Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[WIP] Add Football Player Tracking #109

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

[WIP] Add Football Player Tracking #109

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
161 changes: 65 additions & 96 deletions Football/Football.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,7 @@
from torch.autograd import Variable
from tqdm import tqdm
from PIL import Image
import numpy as np

from Football.utils import *
from Badminton.utilities.datasets import *
Expand All @@ -28,7 +29,7 @@ def __init__(self,
weights_path='config/yolov3.weights',
class_path='config/coco.names',
img_size=416,
conf_thres=0.7,
conf_thres=0.6,
nms_thres=0.4,
tiny=False,
Windows=False):
Expand Down Expand Up @@ -215,6 +216,7 @@ def detect_players_image(self,
return None,None
else:
return out_img, coordinate

def detect_players_video(self, video_path, optimization=False,
frames_skipped_input=1, heatmap=False, output_path=None):

Expand All @@ -224,120 +226,87 @@ def detect_players_video(self, video_path, optimization=False,

fps = cap.get(cv2.CAP_PROP_FPS)
start_time = time.time()

print("Optimization set to %s" %str(optimization))
frames_skipped = frames_skipped_input



if heatmap:
list_of_all_coordinates=[]

if optimization:
print("Identifying players in every %s frame(s)" % str(frames_skipped) if frames_skipped > 1 else '')
count_of_frames=0
pbar=tqdm(total=total_frame)

while(1):
# reading the video frame by frame
ret,frame = cap.read()

if not ret:
break

# initialize a dictionary that maps strings to their corresponding
# OpenCV object tracker implementations
OPENCV_OBJECT_TRACKERS = {
"csrt": cv2.TrackerCSRT_create,
"kcf": cv2.TrackerKCF_create,
"boosting": cv2.TrackerBoosting_create,
"mil": cv2.TrackerMIL_create,
"tld": cv2.TrackerTLD_create,
"medianflow": cv2.TrackerMedianFlow_create,
"mosse": cv2.TrackerMOSSE_create
}
# initialize OpenCV's special multi-object tracker
trackers = cv2.MultiTracker_create()
tracker = OPENCV_OBJECT_TRACKERS['csrt']()
print("Using Tracker:", tracker.__class__.__name__)
pbar = tqdm(total=total_frame)
list_of_all_coordinates = []
i = 0
frame_count = 0
while(1):
# Read video frame
ret, frame = cap.read()
frame_count += 1
if (frame_count % 20 == 0):
i = 0
if not ret:
break
else:
(h, w) = frame.shape[:2]


# The idea is that we will only detect_players_image after every 5 or 'frames_skipped' frames and
# for the other frames we will calculate the weighted average of the previous location and next location to determine the position of the players

if count_of_frames % frames_skipped == 0:
out_frame, all_coordinates = self.detect_players_image(frame,
ret_img=1,
display_detection=False)
# out_frame conatins the new frame with the players detected and all_cooridnates contains the coordinates of the box(es)

if count_of_frames==0: #for the first frame
frame_list=[] #initialize a frame list (size will be frames_skipped)
for f in range(frames_skipped):
frame_list.append(frame)
#ensure first frame detects 2 players-->
if len(all_coordinates)!= 8:
all_coordinates.append(10)
all_coordinates.append(10)
all_coordinates.append(10)
all_coordinates.append(10)
previous_frame_coordiantes = all_coordinates
if heatmap:
current_coords_list=all_coordinates
else:
#for every frame read thereafter
current_coords_list = check_if_two_players_detected(previous_frame_coordiantes,
all_coordinates)
step_list = calculate_step(previous_frame_coordiantes,
current_coords_list,
frames_skipped)
for frame_no in range(1, frames_skipped):
frame_coords = get_frame_coords(previous_frame_coordiantes,
step_list,
frame_no)
frame_list[frame_no] = draw_boxes(frame_list[frame_no],
frame_coords)
if heatmap:
list_of_all_coordinates.append(current_coords_list)
else:
out_video.append(frame_list[frame_no])

previous_frame_coordiantes = current_coords_list

if heatmap:
list_of_all_coordinates.append(current_coords_list)
else:
out_video.append(out_frame)
frame_list[0] = out_frame

pbar.update(frames_skipped)
if cv2.waitKey(1) == ord('q'):
break

else: # For all the other frames that we have skipped
frame_list[count_of_frames % frames_skipped] = frame # Add the skipped frames to the frame_list so that they can be modified with the boxes

count_of_frames+=1

else:
pbar=tqdm(total=total_frame)

# NO OPTIMIZATION

while(1):
# Read video frame
ret, frame = cap.read()

if not ret:
break

(h, w) = frame.shape[:2]
out_frame, all_coordinates = self.detect_players_image(frame,
if i == 0:
print("Running Detector")
frame, all_coordinates = self.detect_players_image(frame,
ret_img=1,
display_detection=False)
while i < len(all_coordinates):
# Every object detected has 4 box properties: x, y, w, h
trackers.add(tracker, frame, tuple(all_coordinates[i:i+4]))
i += 4
print(len(all_coordinates) // 4, "objects added to tracker.")
else:
ok, bbox = trackers.update(frame)
if not ok:
print("Tracking Failed")
i = 0
else:
for newbox in bbox:
p1 = (int(newbox[0]), int(newbox[1]))
p2 = (int(newbox[0] + newbox[2]), int(newbox[1] + newbox[3]))
cv2.rectangle(frame, p1, p2, (200,0,0))

if heatmap:
list_of_all_coordinates.append(all_coordinates)
else:
out_video.append(out_frame)
out_video.append(frame)
# cv2.imshow("Frame", frame)

pbar.update(1)
if cv2.waitKey(1) == ord('q'):
keyPress = cv2.waitKey(1)
if keyPress != -1:
print("Key Pressed:", keyPress)
if keyPress == ord('q'):
break
if keyPress == ord('r'):
print("r pressed, resetting detector")
trackers = cv2.MultiTracker_create()
i = 0

cap.release()
pbar.close()
print("Video Analysed in :" + str(time.time() - start_time) + 's')

if heatmap:
return list_of_all_coordinates
else:
# ------------------------------
# ----------- TODO -----------
# ------------------------------
fourcc = cv2.VideoWriter_fourcc(*'XVID')
if output_path is None:
output_path = video_path.replace(".mp4", "-out.avi")
Expand All @@ -349,4 +318,4 @@ def detect_players_video(self, video_path, optimization=False,
cv2.destroyAllWindows()



2 changes: 1 addition & 1 deletion test_fb.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,3 +1,3 @@
from Football.Football import Detector
obj = Detector()
obj.detect_players_video('Football/images/abc.mp4')
obj.detect_players_video('Football/images/abc.mp4')