Inference on Boot - ESP32 CAM

Help
,
#1

Hi, good morning,

I am trying to use an ESP32 CAM to take photos and store them on an SD card in different folders depending on the classification they receive. In the first folder, I want to place the diseased plants, and in the second, the healthy ones.

I don’t want to use Wi-Fi, so I used the link GitHub - edgeimpulse/example-esp32-cam: Builds and runs an exported image classification impulse on ESP32 Cam to perform the inference at boot, but the code has problems with image_util.h, and I would like to know what I can do to fix it.

Right now I’m able to take the pictures and send to the sd card, but I want to classify the image right before sending the image to the card. Saving the img as “Figure 1 Classification Health”
Thank you.

#2

Hello!

The repo you are referring to has been deprecated for some time now.
The easiest might be to load the esp32_camera example and then implement your saving to sd card function:

Screenshot 2024-05-22 at 11.51.22
Screenshot 2024-05-22 at 11.51.221920×1295 82.4 KB

Best,

Louis

Best

#3

Hey, Louis

Thank you for responding.

I can already take photos with the ESP32 CAM and store them on the SD card. I would like to know how to use my Edge Impulse library to classify the images by naming them with their classes. That way, I would have the photo, classification (X or Y), and I could store it sufficiently on the SD card.

I need the photo to be classified at the moment it is taken.

Thank you for your help.