app.py

#!/usr/bin/env python3
"""
Reference: <https://github.com/methylDragon/opencv-motion-detector>
"""

from __future__ import annotations

import sys

import cv2
import imutils
import numpy as np

device_index = int(sys.argv[1]) if len(sys.argv) > 1 else 0

# =============================================================================
# USER-SET PARAMETERS
# =============================================================================

# Number of frames to pass before changing the frame to compare the current
# frame against
FRAMES_TO_PERSIST = 10

# Minimum boxed area for a detected motion to count as actual motion
# Use to filter out noise or small objects
MIN_SIZE_FOR_MOVEMENT = 2000

# Minimum length of time where no motion is detected it should take
# (in program cycles) for the program to declare that there is no movement
MOVEMENT_DETECTED_PERSISTENCE = 100

# =============================================================================
# CORE PROGRAM
# =============================================================================


# Create capture object
cap = cv2.VideoCapture(device_index)  # Flush the stream
cap.release()
cap = cv2.VideoCapture(device_index)  # Then start the webcam

# Init frame variables
first_frame = None
next_frame = None

# Init display font and timeout counters
font = cv2.FONT_HERSHEY_SIMPLEX
delay_counter = 0
movement_persistent_counter = 0

# LOOP!
while True:

    # Set transient motion detected as false
    transient_movement_flag = False

    # Read frame
    ret, frame = cap.read()
    text = "Unoccupied"

    # If there's an error in capturing
    if not ret:
        print("CAPTURE ERROR")
        continue

    # Resize and save a greyscale version of the image
    frame = imutils.resize(frame, width=750)
    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)

    # Blur it to remove camera noise (reducing false positives)
    gray = cv2.GaussianBlur(gray, (21, 21), 0)

    # If the first frame is nothing, initialise it
    if first_frame is None:
        first_frame = gray

    delay_counter += 1

    # Otherwise, set the first frame to compare as the previous frame
    # But only if the counter reaches the appriopriate value
    # The delay is to allow relatively slow motions to be counted as large
    # motions if they're spread out far enough
    if delay_counter > FRAMES_TO_PERSIST:
        delay_counter = 0
        first_frame = next_frame

    # Set the next frame to compare (the current frame)
    next_frame = gray

    # Compare the two frames, find the difference
    frame_delta = cv2.absdiff(first_frame, next_frame)
    thresh = cv2.threshold(frame_delta, 25, 255, cv2.THRESH_BINARY)[1]

    # Fill in holes via dilate(), and find contours of the thesholds
    thresh = cv2.dilate(thresh, None, iterations=2)
    cnts, _ = cv2.findContours(thresh.copy(), cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

    # loop over the contours
    for c in cnts:

        # Save the coordinates of all found contours
        (x, y, w, h) = cv2.boundingRect(c)

        # If the contour is too small, ignore it, otherwise, there's transient
        # movement
        if cv2.contourArea(c) > MIN_SIZE_FOR_MOVEMENT:
            transient_movement_flag = True

            # Draw a rectangle around big enough movements
            cv2.rectangle(frame, (x, y), (x + w, y + h), (0, 255, 0), 2)

    # The moment something moves momentarily, reset the persistent
    # movement timer.
    if transient_movement_flag:
        movement_persistent_flag = True
        movement_persistent_counter = MOVEMENT_DETECTED_PERSISTENCE

    # As long as there was a recent transient movement, say a movement
    # was detected
    if movement_persistent_counter > 0:
        text = "Movement Detected " + str(movement_persistent_counter)
        movement_persistent_counter -= 1
    else:
        text = "No Movement Detected"

    # Print the text on the screen, and display the raw and processed video
    # feeds
    cv2.putText(frame, str(text), (10, 35), font, 0.75, (255, 255, 255), 2, cv2.LINE_AA)

    # For if you want to show the individual video frames
    # cv2.imshow("frame", frame)
    # cv2.imshow("delta", frame_delta)

    # Convert the frame_delta to color for splicing
    frame_delta = cv2.cvtColor(frame_delta, cv2.COLOR_GRAY2BGR)

    # Splice the two video frames together to make one long horizontal one
    cv2.imshow("frame", np.hstack((frame_delta, frame)))

    # Interrupt trigger by pressing q to quit the open CV program
    ch = cv2.waitKey(1)
    if ch & 0xFF == ord('q'):
        break

# Cleanup when closed
cv2.waitKey(0)
cv2.destroyAllWindows()
cap.release()