Frame rate and integration time are related to each other, and sometimes they are a compromise. If our process line is going, for example, two meters per second, and our samples have a specific size, what is the frame rate we need to have enough spatial resolution?

Let’s calculate that.

One question we always get is how to estimate the achievable sample resolution and sorting speed. This, of course, depends on the speed of the conveyor belt, the sample size, and the required spatial resolution of the sample. But we also need to understand the effect of integration time and frame rate.

First, the frame rate means how many measurements we are doing per second, and it is essential to understand that integration times multiplied by the frame rate should always be one or less.

For example, if we have a frame rate of 100 frames per second, one can use up to ten milliseconds of integration time.

I will give an example. We have a process line where the speed is about

One thousand millimeters per second, and we assume that we have pretty small samples that we need to identify and sort, flakes of the size of 10 x 10 millimeters. Suppose our experiment thinks that the frame integration time is ten milliseconds. In that case, it means that during this integration time, the flake will move about ten millimeters from position T1 to position T2.

And because we have only 100 frames per second, in this case, we would get only one, or even less than one, sample per flake, which means that we cannot recognize the flake.

In practice, we need at least 3 to 4 samples per flake. So, calculating backward, we need 250 frames per second, and the maximum integration time we could use is about 2.5 milliseconds.