Help with test of HC-SR04 o4 or URM-09 sensor versus moving golf ball

Posted by: @allenk

↑

If anyone has a project with hypersonic sensor (HC-SR04 or URM-09) - can I ask for a rough test how far it senses the golf ball.
Something like - if you roll a golf ball in front of the sensor at 10"(25cm), 20"(50cm), 30"(75cm) - does it trigger the sensor ?

I've been working on getting reliable readings from the HC-SR04. So far, for a stationary sensor, I can get repeatable readings to about 20-30cm. After that the readings become erratic. I need a large flat target surface to get more stable readings. I have the sensor mounted on an Adafruit Mini Pan-Tilt kit. If I rotate the sensor (analogous to a moving target) the readings become more erratic but on long target surfaces they'll stabilize over time. But I'm still working out the testing process, i.e., variable rotation, sampling distance, and rate, etc. The sensor does detect a target up to 70cm, but erratically. So, you'll get readings like 20, 20, 20, ..., 70, 50, 65. 69, 73, ...

Thanks for the details

i I think your results confirm, I would not be able to sense the golf ball accurately.

additionally the rounded ball surface might reduce accuracy more.

thank you, time to look at other sensors, designs 🙂

I can pretty much ascertain it will not help you locate a tiny object precisely, if at all...

It's probably doing a big average of the various delays of the reflected wave it's receiving, so if you pick up anything it will rather be the average distance to the background (that's covering most of the field of view of the sensor) and not the tiny speck of reflection from your golf ball that will probably be considered an outlier and discarded.

If you want to have something really precise you'll probably have to resort to using a laser distance sensor, and probably an array of them so they can track a precise tiny object.

If I had to make such a project myself, I would probably use a fixed set of stereo cameras to track anything moving and determine the distance of the object depending on the difference between the left & right images...

But that's immediately another difficulty level than plugging a simple ultrasound sensor! 😆 

@patapom thanks for the insight - it helps greatly.
Shifting approaches to a Time Of Flight sensor - we'll see if it works .... I might also have to adjust from tracking golf ball to tracking large inflatable beach ball ... yeah that will work 🙂 🙂
Thanks everyone for comments, even reading this!

Another option is a radar, but once again it might miss a tiny object entirely unless you can program it to not ignore the ball as an outlier but as the actual object you're after...

You could maybe try this board?

https://www.hlktech.net/index.php?id=1157

You´re absolutely right! I need to check the pins I´ve been using...

Unfortunately, the Xiao wasn´t showing as a USB device even when it wasn´t connected to anything after some (still unknown) mistake I made at some point, whereas it was showing fine when I tried it straight from the box...

But it seems to be manageable now. Although not ideal, I can work with that. 😅

Ignore this post as I made a mistake, wrong thread... 🤣 

(is there a way to delete old messages or is it sealed after a while?)

Thanks for the replies (also glad I am not the only one that does that) 🙂

Everyone I just saw this post and it might help in my project.

Getting Started with LIDAR - YouTube

Using an inexpensive LiDAR sensor can solve the ultrasonic range challenge.
Simply using 3 LiDAR sensors to detect range of a golf ball at 2 or more times would give me speed and direction.

would LiDAR sensors like the TF mini detect a moving golf ball within about 15-20 feet?
Thanks

Posted by: @allenk

↑

would LiDAR sensors like the TF mini detect a moving golf ball within about 15-20 feet?

How will using stationary range sensors help track motion?

AFAICT, only one would produce variable output and then only for a very short period and distance. The others would report the ball is present then gone.

How do the commercial golf swing devices work? I think they have to be able to track the ball in some fashion.

Thanks - my plan was using 3 sensors, detect and triangulate an xyz position, then detect again (within the time still in range), triangulate the next xyz position. With that minimum of 2 detections a speed and direction could be determined. Any input on challenges/issues is greatly appreciated, thanks

Hi, for information: from my tests with the Waveshare ToF mini sensor, I know it can detect a USB cable at 10cm. I was testing the sonsor at 20cm from an object when the cable got in the way. Nice surprise! So far, maximum range of these sensors that I've seen is 300cm (across my office).

Posted by: @allenk

↑

my plan was using 3 sensors, detect and triangulate an xyz position, then detect again (within the time still in range), triangulate the next xyz position.

The "triangulation" implies changing what the sensor is pointing at. For an initial location P1, in order to measure the distance to P2 the sensors must be physically repositioned. Even if P1 and P2 are coaxial with one sensor, the other 2 sensors must move. Otherwise, you'll measure how quickly the object goes out of range at P1. For a golf ball, that's like, what, 20 mm? Coordinating that is...um...a challenge.

I'd recommend you dig into how golf ball launchers work. I initially though they used something similar to laser projectors (that us DMD chips). But I don't know if there's a version that can range find; I suspect there is. The launchers I looked at said they used both lasers and cameras.

Good luck.

TfMccarthy, thanks again

“ measure the distance to P2 “

that’s the exact issue I am working thru.

the LiDar sensors are stated range of 6m, and my bad swing of 70-90 mph… I might/hope to get P1 and P2 both in the same sensors (same direction not counting wayward balls) before going out of range. I think this is what the inexpensive monitors are doing. But the high end ones I think are using fast cameras.

assuming my two point track and speed works the 3 sensors, board, etc - gets close to the cost of used monitors.…. But this challenge is more fun 🙂 🙂

thanks