2024-09-28 08:59 AM
Hello
I need to trigger an action whenever a small object (aprox 1 cm diameter) is at a fixed distance (around 10 cm) of the sensor.
Using VL53L0X the readings I get are very much affected by the background, for example if the object is in range and the background (1 m behind) changes from dark to white, the reading changes to indicates a much larger distance.
What can be done to improve this behavior? Absolute accuracy is not important, but repeteability is.
Solved! Go to Solution.
2024-09-30 07:37 AM
The light from the sensor goes out in a 27 degree cone. And you get the average of all the photon arrival times.
So at 10cm, the Field of View is a circle 5cm in diameter. A small part of that is your 1cm target and the rest is what you get from the background.
What to do.
1) Use the VL53L4CD. It has an FoV of 18 degrees. So at 10cm the diameter of the Field of View is 3mm. Maybe still too large. But if you can make your target out of something reflective 3M retro reflective tape perhaps, you will do better. By having a reflective target, most of the photons will come from your target and not your background.
I do have to admit this is not *** proof. A very reflective background, and you are back to square one. But this is the easiest way to go.
2) Switch to the VL53L4CX. This sensor also has a 18 degree FoV, but works differently. It uses histograms. And it can return multiple targets within the FoV. So you get your one target at 10cm and you will also detect your background at 1M. The two targets do have to be 80cm distant from each other - or you get a mix of the two.
3) If you have a bit more money, you can go with the VL53L1CX. This part has a lens and can you create your own Region of interest within the Field of view. By narrowing the area you are looking in, you can look only for your target.
4) The VL53L4CB has a mix of the lens AND the histograms. It's our ultimate Single zone sensor. With a reduced Region of interest AND the ability to 'see' both the target and the background, you will have a solution.
But before you do anything else, see if you can use your current system and a more reflective target. That might solve you issue.
All the parts I mentioned are footprint compatible. And the software migration is not too bad.
- john
2024-09-30 07:37 AM
The light from the sensor goes out in a 27 degree cone. And you get the average of all the photon arrival times.
So at 10cm, the Field of View is a circle 5cm in diameter. A small part of that is your 1cm target and the rest is what you get from the background.
What to do.
1) Use the VL53L4CD. It has an FoV of 18 degrees. So at 10cm the diameter of the Field of View is 3mm. Maybe still too large. But if you can make your target out of something reflective 3M retro reflective tape perhaps, you will do better. By having a reflective target, most of the photons will come from your target and not your background.
I do have to admit this is not *** proof. A very reflective background, and you are back to square one. But this is the easiest way to go.
2) Switch to the VL53L4CX. This sensor also has a 18 degree FoV, but works differently. It uses histograms. And it can return multiple targets within the FoV. So you get your one target at 10cm and you will also detect your background at 1M. The two targets do have to be 80cm distant from each other - or you get a mix of the two.
3) If you have a bit more money, you can go with the VL53L1CX. This part has a lens and can you create your own Region of interest within the Field of view. By narrowing the area you are looking in, you can look only for your target.
4) The VL53L4CB has a mix of the lens AND the histograms. It's our ultimate Single zone sensor. With a reduced Region of interest AND the ability to 'see' both the target and the background, you will have a solution.
But before you do anything else, see if you can use your current system and a more reflective target. That might solve you issue.
All the parts I mentioned are footprint compatible. And the software migration is not too bad.
- john
2024-10-02 08:41 AM
Related question: can the maximum detection distance be somehow reduced? That is, prevent objects farther away from 50 cm to be detected? In other words, I want to detect objects at no more than 50 cm