mouse_tracker2.py

# USAGE

from collections import deque
import numpy as np
import argparse
import imutils
import cv2
import random
from datetime import datetime
from scipy import misc
import math

date_print = datetime.strftime(datetime.now(), '%Y-%m-%d_%H-%M-%S')
print date_print

# construct the argument parse and parse the arguments
ap = argparse.ArgumentParser()
ap.add_argument("-v", "--video",
	help="path to the (optional) video file")

args = vars(ap.parse_args())

# changed: no max length 
pts = deque()
time = deque()

counter = 0
#(dX, dY) = (0, 0)

# Start time? 
start_time = datetime.now()


# if a video path was not supplied, grab the reference
# to the webcam
if not args.get("video", False):
	camera = cv2.VideoCapture(0)
	 
	# grab the current frame
	(grabbed, frame) = camera.read()
	average_image = frame
	#frame_first = imutils.resize(frame_first, width=600)
	average_image = cv2.cvtColor(average_image,cv2.COLOR_BGR2GRAY)
	average_image = cv2.GaussianBlur(average_image, (13, 13), 0)

# otherwise, grab a reference to the video file
else:
	camera = cv2.VideoCapture(args["video"])
	
	# extract background area from randomly sampled frames of the video
	frame_count = int(camera.get(7))
	print "Number of frames: " + str(frame_count)

	for i in range(100): # take 100 randomly sampled frames from the video
		random_frame = int(random.random()*frame_count)
		camera.set(1,random_frame)
		(grabbed, frame) = camera.read()

		if not grabbed: 
			break		
		frame = np.array(cv2.cvtColor(frame,cv2.COLOR_BGR2GRAY),ndmin=3)
		if i == 0:
			average_image = frame
					
		else:
			average_image = np.concatenate((average_image, frame), axis=0)

		#print average_image.shape

	average_image = average_image.mean(axis=0)
	average_image  = np.array(average_image, np.uint8)
	average_image = cv2.medianBlur(average_image, 13)
	cv2.imwrite("export/" + args.get("video") + ".png",average_image)
	

# set frame back to beginning of video:
camera.set(1,1)


# define 
def imclearborder(imgBW, radius):

    # Given a black and white image, first find all of its contours
    imgBWcopy = imgBW.copy()
    contours = cv2.findContours(imgBWcopy.copy(), cv2.RETR_LIST, cv2.CHAIN_APPROX_SIMPLE)[-2]

    # Get dimensions of image
    imgRows = imgBW.shape[0]
    imgCols = imgBW.shape[1]    

    contourList = [] # ID list of contours that touch the border

    # For each contour...
    for idx in np.arange(len(contours)):
        # Get the i'th contour
        cnt = contours[idx]

        # Look at each point in the contour
        for pt in cnt:
            rowCnt = pt[0][1]
            colCnt = pt[0][0]

            check1 = (rowCnt >= 0 and rowCnt < radius) or (rowCnt >= imgRows-1-radius and rowCnt < imgRows)
            check2 = (colCnt >= 0 and colCnt < radius) or (colCnt >= imgCols-1-radius and colCnt < imgCols)

            if check1 or check2:
                contourList.append(idx)
                break

    for idx in contourList:
        cv2.drawContours(imgBWcopy, contours, idx, (0,0,0), -1)
        print 'deleted ROI'

    return imgBWcopy

frame_zaehler = 0
# keep looping
while True:
	frame_zaehler = frame_zaehler + 1
	 
	# grab the current frame
	(grabbed, frame) = camera.read()

	if args.get("video") and not grabbed:
		break

	if frame_zaehler == 1:
		frame_sicherung = frame.copy()


	frame_grey = frame
	frame_grey = cv2.cvtColor(frame_grey,cv2.COLOR_BGR2GRAY)
	frame_grey = cv2.medianBlur(frame_grey, 13)
	frame_grey = average_image - frame_grey 
	frame_grey[frame_grey < 50] = 255

	#frame_grey(frame_grey < 50) = 255
	threshold_lower = 165
	threshold_upper = 255
	ret1,mask = cv2.threshold(frame_grey,int(threshold_lower),int(threshold_upper),cv2.THRESH_BINARY)
	#ret, mask = cv2.threshold(frame_grey,0,255,cv2.THRESH_BINARY+cv2.THRESH_OTSU)
	mask = 255-mask
	kernel = np.ones((3,3),np.uint8)
	mask = cv2.erode(mask,kernel,iterations=1)
	mask = cv2.morphologyEx(mask, cv2.MORPH_OPEN, kernel)

 	Iclear = imclearborder(mask, 5)

	cnts = cv2.findContours(Iclear.copy(), cv2.RETR_EXTERNAL,
		cv2.CHAIN_APPROX_SIMPLE)[-2]

	center = None
	distance = []
	sorted_idx = 0

	#  proceed if at least one contour was found
	if len(cnts) > 0:
		# sort for distance to center
		if len(cnts) > 1:
			for idx in xrange(len(cnts)):
				((x, y), radius) = cv2.minEnclosingCircle(cnts[idx])
				# get distance euclidian
				calc_distance = math.sqrt((x-frame.shape[0]/2)**2+(y-frame.shape[1]/2)**2)
				distance.append(calc_distance)
			sorted_idx = distance.index(min(distance))


		M = cv2.moments(cnts[sorted_idx])
		((x, y), radius) = cv2.minEnclosingCircle(cnts[sorted_idx])

		# why? because M (moments) are much smoother ... 
		# and this double calculation doesn't slow down algorithm too much 
		try:
  			center = (int(M["m10"] / M["m00"]), int(M["m01"] / M["m00"]))
		except: 
			center = [int(x),int(y)]

		#  proceed if radius has minimum size
		if radius > 10:
			
			cv2.circle(frame, (center[0], center[1]), 4,
			 	(0, 255, 0), 2)

			# TIME IN MILLISECONDS!
  			movie_ms = camera.get(0)
			time_pts = str(movie_ms) + ',' + str(center)[1:-1]
			pts.append(time_pts)

			
   		cv2.putText(frame, 'x: {0:.{2}f}, y: {1:.{2}f}'.format(x,y,1),
			(10, frame.shape[0] - 10), cv2.FONT_HERSHEY_PLAIN,
			1, (255, 255, 255), 1)


	time_difference = datetime.now() - start_time
	time_difference = divmod(time_difference.total_seconds(), 60)[0]*60 + divmod(time_difference.total_seconds(), 60)[1]

	cv2.putText(frame, "Elapsed time[s]: {0:.{2}f}, fps: {1:.{2}f}".format(time_difference,frame_zaehler/time_difference,1),
		(10, frame.shape[0] - 25), cv2.FONT_HERSHEY_PLAIN,
		1, (255, 255, 255), 1)	


	# draw progress bar 
 	ratio_movie_len = camera.get(2)
 	cv2.line(frame,(0,0),(int(frame.shape[1]*ratio_movie_len),0),(255,255,255),4,8)
	cv2.putText(frame, "{0:.{1}f}%".format(ratio_movie_len*100,1),
		(int(frame.shape[1]*ratio_movie_len)-47, 15), cv2.FONT_HERSHEY_PLAIN,
		1, (255, 255, 255), 1)	


	# show the frame to our screen and increment the frame counter
	cv2.imshow("Frame + Trace", frame)
	#cv2.imshow("Cleared Mask",Iclear)


	key = cv2.waitKey(1) & 0xFF
	counter += 1

	# if the 'q' key is pressed, stop the loop
	if key == ord("q"):
		break


# finished
print "Elapsed time: " + str(time_difference) + "s, FPS at the end: " + str(frame_zaehler/time_difference)

# Save screenshot
# draw path 
# loop over the set of tracked points
for i in np.arange(1, len(pts)):
# 	# if either of the tracked points are None, ignore
# 	# them
	if pts[i - 1] is None or pts[i] is None:
		continue
	cv2.line(frame_sicherung, (int(pts[i-1].split(',')[1]),int(pts[i-1].split(',')[2])), 
		(int(pts[i].split(',')[1]),int(pts[i].split(',')[2])), (0,255, 0), 1)

# draw circle at start and end point 
cv2.circle(frame_sicherung, (int(pts[i].split(',')[1]),int(pts[i].split(',')[2])), 3,
			 	(255, 0, 0), 2)
cv2.circle(frame_sicherung, (int(pts[0].split(',')[1]),int(pts[0].split(',')[2])), 3,
			 	(255,255,255), 2)

big = misc.imresize(frame_sicherung, 2.5)
cv2.imwrite("export/" + args.get("video") + "_trace.png",big)



# cleanup the camera and close any open windows
camera.release()
cv2.destroyAllWindows()

# Save pts
with open("export/" + args.get("video") + "_" + date_print + ".csv","w") as saved_pts:
    for line in pts:
        #print line
        coord = line
        saved_pts.write("%s\n" %coord)
        
#close file again
saved_pts.close()