Periodic_Approach.py

import cv2
import pandas as pd
import numpy as np
from ultralytics import YOLO
import time
# Load the YOLO models
model1 = YOLO('yolov8m.pt')#pre-trained model
model=YOLO('best.pt')#construction dataset trained model

# Define a callback function to get the RGB values of the pixel where the mouse is moved over
def RGB(event, x, y, flags, param):
    if event == cv2.EVENT_MOUSEMOVE :
        colorsBGR = [x, y]
        print(colorsBGR)

# Create a window and set the mouse callback function
cv2.namedWindow('RGB')
cv2.setMouseCallback('RGB', RGB)

# Open the video file
cap = cv2.VideoCapture('22 5sec.mp4')
#total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))

# Read the list of classes from a file
my_file = open("classes.txt", "r")
data = my_file.read()
class_list = data.split("\n")
print("Class List:")
print(class_list)
total_frames = int(cap.get(cv2.CAP_PROP_FRAME_COUNT))
count = 0
prev_results = None
worker_results_dict = {}

size = (1020,500)
# Create a video writer to save the output
#result = cv2.VideoWriter('filename.mp4', cv2.VideoWriter_fourcc(*'MJPG'), 10, size)
stime=time.time()
while True:
    if(total_frames==count):
        break
    # Read a frame from the video
    ret, frame = cap.read()
    
    # Increment the frame counter
    count += 1
    print("Frame no.: "+ str(count))
#    if(total_frames==count):
#        break
    # Resize the frame to the desired size
    frame = cv2.resize(frame, size)
    
    # Wait for a key press
    #cv2.waitKey(33)
    
    # Run the YOLO model on the frame
    if count % (30*0.5) == 0 or count == 1:
        print("Prediction")
        results = model.predict(frame)
        prev_results = results
        # Store the results for later use
        
    # Extract the bounding boxes from the YOLO results
    a = results[0].boxes.boxes
    px = pd.DataFrame(a).astype("float")
    for index,row in px.iterrows():
        # Extract the class id and confidence score for the bounding box
        class_id = int(row[5])
        conf = row[4]
        # Check if the class is a person and the confidence score is above a threshold
        if conf > 0.3 and class_id == 5:
            # Extract the coordinates of the bounding box
            x1 = int(row[0])
            y1 = int(row[1])
            x2 = int(row[2])
            y2 = int(row[3])
            # Draw a rectangle around the bounding box
            cv2.rectangle(frame, (x1, y1), (x2, y2), (0, 255, 0), 1)
            # Extract the frame of the worker from the bounding box
            worker_frame = frame[y1:y2, x1:x2, :]
            # Run the YOLO model on the worker frame
            if count % (30*0.5) == 0 or count == 1:
                worker_results = model.predict(worker_frame)
                worker_results_dict[(x1, y1)] = worker_results
                # Store the results for later use
            else:
                worker_results = worker_results_dict.get((x1, y1), None)
                # Get the results from dictionary if available
            # Extract the bounding boxes from the worker YOLO results
            worker_a = worker_results[0].boxes.boxes
            worker_px = pd.DataFrame(worker_a).astype("float")
            helmet_wearing = False
            vest_wearing = False
            mask_wearing = False
            # Check if the confidence is above a certain threshold and the class is helmet or vest or mask
            for _, worker_row in worker_px.iterrows():
                label=""
                worker_class_id = int(worker_row[5])
                worker_conf = worker_row[4]
                if worker_conf > 0.25 and worker_class_id in [0,7]:
                    if worker_class_id == 0:
                        helmet_wearing = True
                    elif worker_class_id == 7:
                        vest_wearing = True
                #if worker_conf > 0.1 and worker_class_id in [1]:
                mask_wearing = True
            #Labeling the person accroding to detection
            label+=str(index)##
            if helmet_wearing and vest_wearing and mask_wearing:
                label += "yes"
                color = (0, 255, 0)#green
            else:
                label +="no "
                if not helmet_wearing:
                    label+="helmet "
                if not vest_wearing:
                    label+="vest "
                if not mask_wearing:
                    label+="mask"
                color = (0, 0, 255)#red
            cv2.rectangle(frame, (x1, y1), (x2, y2), color, 1)
            cv2.putText(frame, label, (x1, y1), cv2.FONT_HERSHEY_COMPLEX, 0.5, color, 1)

    #displaying the frame
        cv2.imshow("RGB", frame)

    #check for escape key
        if cv2.waitKey(1) & 0xFF == 27:
            break
etime=time.time()

print("Elapsed time: {:.2f} seconds", format(etime-stime))
#releasing the video capture object and closing all windows
cap.release()
cv2.destroyAllWindows()