Introduction to the ESP32-S31 with Vision: Assignment 2

Capture and display your first frames

#

In this assignment, you will create two ESP-Vision camera applications. The first continuously sends camera frames to the IDE preview. The second displays the live camera image on the ESP32-S31-Korvo-1 LCD.

How the script works

#

The application uses the sensor module:

Both applications use QVGA (320 × 240), which provides a responsive preview while keeping memory requirements moderate.

Step 1: Stream the camera to the IDE preview

#

Create camera_preview.py:

# SPDX-FileCopyrightText: 2026 Espressif Systems (Shanghai) CO LTD
# SPDX-License-Identifier: Apache-2.0

import sensor
import time


# Initialize the board camera.
sensor.reset()
# Capture frames in 16-bit RGB color at 320 x 240.
sensor.set_pixformat(sensor.RGB565)
sensor.set_framesize(sensor.QVGA)
# Let automatic exposure and white balance settle.
sensor.skip_frames(time=1000)

while True:
    # Capture a frame and send it to the connected IDE preview.
    img = sensor.snapshot()
    img.flush()
    # Yield briefly to keep the IDE responsive.
    time.sleep_ms(20)

This is the ESP-Vision helloworld.py example. img.flush() sends every captured frame to the connected IDE preview. The 20 ms delay yields between frames and keeps the development interface responsive.

Step 2: Open the IDE preview

#

1. Connect the Web IDE to the ESP32-S31-Korvo-1.
2. Create camera_preview.py.
3. Paste the script and save it to the device.
4. Select Run.
5. Open the camera preview in the Web IDE.
6. Confirm that the live camera image appears.
7. Stop the script before continuing.

Step 3: Show the camera on the LCD

#

Create a second file named lcd_preview.py. This application initializes the LCD and continuously writes camera frames to it:

import display
import sensor
import time


def main():
    # Initialize the board LCD.
    lcd = display.Display()
    try:
        # Initialize and configure the camera for 320 x 240 RGB frames.
        sensor.reset()
        sensor.set_pixformat(sensor.RGB565)
        sensor.set_framesize(sensor.QVGA)
        # Let automatic exposure and white balance settle.
        sensor.skip_frames(time=1000)
        print("LCD_PREVIEW:READY")

        while True:
            # Capture a frame and draw it on the LCD.
            frame = sensor.snapshot()
            lcd.write(frame)
            # Yield briefly between frames.
            time.sleep_ms(20)
    finally:
        # Release both peripherals when the script stops or fails.
        sensor.shutdown()
        lcd.deinit()


main()

Run lcd_preview.py. The live camera image appears on the board’s LCD instead of the IDE preview. Stop the script to verify that both the camera and LCD are released cleanly.

Complete working examples

#

# SPDX-FileCopyrightText: 2026 Espressif Systems (Shanghai) CO LTD
# SPDX-License-Identifier: Apache-2.0

import sensor
import time


# Initialize the board camera.
sensor.reset()
# Capture frames in 16-bit RGB color at 320 x 240.
sensor.set_pixformat(sensor.RGB565)
sensor.set_framesize(sensor.QVGA)
# Let automatic exposure and white balance settle.
sensor.skip_frames(time=1000)

while True:
    # Capture a frame and send it to the connected IDE preview.
    img = sensor.snapshot()
    img.flush()
    # Yield briefly to keep the IDE responsive.
    time.sleep_ms(20)

import display
import sensor
import time


def main():
    # Initialize the board LCD.
    lcd = display.Display()
    try:
        # Initialize and configure the camera for 320 x 240 RGB frames.
        sensor.reset()
        sensor.set_pixformat(sensor.RGB565)
        sensor.set_framesize(sensor.QVGA)
        # Let automatic exposure and white balance settle.
        sensor.skip_frames(time=1000)
        print("LCD_PREVIEW:READY")

        while True:
            # Capture a frame and draw it on the LCD.
            frame = sensor.snapshot()
            lcd.write(frame)
            # Yield briefly between frames.
            time.sleep_ms(20)
    finally:
        # Release both peripherals when the script stops or fails.
        sensor.shutdown()
        lcd.deinit()


main()

What you will see

#

• The first application shows a live 320 × 240 camera image in the IDE preview.
• In the second stage, the board LCD shows the live camera image and the output reports LCD_PREVIEW:READY.

What you learned

#

In this assignment, you:

• Initialized the board camera with the ESP-Vision sensor module
• Selected a pixel format and frame size
• Streamed live camera frames to the IDE with img.flush()
• Created a separate live LCD application with lcd.write(frame)
• Released camera and display resources during cleanup
• Used the Web IDE to edit, run, preview, and stop camera applications

Next step

#

Lecture 1: Available models for ESP-Vision

Return to the workshop main page