https://github.com/johnbigeon/thermal_cam_micropython
Micropython code to display MLX90640 thermal image on ST7735 LCD screen
https://github.com/johnbigeon/thermal_cam_micropython
esp32-c3 micropython mlx90640 st7735 thermal-camera
Last synced: 11 days ago
JSON representation
Micropython code to display MLX90640 thermal image on ST7735 LCD screen
- Host: GitHub
- URL: https://github.com/johnbigeon/thermal_cam_micropython
- Owner: JohnBigeon
- License: mit
- Created: 2025-05-24T09:58:37.000Z (about 1 month ago)
- Default Branch: main
- Last Pushed: 2025-06-09T13:25:35.000Z (18 days ago)
- Last Synced: 2025-06-09T14:26:43.837Z (18 days ago)
- Topics: esp32-c3, micropython, mlx90640, st7735, thermal-camera
- Language: Python
- Homepage:
- Size: 45.9 MB
- Stars: 0
- Watchers: 0
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Thermal imagery MLX90640 and ST7735 with micropython
## Concepts
The concept of this project is to display the MLX90640 imagery on a ST7735 LCD screen on esp32-c3 board with micropython.
Display of the MLX90640 image via a ST7735 LCD display.
Thermal image display via native micropython strategy.
Thermal image display via native micropython strategy.
## Components
### Preparation of MLX90640 for i2c communication
To enable the i2c communication of the MLX90640, the i2c pins should be soldered as mentioned here [https://github.com/adafruit/Adafruit_CircuitPython_MLX90640/issues/19].
### Wiring
Connect the devices as follow:
- pin_cam_sda = 20
- pin_cam_scl = 21
- pin_scl = 5
- pin_sda = 2
- pin_res = 0
- pin_dc = 3
- pin_cs = 9
### Breadboard
Breadboard integration for debug purpose.
## Software versionning
Different version of this software is proposed:
- *Native_micropython* is based on a micropython firmware with uploaded librairies. It's obviously slow since my code send each pixel individually (via x,y positions). In theory, I should be able to send stream of data, but not able to find it via micropython.
- *Precompiled_micropython* is based on a micropython firmware with embedded ST7735 [https://github.com/russhughes/st7789_mpy]. It's based on the driver written in C, which should be more efficient (i.e.: should send stream of data instead).
- *Precompiled_optimized* is based on a micropython firmware with embedded ST7735 [https://github.com/russhughes/st7789_mpy]. Here, the colorometry conversion is optimized too.
### Native micropython installation
The main script *main.py* should load these two librairies and upload the image received from the thermal camera on the LCD display.
#### Micropython firmware
First, upload the generic esp32-c3 firmware via:
```
esptool --port /dev/ttyACM0 erase_flash
esptool --port /dev/ttyACM0 --baud 460800 write_flash 0 ESP32_GENERIC_C3-20250415-v1.25.0.bin
```
#### Upload typing
Execute the following script from you linux terminal:
```
sudo apt install micropython-mpremote
mpremote mip install github:josverl/micropython-stubs/mip/typing.mpy
```
you should see:
```
Install github:josverl/micropython-stubs/mip/typing.mpy
Downloading github:josverl/micropython-stubs/mip/typing.mpy to /lib
Installing: /lib/typing.mpy
Done
```
### Precompiled micropython installation
Follow that:
https://github.com/micropython/micropython/tree/master/ports/esp32
#### Clone ESP-IDF
```
git clone -b v5.4.1 --recursive https://github.com/espressif/esp-idf.git
cd esp-idf
git submodule update --init --recursive
./install.sh esp32c3
source export.sh
```
#### Clone MicroPython
```
cd ..
git clone https://github.com/micropython/micropython.git
cd micropython
git submodule update --init --recursive
make -C mpy-cross
cd ports/esp32
mkdir -p /home/jbi/Micropython_prod/micropython/ports/esp32/user_c_modules
cd /home/jbi/Micropython_prod/micropython/ports/esp32/user_c_modules
git clone https://github.com/russhughes/st7789_mpy.git
```
```
nano micropython.cmake
```
add
```
include(${CMAKE_CURRENT_LIST_DIR}/st7789_mpy/micropython.cmake)
```
```
export IDF_TARGET=esp32c3
make BOARD=ESP32_GENERIC_C3 USER_C_MODULES=/home/jbi/Micropython_prod/micropython/ports/esp32/user_c_modules clean
make BOARD=ESP32_GENERIC_C3 USER_C_MODULES=/home/jbi/Micropython_prod/micropython/ports/esp32/user_c_modules -j$(nproc)
```
#### Add to the frozen
```
cd /home/jbi/Micropython_prod/micropython/ports/esp32
git clone https://github.com/russhughes/st7789_mpy.git user_c_modules/st7789_mpy
cd user_c_modules/st7789_mpy
```
#### Build with the C extension
```
cd ../../
make BOARD=ESP32_GENERIC_C3 USER_C_modules=user_c_modules/st7789_mpy clean
make BOARD=ESP32_GENERIC_C3 USER_C_modules=user_c_modules/st7789_mpy -j$(nproc)
```
#### Upload the firmware
```
cd build-ESP32_GENERIC_C3/
esptool --no-stub --chip esp32c3 --port COM69 --baud 115200 write_flash -z 0x0 .\firmware.bin
```
#### Source of the code
The display ST7735 and the sensor MLX90640 are based on the librairies deployed here:
- ST7735 [https://github.com/boochow/MicroPython-ST7735]
- MLX90640 [https://github.com/michael-sulyak/micropython-mlx90640]
### Drawback of different strategies
The main actual drawback of the *native_micropython* is the slow fps display and the colorometry conversion (around 13s per frame).
For fast fps, please refer to the *all .cpp* script here [https://github.com/weinand/thermal-imaging-camera].
The issue with *Native_micropython* has been discussed here:
- [https://www.reddit.com/r/raspberrypipico/comments/16q4jkj/help_finding_a_fast_library_for_18inch_st7735s/]
- [https://www.reddit.com/r/raspberrypipico/comments/zx32ak/very_slow_refresh_rate_for_st7735/]
For precompiled and precompiled and optimized methods, 4.6s and 1.5 s are actually measured.
## Debug purpose
If the thermal camera returns no image, please follow the Waveshare tutorial here [https://www.waveshare.com/wiki/MLX90640-D55_Thermal_Camera]