Download OpenCV Social Distancing Alert System Project<\/h3>\n
Please download the source code of OpenCV Social Distancing Alert System: OpenCV Social Distancing Alert System Project Code.<\/strong><\/a><\/p>\n Open windows cmd as administrator<\/strong><\/p>\n 1. To install the opencv library run the command from the cmd.<\/p>\n To implement it follow the below step.<\/strong><\/p>\n 1. First of all we are importing all the necessary libraries that are required during implementation.<\/p>\n 2. The function detect_people detects people in a given frame. It converts the frame to grayscale and applies a cascade classifier specifically designed to detect people. It then returns the coordinates and dimensions of the bounding boxes around the detected people in the frame.<\/p>\n Output of this step<\/strong><\/p>\n 3. The function compute_distances calculates the distances between detected individuals based on their bounding boxes. It finds the centroid (center point) of each bounding box and computes the Euclidean distance between all pairs of centroids. The function returns the distances as a matrix and also provides a list of the centroids.<\/p>\n Output of this step<\/strong><\/p>\n 4. The violate_social_distancing function counts violations of social distancing based on the distances between individuals. It returns the total count of violations.<\/p>\n 5. The code sets the file path to the Haarcascade XML file for detecting full-body patterns. It then initializes a classifier object using that XML file, which can be used to detect full-body images.<\/p>\n 6. Pass the path of the video file in the VideoCapture function.<\/p>\n 7. Start the While loop.<\/p>\n 8. The code snippet reads the next frame from a video source. If no frame is successfully read or the video has ended, the code exits the loop and terminates.<\/p>\n 9. In this code snippet, the frame is resized to a size of 512×512 pixels. People are then detected in the resized frame using a cascade classifier. The distances between the detected individuals are calculated, and the number of violations of social distancing is determined based on these distances.<\/p>\n 10. For each pedestrian detected, a green rectangle is drawn around them on the frame using their coordinates. This visually highlights the detected pedestrians.<\/p>\n Output of this step<\/strong><\/p>\n 11. For each centroid representing a detected individual, a red circle is drawn on the frame at their coordinates. This provides a visual representation of the centroids of the detected individuals.<\/p>\n Output of this step<\/strong><\/p>\n 12. The loop iterates over each pedestrian detection in the pedestrian_detections list, allowing for individual processing or analysis of each detected pedestrian.<\/p>\n 13. The code snippet compares each pair of pedestrian detections to check if their distance is less than 50. If a violation is detected, it visually highlights the violating pedestrians by drawing red rectangles around them and displaying a “Violating” label near each of them on the frame. It also displays an alert message indicating a social distancing violation.<\/p>\n Note:-\u00a0<\/strong> write this for loop under the step 11th for loop.<\/p>\n 14. The code displays the processed frame in a window titled “DataFlair” and waits for a key press. If the pressed key is ‘q’, the program stops running.<\/p>\n Note:-<\/strong> Write steps 8-14 under the while loop of step 7.<\/p>\n 15. The code stops using the video capture and closes all OpenCV windows.<\/p>\nInstallation<\/h3>\n
pip install opencv-python<\/pre>\n
Let\u2019s Implement It<\/h3>\n
import cv2\r\nimport numpy as np\r\nimport time<\/pre>\n
