"""the tracker type depend on your choice"""

def __init__(self, type):

self.net = tfnet()

self.colors = []

self.trackerType = type

self.new_frame = []

self.bboxes = []

"""Create a tracker based on tracker name"""

def createTrackerByName(self):

trackerType = self.trackerType

trackerTypes = ['BOOSTING', 'MIL','KCF', 'TLD', 'MEDIANFLOW', 'MOSSE', 'CSRT']

if trackerType == trackerTypes[0]:

tracker = cv2.TrackerBoosting_create()

elif trackerType == trackerTypes[1]:

tracker = cv2.TrackerMIL_create()

elif trackerType == trackerTypes[2]:

tracker = cv2.TrackerKCF_create()

elif trackerType == trackerTypes[3]:

tracker = cv2.TrackerTLD_create()

elif trackerType == trackerTypes[4]:

tracker = cv2.TrackerMedianFlow_create()

elif trackerType == trackerTypes[5]:

tracker = cv2.TrackerMOSSE_create()

elif trackerType == trackerTypes[6]:

tracker = cv2.TrackerCSRT_create()

else:

# 这里应该主动抛出异常

tracker = None

print('Incorrect tracker name')

print('Available trackers are:')

for t in trackerTypes:

print(t)

raise NameError('trackers name error !')

return tracker

def detection(self,frame):

# obtain box through yolo

bboxes_person, bboxes_bicycle = self.net.object_detection(frame)

bboxes = overlap(bboxes_person,bboxes_bicycle)

return bboxes

def create_MutilTracker(self, frame, bboxes):

# Initialize mutiltracker with first frame and bounding box

multiTracker = cv2.MultiTracker_create()

for bbox in bboxes:

multiTracker.add(self.createTrackerByName(), frame, bbox)

self.colors.append((randint(0, 255), randint(0, 255), randint(0, 255)))

return multiTracker

def match(self):

pass

"""object detection every num frame"""

def tracking(self, num=30, wait_frame_num=30, filename="videos/riding.mp4", confidence=0.02, camera=False):

# Read video

# detect every num frame. wait wait_frame_num if there is no bicycle.we can define filename as camera.

if camera is not False:

video = cv2.VideoCapture(camera)

else:

video = cv2.VideoCapture(filename)

# Exit if video not opened.

if not video.isOpened():

print("Could not open video !")

sys.exit()

# establish output file, the same as video.

width = video.get(cv2.CAP_PROP_FRAME_HEIGHT)

height = video.get(cv2.CAP_PROP_FRAME_HEIGHT)

# width = video.set(cv2.cv.CV_CAP_PROP_FRAME_HEIGHT, 1080)

# video.set(cv2.cv.CV_CAP_PROP_FRAME_WIDTH, 1920)

fourcc = cv2.VideoWriter_fourcc(*'DIVX')

if not os.path.exists('/home/nvidia/Documents/yolo_bicycle/output'):

os.mkdir('/home/nvidia/Documents/yolo_bicycle/output')

out = cv2.VideoWriter('/home/nvidia/Documents/yolo_bicycle/output/'+self.trackerType+'.avi',fourcc, 20.0, (int(width), int(height)))

# define some variable using in the below

ret = True

num_frame = -1

bicycle_flag = False

change_flag = False

no_bicycle_frame = 0

multitrackers = []

while ret:

# Read a new frame

ret, frame = video.read()

self.new_frame = frame

if not ret:

break

'''detect whether there is bicycle

if not, wait 30 frame and detect again

if there is, start analysing'''

# To be honest, the logic is not clear enough.

if bicycle_flag is False and no_bicycle_frame > 0:

no_bicycle_frame -= 1

elif bicycle_flag is False and no_bicycle_frame <= 0:

self.bboxes = self.detection(frame)

if len(self.bboxes) > 0:

bicycle_flag = True

change_flag = True

else:

bicycle_flag = False

no_bicycle_frame = wait_frame_num

#print(bicycle_flag)

if bicycle_flag is True:

# initinal multitracker

num_frame += 1

if num_frame % num == 0:

bicycle_group = 0

if change_flag is True:

change_flag = False

else:

self.bboxes = self.detection(frame)

bicycle_groups = int(len(self.bboxes) / 2)

#print(bicycle_groups)

self.colors = []

# 这里没有匹配车与人，后续增加一个车与人的匹配，根据重合度匹配。match

# 这里创建了多个跟踪器，一个跟踪器跟踪一对人和车

while bicycle_group < bicycle_groups:

current_group = []

multitrackers = []

current_group.append(self.bboxes[bicycle_group*2])

current_group.append(self.bboxes[bicycle_group*2 + 1])

multitrackers.append(self.create_MutilTracker(frame, current_group))

bicycle_group +=1

num_frame = 0

continue

# Start timer

timer = cv2.getTickCount()

# Update tracker

multitracker_num = 0

successes = []

update_bboxes = []

for multitracker in multitrackers:

#if multitracker is not None:

success, update_bbox = multitracker.update(frame)

successes.append(success)

update_bboxes.append(update_bbox)

multitracker_num += 1

# Calculate Frames per second (FPS)

fps = cv2.getTickFrequency() / (cv2.getTickCount() - timer)

for update_bbox in update_bboxes:

bbox_leave = Judgement(update_bbox, confidence)

# Draw bounding box

bboxes_num = 0

for success in successes:

if success :

# Tracking success

for i, newbox in enumerate(update_bboxes[bboxes_num]):

p1 = (int(newbox[0]), int(newbox[1]))

p2 = (int(newbox[0] + newbox[2]), int(newbox[1] + newbox[3]))

self.new_frame = cv2.rectangle(frame, p1, p2, self.colors[bboxes_num], 2, 1)

bboxes_num += 1

# 这里需要处理跟踪失败，给一个提示即可。

#else :

# Tracking failure

#self.new_frame = cv2.putText(frame, "Tracking failure detected", (100,80), cv2.FONT_HERSHEY_SIMPLEX,0.75,(0,0,255),2)

# Display tracker type on frame

cv2.putText(self.new_frame, self.trackerType + " Tracker", (100,20), cv2.FONT_HERSHEY_SIMPLEX, 0.75, (50,170,50),2)

# Display FPS on frame

cv2.putText(self.new_frame, "FPS : " + str(int(fps)), (100,50), cv2.FONT_HERSHEY_SIMPLEX, 0.75, (50,170,50), 2)

# Display result

cv2.namedWindow("Video")

out.write(self.new_frame)

cv2.imshow("Tracking", self.new_frame)

k = cv2.waitKey(1) & 0xff

if k == 27 :

break

out.release()

video.release()