I am currently working on a small project in which I visualise growth data of a plant. For the monitoring, I use various known sensors for condition monitoring. But so far I am missing the monitoring of the growth itself, i.e. the height and the volume of the plant. I had thought about possibly connecting an ESP32 cam and counting the green pixels there somehow via the IDE. But I'm not getting anywhere with that. Does anyone have any experience of how to monitor the growth itself?
OpenCV runs on the Raspberry Pi, but not on the ESP32.
There are robots with the ESP32-CAM, but they are connected to a smartphone or computer. The pattern recognition is not done on the ESP32.
There all kind of projects on a ESP32 for pattern recognition on the ESP32 itself. Perhaps the basic functions of this one can be used for a plant as well: GitHub - espressif/esp-who: Face detection and recognition framework.
What about a laser scanner ?
If the soil humidity stays the same, then the extra weight is from the plant above the soil. Perhaps the total weight is an indication of how big the plant is.
What if the base is suddenly tilted, then the little wind motion by the plant is an indication of its size.
Could you measure the capacitance of the plant to air ? That increases when its size increases.
A scan of the plant, using LIDAR could give a rough plant outline, especially if the LIDAR is moved around the plant. Once the initial scan is made, and area data has been, calculated the scan can be stored and when time of the next scan comes both scans can be compared for differences, and those differences would represent a change.
If the stem of the plant is growing longer, and the plant is fairly substantial, you could hack a cheap digital Vernier (there are examples of hacking one on the 'net) and clamp it to the stem.
Thank you for your helpful answers! I will now first record the growth height via an ultrasound sensor. Then I will integrate the ESP32 Cam and try to send photos to my Raspberry PI using Node Red and save them there.
If you are saving photos somewhere, you might as well process them on a PC. There is lots of public domain software that could be used for analysis, ImageJ, for example.
A search for "imagej plant growth" demonstrates that many people use it to quantify plant growth.
Here is one how-to guide: Simple Imaging Techniques for Plant Growth Assessment
If you doing a RPi4, then that’s where the camera should be connected to.
A RPi doing animal detection GitHub - EdjeElectronics/TensorFlow-Lite-Object-Detection-on-Android-and-Raspberry-Pi: A tutorial showing how to train, convert, and run TensorFlow Lite object detection models on Android devices, the Raspberry Pi, and more!