Hey all,
I have been working on this project for the better part of the year so far. It uses a VC0706 camera to take a picture, and a Photon to send it over WiFi to an external server. I have also placed an SD card on it to save images locally. Once that image is on a server, the image is classified using TensorFlow. The goal is to identify bicycles in my application, but this same sensor could be used to identify a variety of things once the model is retrained. This is version 1 of this project, so things are not perfect, and there are probably a lot of things that I could do better. Particularly, the speed at which the SD card saves an image is slow…
I got a ton of help from the forums here by reading different posts and replies from tons of people. I have everything working now, and I think I can give something back.
I have posted my sanitized code on GitHub. It includes my listener script for gathering the image data, as well as the project itself. I’ll eventually include my TensorFlow model with custom scripts as well at a later date.
Please be kind, this is my first big project like this! I have no problem updating some things to make this work better.
I saw tons of people had trouble with SD cards and the VC0706 image sensor specifically, and got my stuff to work after reading through dozens of these posts. My project uses a modified version of the Adafruit library to use the image sensor. I can try to answer some questions if you need it, otherwise the library itself there should work! I have also included the Python listener I created to get the data input from the Photon and save it to the server. Right now, it can handle 2 connections at a time, but can be modified to gather from more connections.
Please let me know if you have questions!