MicroPython 1.28 documentation

1

Open the IDE

Download and install OpenMV IDE for Windows, macOS, or Linux and launch the IDE.

2

Connect your cam

Plug the OpenMV Cam into your computer via USB. A blue heartbeat LED blinks when it's ready.

3

Run your first script

Click the socket-icon connect button in the IDE, then hit the green play arrow to run your first script.

Hello world

Examples

import csi
import time
import ml
from ml.postprocessing.ultralytics import YoloV8

csi0 = csi.CSI()
csi0.reset()
csi0.pixformat(csi.RGB565)
csi0.framesize(csi.VGA)

# Built-in single-class person detector model.
model = ml.Model("/rom/yolov8n_192.tflite",
                 postprocess=YoloV8(threshold=0.4))
clock = time.clock()

while True:
    clock.tick()
    img = csi0.snapshot()
    # predict returns a list per class of ((x, y, w, h), score) tuples.
    for class_dets in model.predict([img]):
        for rect, score in class_dets:
            img.draw_rectangle(rect, color=(0, 255, 0))
    print(clock.fps(), "fps")

Real-time person tracking

The on-board YOLOv8 model is a single-class person detector — int8 quantised and shipped in ROM.

Loaded from /rom/yolov8n_192.tflite — no SD card or download needed.

Runs in real time on NPU-equipped boards — the OpenMV N6 and AE3.

Train your own YOLOv8 in Ultralytics — same three lines of glue.

import csi
import math
import time

csi0 = csi.CSI()
csi0.reset()
csi0.pixformat(csi.RGB565)
csi0.framesize(csi.QVGA)
csi0.auto_gain(False)
csi0.auto_whitebal(False)

clock = time.clock()

while True:
    clock.tick()
    img = csi0.snapshot()
    for tag in img.find_apriltags():
        img.draw_detection(tag, color1=(255, 0, 0), color2=(0, 255, 0))
        deg = math.degrees(tag.rotation)
        print("ID %d  rotation %.1f deg" % (tag.id, deg))
    print(clock.fps(), "fps")

Locate and identify AprilTags

AprilTags are 2D fiducial markers — robust to motion blur and partial occlusion, and they give you full 3D pose.

Built-in detector — no model file or training needed.

Returns ID plus full 6-DoF pose — x/y/z translation and x/y/z rotation.

Use for robotics calibration, AR markers, and indoor positioning.

import csi
import time
import ml
from ml.postprocessing.mediapipe import BlazeFace

csi0 = csi.CSI()
csi0.reset()
csi0.pixformat(csi.RGB565)
csi0.framesize(csi.VGA)
csi0.window((400, 400))  # square window for best results

model = ml.Model("/rom/blazeface_front_128.tflite",
                 postprocess=BlazeFace(threshold=0.4))
clock = time.clock()

while True:
    clock.tick()
    img = csi0.snapshot()
    for rect, score, keypoints in model.predict([img]):
        img.draw_rectangle(rect, color=(0, 0, 255))
        ml.utils.draw_keypoints(img, keypoints, color=(255, 0, 0))
    print(clock.fps(), "fps")

Detect faces with BlazeFace

Google's BlazeFace is a lightweight TensorFlow Lite face detector that returns bounding boxes plus six landmarks per face.

Loaded from /rom/blazeface_front_128.tflite — pre-quantised, no download needed.

Six keypoints per face: eyes, nose, mouth, and ears.

No privacy concerns — frames never leave the camera.

import csi
import time

csi0 = csi.CSI()
csi0.reset()
csi0.pixformat(csi.RGB565)
csi0.framesize(csi.QVGA)
csi0.auto_gain(False)

clock = time.clock()

while True:
    clock.tick()
    img = csi0.snapshot()
    for code in img.find_qrcodes():
        img.draw_rectangle(code.rect, color=(255, 0, 0))
        print(code.payload)
    print(clock.fps(), "fps")

Scan QR codes from a live feed

The built-in QR decoder handles tilted, distorted, and partially occluded codes.

Each result also exposes version, ECC level, and corner coordinates.

Numeric, alphanumeric, binary, and Kanji data modes.

Returns the decoded payload as a Python string — ready to use.

import csi
import time

csi0 = csi.CSI()
csi0.reset()
csi0.pixformat(csi.RGB565)
csi0.framesize(csi.QVGA)
csi0.auto_gain(False)
csi0.auto_whitebal(False)

# LAB thresholds: (L_min, L_max, A_min, A_max, B_min, B_max)
thresholds = [
    (30, 100, 15, 127, 15, 127),   # red
    (30, 100, -64, -8, -32, 32),   # green
]

clock = time.clock()

while True:
    clock.tick()
    img = csi0.snapshot()
    for blob in img.find_blobs(thresholds, pixels_threshold=200):
        img.draw_rectangle(blob.rect, color=(255, 0, 0))
        img.draw_cross((blob.cx, blob.cy))
    print(clock.fps(), "fps")

Find blobs of color

find_blobs returns connected pixel regions matching one or more LAB thresholds.

Tune thresholds for your lighting — disable auto-gain and auto-whitebal first.

Pass multiple thresholds for multi-color tracking in one call.

pixels_threshold filters tiny detections; merge=True joins overlapping blobs.

import csi
import time

csi0 = csi.CSI()
csi0.reset()
csi0.pixformat(csi.GRAYSCALE)
csi0.framesize(csi.VGA)
csi0.window((640, 80))  # narrow strip for fast linear scanning
csi0.auto_gain(False)
csi0.auto_whitebal(False)

clock = time.clock()

while True:
    clock.tick()
    img = csi0.snapshot()
    for code in img.find_barcodes():
        img.draw_rectangle(code.rect, color=(0, 255, 0))
        print(code.payload, "(quality %d)" % code.quality)
    print(clock.fps(), "fps")

Read 1D barcodes

Find 1D barcodes anywhere in the frame and decode their payloads.

Powered by the ZBar library — recognises EAN, UPC, Code 39/93/128, Codabar, ITF, ISBN, and DataBar.

Use a windowed strip in grayscale for the fastest linear scan.

Each result has format, payload, rotation, corners, and a bounding rect.

import csi
import time
import ml
from ml.postprocessing.mediapipe import HandLandmarks

csi0 = csi.CSI()
csi0.reset()
csi0.pixformat(csi.RGB565)
csi0.framesize(csi.VGA)
csi0.window((400, 400))  # square window for the model

# Connections between the 21 keypoints — palm + 5 fingers.
hand_lines = ((0, 1), (1, 2), (2, 3), (3, 4), (0, 5), (5, 6),
              (6, 7), (7, 8), (5, 9), (9, 10), (10, 11), (11, 12),
              (9, 13), (13, 14), (14, 15), (15, 16), (13, 17), (17, 18),
              (18, 19), (19, 20), (0, 17))

model = ml.Model("/rom/hand_landmarks_full_224.tflite",
                 postprocess=HandLandmarks(threshold=0.4))
clock = time.clock()

while True:
    clock.tick()
    img = csi0.snapshot()
    # predict returns a list per hand: index 0 = left, index 1 = right.
    for detections in model.predict([img]):
        for rect, score, keypoints in detections:
            ml.utils.draw_skeleton(img, keypoints, hand_lines,
                                   kp_color=(255, 0, 0),
                                   line_color=(0, 255, 0))
    print(clock.fps(), "fps")