|
|
import cv2
|
|
|
import numpy as np
|
|
|
from collections import deque, defaultdict
|
|
|
import time
|
|
|
|
|
|
|
|
|
trace_len = 20
|
|
|
min_area = 500
|
|
|
|
|
|
|
|
|
object_traces = defaultdict(lambda: deque(maxlen=trace_len))
|
|
|
long_term_memory = defaultdict(list)
|
|
|
next_object_id = 1
|
|
|
object_centroids = {}
|
|
|
|
|
|
def count_direction_changes(trace):
|
|
|
count = 0
|
|
|
for i in range(2, len(trace)):
|
|
|
v1 = np.array(trace[i - 1]) - np.array(trace[i - 2])
|
|
|
v2 = np.array(trace[i]) - np.array(trace[i - 1])
|
|
|
if np.dot(v1, v2) < 0:
|
|
|
count += 1
|
|
|
return count
|
|
|
|
|
|
def extract_features(trace):
|
|
|
if len(trace) < 2:
|
|
|
return [0, 0, 0, 0]
|
|
|
dx = trace[-1][0] - trace[0][0]
|
|
|
dy = trace[-1][1] - trace[0][1]
|
|
|
total_distance = sum(np.linalg.norm(np.array(trace[i]) - np.array(trace[i-1])) for i in range(1, len(trace)))
|
|
|
avg_speed = total_distance / (len(trace) + 1e-6)
|
|
|
direction_changes = count_direction_changes(trace)
|
|
|
return [dx, dy, avg_speed, direction_changes]
|
|
|
|
|
|
def ai_brain(trace, memory):
|
|
|
if len(trace) < 3:
|
|
|
return "Unknown"
|
|
|
dx, dy, speed, changes = extract_features(trace)
|
|
|
|
|
|
if len(memory) >= 5 and memory.count("Erratic") > 3:
|
|
|
return "Suspicious"
|
|
|
if speed > 150 and changes > 4:
|
|
|
return "Erratic"
|
|
|
if speed < 5 and changes == 0:
|
|
|
return "Idle"
|
|
|
return "Normal"
|
|
|
|
|
|
def get_color(i):
|
|
|
np.random.seed(i)
|
|
|
return tuple(int(x) for x in np.random.randint(100, 255, 3))
|
|
|
|
|
|
|
|
|
cap = cv2.VideoCapture(0)
|
|
|
ret, prev = cap.read()
|
|
|
prev_gray = cv2.cvtColor(prev, cv2.COLOR_BGR2GRAY)
|
|
|
prev_gray = cv2.GaussianBlur(prev_gray, (21, 21), 0)
|
|
|
|
|
|
while True:
|
|
|
ret, frame = cap.read()
|
|
|
if not ret:
|
|
|
break
|
|
|
gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
|
|
|
gray_blur = cv2.GaussianBlur(gray, (21, 21), 0)
|
|
|
|
|
|
|
|
|
delta = cv2.absdiff(prev_gray, gray_blur)
|
|
|
thresh = cv2.threshold(delta, 25, 255, cv2.THRESH_BINARY)[1]
|
|
|
thresh = cv2.dilate(thresh, None, iterations=2)
|
|
|
|
|
|
contours, _ = cv2.findContours(thresh, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)
|
|
|
current_centroids = []
|
|
|
|
|
|
for cnt in contours:
|
|
|
if cv2.contourArea(cnt) < min_area:
|
|
|
continue
|
|
|
(x, y, w, h) = cv2.boundingRect(cnt)
|
|
|
cx, cy = x + w // 2, y + h // 2
|
|
|
current_centroids.append((cx, cy))
|
|
|
matched_id = None
|
|
|
|
|
|
|
|
|
for object_id, last_centroid in object_centroids.items():
|
|
|
if np.linalg.norm(np.array([cx, cy]) - np.array(last_centroid)) < 50:
|
|
|
matched_id = object_id
|
|
|
break
|
|
|
|
|
|
if matched_id is None:
|
|
|
matched_id = next_object_id
|
|
|
next_object_id += 1
|
|
|
|
|
|
object_centroids[matched_id] = (cx, cy)
|
|
|
object_traces[matched_id].append((cx, cy))
|
|
|
trace = object_traces[matched_id]
|
|
|
|
|
|
behavior = ai_brain(trace, [m['status'] for m in long_term_memory[matched_id]])
|
|
|
long_term_memory[matched_id].append({'status': behavior, 'timestamp': time.time()})
|
|
|
|
|
|
color = get_color(matched_id)
|
|
|
cv2.rectangle(frame, (x, y), (x + w, y + h), color, 2)
|
|
|
cv2.putText(frame, f"ID {matched_id}", (x, y - 20), cv2.FONT_HERSHEY_SIMPLEX, 0.6, color, 2)
|
|
|
cv2.putText(frame, f"Behavior: {behavior}", (x, y - 5), cv2.FONT_HERSHEY_SIMPLEX, 0.5, (0, 255, 255), 1)
|
|
|
|
|
|
|
|
|
inactive_ids = [obj_id for obj_id in object_centroids if obj_id not in [id for id, _ in object_centroids.items()]]
|
|
|
for iid in inactive_ids:
|
|
|
object_centroids.pop(iid, None)
|
|
|
|
|
|
prev_gray = gray_blur.copy()
|
|
|
cv2.imshow("Motion AI", frame)
|
|
|
if cv2.waitKey(1) & 0xFF == ord("q"):
|
|
|
break
|
|
|
|
|
|
cap.release()
|
|
|
cv2.destroyAllWindows()
|
|
|
|
