translator.
First of all. You need to understand how visual ballistic works.
- The ball is spun on the ball track with some sort of backspin and each dealer is different
- The ball leaves the ball track and this can be predicted when and where it leaves on a ball track to a certain extent especailly if a wheel is tilted or has some sort of defect on the ball track guiding the ball to leave the ball track at exactly the same place every time.
- The ball would approach the rotor at an angle which can be predicted to a certain extent and also hit the pocket/fret/cone and scatter and this scatter could also be predicted to a certain extent.
- We need to get data such as barometric pressure, rotor speed, which deflector the ball drop and whether it is consistent, the ball timing on the ball track whether the same speed gives similar total time or correlate to the type of spin the dealer makes. The first number the ball hit the pocket/fret and the final number and also the angle the ball approach the rotor.
I am more interested in how are you guys are going to stream that 240fps? Equipment is easy and cheap but your idea clearly is streaming that 240fps real time over the internet.