python yolo object detect

import cv2 as cv

import time

Conf_threshold = 0.5

NMS_threshold = 0.4

COLORS = [(0, 255, 0), (0, 0, 255), (255, 0, 0),

          (255, 255, 0), (255, 0, 255), (0, 255, 255)]

class_name = []

with open('classes.txt', 'r') as f:

    class_name = [cname.strip() for cname in f.readlines()]

# print(class_name)

#net = cv.dnn.readNet('yolov4-tiny.weights', 'yolov4-tiny.cfg')

#net = cv.dnn.readNet('yolov4.weights', 'yolov4.cfg')

net = cv.dnn.readNet('yolov3.weights', 'yolov3.cfg')

#net.setPreferableBackend(cv.dnn.DNN_BACKEND_CUDA)

#net.setPreferableTarget(cv.dnn.DNN_TARGET_CUDA_FP16)

cap=cv.VideoCapture(0)

#cap = cv.VideoCapture('output.avi')

ret, frame = cap.read()

img_size_h, img_size_w, channels = frame.shape

    

model = cv.dnn_DetectionModel(net)

#model.setInputParams(size=(416, 416), scale=1/255, swapRB=True)

model.setInputParams(size=(img_size_w, img_size_h), scale=1/255, swapRB=True)

starting_time = time.time()

frame_counter = 0

while True:

    ret, frame = cap.read()

    frame_counter += 1

    if ret == False:

        break

    classes, scores, boxes = model.detect(frame, Conf_threshold, NMS_threshold)

    for (classid, score, box) in zip(classes, scores, boxes):

        if class_name[int(classid)] != "person":

            continue

        color = COLORS[int(classid) % len(COLORS)]

        label = "%s : %.2f" % (class_name[int(classid)], score *100)

        #label = "%f" % round(score *100 , 2)

        cv.rectangle(frame, box, color, 1)

        cv.putText(frame, label, (box[0], box[1]-10),

                   cv.FONT_HERSHEY_COMPLEX, 0.3, color, 1)

    endingTime = time.time() - starting_time

    fps = "FPS: %.2f" % (frame_counter/endingTime)

    # print(fps)

    cv.putText(frame, fps, (20, 50),

               cv.FONT_HERSHEY_COMPLEX, 0.7, (0, 255, 0), 2)

    cv.imshow('frame', frame)

    key = cv.waitKey(1)

    if key == ord('q'):

        break

cap.release()

cv.destroyAllWindows()