Скачать или смотреть DIY smart camera #2: Using the Azure Computer Vision Resource with the ESP32-CAM.

This is an esp32-cam project that connects to the Microsoft Azure Computer Vision Resource. In this project I use the "Describe Image" API, which creates a caption and some tags for the image. It seems to work quite well for a wide range of every day objects.

The previous video with the stand alone neural network is here:
   • diy smart camera: ESP32-CAM with TFT  

The code for this is on GitHub:
https://github.com/jonathanrandall/di...

To get this working requires:
1. Use of esp32 board version 1.0.6.
2. Change the User_setup.h file in the TFT_eSPI library to reflect the pins you are using for SPI connection to TFT screen (i've provided my copy on the gihub repo).

Azure computer vision resource API reference:
https://westcentralus.dev.cognitive.m...

You can create a free Azure account here:
https://azure.microsoft.com/

The webserver part for this is adapted the from Robot Zero One selfie-cam project:
https://robotzero.one/esp32-cam-tft-s...

Inspiring project using clarifai cloud:
https://microcontrollerslab.com/esp32...

Note: if you want to connect the esp32 to an iphone hotspot, you need to make sure there are no spaces or apostrophes' in the ssid,. You can change this in settings - general - about. Also, you need to have 'maximum compatibility" selected in your hotspot settings.