I'm doing a project where I measure the X and/or Y movement of multiple bamboo 'stalks' in the wind and translate that into parameters I'm using in sound/music projects.
My first experiments were with accelometers, I got excellent results with an ADXL 354 sensor attached around 1 meter height, sensing the movement with low noise and high accuracy.
Now, I'm actually thinking of changing the type of sensor. Is there is not a simpler, more low-cost solution? I've been thinking about distance sensors that are attached at the bottom of the stalk, the part that moves much less in the wind, but allows maybe a contact-sensor like a linear potentiometer to be attached to physically measure the movement in one axis or perhaps a very short distance optical sensor with a sort of 'wall' around the bamboo stalk placed to measure the distance?
I'm wondering if there are any thoughts from members here on what feasible options could there be, keeping in mind, high accuracy over a short distance (mm's), low-cost. It doesn't matter to me if it is a contact or non-contact way of measuring distance.
How can strain gauges be glued to bamboo ? I think a metal pole with a pin is required and the bamboo stick is placed on the metal pin. Those metal poles with strain gauges do exist and they can measure the force in X and Y direction. I don't know what they are called
A strain gauge is accurate.
A cheap solution which does the same is two piezo elements at the bottom in a semi-flexible base (silicone for example).
It is not accurate at all, but it might work.
A magnet inside the bamboo or attached to the bamboo and hall sensors at the X and Y axis. The magnet and the hall sensors can be put at the bottom in a semi-flexible base.
I can think of more things (optical, resistive, inductive, capacitive), however, a accelerometer with analog outputs might be the best and most simple solution after all.
You can reduce the noise by using the average of a few analogRead(). How the accelerometer is powered and the "analog reference" of the Arduino must also be correct.
Thanks everyone, I love how there is a variety of approaches and it let's me think in different ways.
I've worked with linear hall-effect sensors before, that can definitely be an approach. The strain gauges and Velostat are interesting, first, I thought, it might be a bit to 'invasive', but it can be attached with a temporary method perhaps. I agree, in the end, maybe the accelerometers might just be the most ideal setup possible, quick to attach, quite accurate etc.. I know this is the approach they use in scientific projects where they want to measure trees moving in the wind.
I have thought of video, but I haven't gone that much into that part of programming, and I like the more 'low-tech' approach of using sensors.
There are perhaps 1,000 different varieties of bamboo. If you are trying to measure the construction type bamboo, there will be no flexing unless you measure the upper, still growing section.
Paul
Using a 32bit MCU, like the Adruino 33 BLE or an ESP32 and the TensorFLow Lite or ESP32 TensorFLow Lite library. Add a camera, like the ESP32-CAM. Once the tensor has been trained to 'see' the plant in no wind, a result set can be obtained of the plants movements during wind.