def detect_people(frame, cascade):\r\n gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)\r\n detections = cascade.detectMultiScale(gray, scaleFactor=1.1, minNeighbors=5, minSize=(30, 30))\r\n return detections<\/pre>\n
![\"function](\"https:\/\/data-flair.training\/blogs\/wp-content\/uploads\/sites\/2\/2023\/12\/function-detect-people.webp\")
<\/a><\/p>\ndef compute_distances(detections):\r\n num_detections = len(detections)\r\n distances = np.zeros((num_detections, num_detections))\r\n centroids = []\r\n for i, (x1, y1, w1, h1) in enumerate(detections):\r\n cx1, cy1 = x1 + w1 \/\/ 2, y1 + h1 \/\/ 2\r\n centroids.append((cx1, cy1))\r\n for j, (x2, y2, w2, h2) in enumerate(detections):\r\n cx2, cy2 = x2 + w2 \/\/ 2, y2 + h2 \/\/ 2\r\n distances[i, j] = np.sqrt((cx1 - cx2)**2 + (cy1 - cy2)**2)\r\n return distances, centroids<\/pre>\n
![\"function](\"https:\/\/data-flair.training\/blogs\/wp-content\/uploads\/sites\/2\/2023\/12\/function-compute-distances.webp\")
<\/a><\/p>\ndef violate_social_distancing(distances, threshold=50):\r\n num_detections = distances.shape[0]\r\n violations = 0\r\n for i in range(num_detections):\r\n for j in range(i+1, num_detections):\r\n if distances[i, j] < threshold:\r\n violations += 1\r\n return violations<\/pre>\n
cascade_path = \"C:\/Users\/yoges\/OneDrive\/Desktop\/dataflair\/Harcascade\/opencv-master\/data\/haarcascades\/haarcascade_fullbody.xml\"\r\ncascade_classifier = cv2.CascadeClassifier(cascade_path)<\/pre>\n
cap = cv2.VideoCapture(\"C:\/Users\/yoges\/OneDrive\/Desktop\/dataflair\/Social Distancing\/pedestrian.avi\")<\/pre>\n
while True:<\/pre>\n
ret, frame = cap.read()\r\n if not ret:\r\n break<\/pre>\n
frame = cv2.resize(frame, (512, 512))\r\npedestrian_detections = detect_people(frame, cascade_classifier)\r\ndistances, centroids = compute_distances(pedestrian_detections)\r\nviolations = violate_social_distancing(distances)<\/pre>\n
for (x, y, w, h) in pedestrian_detections:\r\n cv2.rectangle(frame, (x, y), (x + w, y + h), (0, 255, 0), 2)<\/pre>\n
![\"green](\"https:\/\/data-flair.training\/blogs\/wp-content\/uploads\/sites\/2\/2023\/12\/green-rectangle.webp\")
<\/a><\/p>\nfor (cx, cy) in centroids:\r\n cv2.circle(frame, (cx, cy), 5, (0, 0, 255), -1)<\/pre>\n
![\"red](\"https:\/\/data-flair.training\/blogs\/wp-content\/uploads\/sites\/2\/2023\/12\/red-circle-.webp\")
<\/a><\/p>\nfor i in range(len(pedestrian_detections)):<\/pre>\n
for j in range(i+1, len(pedestrian_detections)):\r\n if distances[i, j] < 50:\r\n x1, y1, w1, h1 = pedestrian_detections[i]\r\n x2, y2, w2, h2 = pedestrian_detections[j]\r\n cv2.rectangle(frame, (x1, y1), (x1 + w1, y1 + h1), (0, 0, 255), 2)\r\n cv2.rectangle(frame, (x2, y2), (x2 + w2, y2 + h2), (0, 0, 255), 2)\r\n\r\n cv2.rectangle(frame, (x1 + w1, y1), (x1 + w1 + 150, y1 - 30), (0, 0, 255), cv2.FILLED)\r\n cv2.putText(frame, \"Violating\", (x1 + w1, y1 - 5), cv2.FONT_HERSHEY_SIMPLEX, 0.7, (255, 255, 255), 2)\r\n cv2.rectangle(frame, (x2 + w2, y2), (x2 + w2 + 150, y2 - 30), (0, 0, 255), cv2.FILLED)\r\n cv2.putText(frame, \"Violating\", (x2 + w2, y2 - 5), cv2.FONT_HERSHEY_SIMPLEX, 0.7, (255, 255, 255), 2)\r\n alert_message = \"ALERT! Social Distancing Violation\"\r\n cv2.putText(frame, alert_message, (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.7, (0, 255, 255), 2)<\/pre>\n
cv2.imshow(\"DataFlair\", frame)\r\n if cv2.waitKey(1) & 0xFF == ord('q'):\r\n break<\/pre>\ncap.release()\r\ncv2.destroyAllWindows()<\/pre>\n
OpenCV Social Distancing Alert System Output<\/h3>\n
![\"capture](\"https:\/\/data-flair.training\/blogs\/wp-content\/uploads\/sites\/2\/2023\/12\/capture-and-closes-all.webp\")
<\/a><\/p>\nOpenCV Social Distancing Alert System Video Output<\/h3>\n
![\"social](\"https:\/\/data-flair.training\/blogs\/wp-content\/uploads\/sites\/2\/2023\/12\/social-distancing-alert-system-using-opencv-video.webp\")
<\/a><\/p>\n