water meter / Wasserzähler - ESP32-CAM

thingiverse

This is a system for taking pictures of a water meter to monitor household water consumption as part of a home automation system. It enables users to check and track their water usage. Update 23.02.2020: A tool has been made available to help create the configuration file (config.ini) at http://wasserzaehler.ignorelist.com. This tool is still in its beta phase, but it works well for creating references and ROIs for number recognition. The system consists of three main components: 1. Round clamping adapter to my water meter - a screwless fixation system on the water meter that can be adjusted to fit individual geometries. 2. Tube towards camera - made up of two parts to allow for height adjustments. 3. Camera holder for ESP32-CAM with OV2640, which has an inbuilt flash LED and does not require additional illumination. The 3D printed parts are connected using a self-made spherical bayonet connection (available at https://www.thingiverse.com/thing:3270713). The camera holder contains the electronics (ESP32-CAM with OV2640) and is based on the scalable box (https://www.thingiverse.com/thing:3202128), consisting of three parts: 1. Bottom with bayonet connection to tube. 2. Inlay for fixing ESP32-CAM and USB connector for power supply. 3. Top cover (with venting holes and USB connector). The only wiring required is the power supply for the ESP32-CAM module (5V or 3.3V and GND). It's recommended to use a 5V supply, as it provides more stability. To use the system as a water meter, the fixed focus objective of the OV2640 needs to be adjusted, as it only produces sharp images at distances greater than ~40cm. This adjustment will result in an image of the water meter that is too small for automated readout processing. **ATTENTION: This will void any warranty as the sealing of the lens objective is broken!** The fixation glue on the OV2640 lens was removed using a sharp knife (see picture). After this, the objective could be screwed in and out. Rotating it by about a quarter counterclockwise resulted in a focus plane about 10cm to the lens. The user needs to determine their best setting with some trial and error for their specific environment. To use the system, two parts of code are required: 1. Http-server on the ESP32-CAM module: https://github.com/jomjol/water-meter-picture-provider 2. Server for image processing and readout of the camera image: https://github.com/jomjol/water-meter-system-complete As the installation of the server involves Tensorflow and OpenCV, it can be a bit tricky. A docker container is provided at https://hub.docker.com/r/jomjol/wasserzaehler for an intel processor driven environment (intel - jomjol/wasserzaehler:latest) and one for an Raspberry PI (armv7 - jomjol/wasserzaehler:raspi-latest). There's an intensive discussion on the usage and details in the iobroker forum. This discussion is in German. An overview of the whole system with theoretical background can be found at https://github.com/jomjol/water-meter-picture-provider. The construction was done using FreeCAD, and the FCStd files are available for download. Although they're not self-explanatory or well-commented. Another adapter for a water meter is also available, thanks to user Stev8176 from the iobroker forum. For more information, don't hesitate to contact me! Have fun, jomjol P.S.: This system is an improved version of https://www.thingiverse.com/thing:3238162. Improvement: - Using ESP32-CAM with OV2640 instead of ESP8266 and ArduCAM. By this, the additional external illumination is not needed, and the electronics are significantly cheaper.