如何在不选择roi的情况下使用opencv Tracker参数

Question

有人可以告诉我在检测到行人后应该将哪个第二个参数传递给tracker.init（） ？ 互联网上只有已选择ROI的参数。 我试图传递一个变量rects ，但这给了我一个错误

import numpy as np
import cv2
import sys
from imutils.object_detection import non_max_suppression

cap = cv2.VideoCapture("video.mp4")
hog_descriptor = cv2.HOGDescriptor()
hog_descriptor.setSVMDetector(cv2.HOGDescriptor_getDefaultPeopleDetector())

tracker = cv2.TrackerKCF_create()
ret, frame = cap.read()

while (1):
    ret, frame = cap.read()
    (rects, weights) = hog_descriptor.detectMultiScale(frame, winStride=(16, 16),
                                                       padding=(8, 8), scale=1.05)




    rects = np.array([[x, y, x + w, y + h] for (x, y, w, h) in rects])
    pick = non_max_suppression(rects, probs=None, overlapThresh=0.65)
    for (xA, yA, xB, yB) in rects:
        cv2.rectangle(frame, (xA, yA), (xB, yB), (0, 255, 0), 2)
    ret = tracker.init(frame)

    ret, = tracker.update(frame)
cv2.imshow("video", frame)
if cv2.waitKey(1) == 27:
    break

cap.release()
cv2.destroyAllWindows()

Answer 1

根据opencv示例代码https://github.com/opencv/opencv_contrib/blob/master/modules/tracking/samples/tracker.py，tracker.init的第二个参数是边界框：

cv.namedWindow("tracking")
camera = cv.VideoCapture(sys.argv[1])
ok, image=camera.read()
...
bbox = cv.selectROI("tracking", image)
tracker = cv.TrackerMIL_create()
...
ok = tracker.init(image, bbox)

因此，如果要将检测到的行人用作ROI，则应使用：

bbox = rects[index]
tracker.init(image, bbox)

其中index是要跟踪的检测到的行人数量