Skip to content

Instantly share code, notes, and snippets.

Created September 26, 2021 19:33
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
Star You must be signed in to star a gist
What would you like to do?
OpenMV Electronic Component CNN Live Classification with LCD
OpenMV Live Image Inference
Continuously captures images and performs image using provided TFLite model
file. Outputs probabilities in console. Displays preview on LCD backpack along
with predicted label.
Author: EdgeImpulse, Inc.
Modified: Shawn Hymel
Date: September 26, 2021
License: Apache-2.0 (
import sensor, image, time, tf, lcd
# Settings
model_file = "trained.tflite" # Location of TFLite model file
labels_file = "labels.txt" # Location of labels file
width = 48 # Width of frame (pixels)
height = 48 # Height of frame (pixels)
pixel_format = sensor.GRAYSCALE # This model only supports grayscale
# Configure camera
sensor.set_pixformat(pixel_format) # Set pixel format to RGB565 or GRAYSCALE
sensor.set_framesize(sensor.QVGA) # Set frame size to QVGA (320x240)
sensor.set_windowing((width, height)) # Crop sensor frame to this resolution
sensor.skip_frames(time = 2000) # Let the camera adjust
# Extract labels from labels file
labels = [line.rstrip('\n').rstrip('\r') for line in open(labels_file)]
# Set up LCD
# Start clock (for measureing FPS)
clock = time.clock()
# Main while loop
inference_count = 0
# Update timer
# Get image from camera
img = sensor.snapshot()
# Do inference. OpenMV tf classify returns a list of prediction objects.
objs = tf.classify(model_file, img)
# We should only get one item in the predictions list, so we extract the
# output probabilities from that.
predictions = objs[0].output()
# Find label with the highest probability
max_val = max(predictions)
max_idx = predictions.index(max_val)
# Draw label with highest probability to image viewer
labels[max_idx] + "\n{:.2f}".format(round(max_val, 2)),
mono_space = False)
img.draw_string(0, 38, "FPS: {:.1f}".format(round(clock.fps(), 2)), mono_space = False)
# Scale image and display on LCD
lcd.display(img, (16, 32), x_scale=2.0, y_scale=2.0)
# Slow printing to the console by only printing every few inferences
#inference_count += 1
#if inference_count >= 10:
#inference_count = 0
## Print all the probabilities
#for i, label in enumerate(labels):
#print(str(label) + ": " + str(predictions[i]))
## Uncomment this if you want to see FPS measurement
#print("FPS:", clock.fps())
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment