forked from henrikmidtiby/MarkerLocator
/
MarkerTracker.py
156 lines (128 loc) · 6.82 KB
/
MarkerTracker.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
# -*- coding: utf-8 -*-
"""
Marker tracker for locating n-fold edges in images using convolution.
@author: Henrik Skov Midtiby
"""
import cv2
import numpy as np
import math
class MarkerTracker:
'''
Purpose: Locate a certain marker in an image.
'''
def __init__(self, order, kernelSize, scaleFactor):
self.kernelSize = kernelSize
(kernelReal, kernelImag) = self.generateSymmetryDetectorKernel(order, kernelSize)
self.order = order
self.matReal = np.zeros((kernelSize, kernelSize), dtype=np.float32)
self.matImag = np.zeros((kernelSize, kernelSize), dtype=np.float32)
for i in range(kernelSize):
for j in range(kernelSize):
self.matReal[i, j] = kernelReal[i][j] / scaleFactor
self.matImag[i, j] = kernelImag[i][j] / scaleFactor
self.lastMarkerLocation = (None, None)
self.orientation = None
(kernelRealThirdHarmonics, kernelImagThirdHarmonics) = self.generateSymmetryDetectorKernel(3*order, kernelSize)
self.matRealThirdHarmonics = np.zeros((kernelSize, kernelSize), np.float32)
self.matImagThirdHarmonics = np.zeros((kernelSize, kernelSize), np.float32)
for i in range(kernelSize):
for j in range(kernelSize):
self.matRealThirdHarmonics[i, j] = kernelRealThirdHarmonics[i][j] / scaleFactor
self.matImagThirdHarmonics[i, j] = kernelImagThirdHarmonics[i][j] / scaleFactor
self.quality = 0
def generateSymmetryDetectorKernel(self, order, kernelsize):
valueRange = np.linspace(-1, 1, kernelsize);
temp1 = np.meshgrid(valueRange, valueRange)
kernel = temp1[0] + 1j*temp1[1];
magni = abs(kernel);
kernel = kernel**order;
kernel = kernel*np.exp(-8*magni**2);
return (np.real(kernel), np.imag(kernel))
def allocateSpaceGivenFirstFrame(self, frame):
framewidth=frame.shape[1]
frameheight=frame.shape[0]
self.newFrameImage32F = np.zeros((frameheight, framewidth,3), dtype=np.float32)
self.newFrameImage32F = np.zeros((frameheight, framewidth,3), dtype=np.float32)
self.frameReal = np.zeros((frameheight,framewidth,1), dtype=np.float32)
self.frameImag = np.zeros((frameheight,framewidth,1), dtype=np.float32)
self.frameRealThirdHarmonics = np.zeros((frameheight,framewidth,1), dtype=np.float32)
self.frameImagThirdHarmonics = np.zeros((frameheight,framewidth,1), dtype=np.float32)
self.frameRealSq = np.zeros((frameheight,framewidth,1), dtype=np.float32)
self.frameImagSq = np.zeros((frameheight,framewidth,1), dtype=np.float32)
self.frameSumSq = np.zeros((frameheight,framewidth,1), dtype=np.float32)
def locateMarker(self, frame):
self.frameReal = frame
self.frameImag = frame
self.frameRealThirdHarmonics = frame
self.frameImagThirdHarmonics = frame
# Calculate convolution and determine response strength.
self.frameReal = cv2.filter2D(self.frameReal, cv2.CV_32F, self.matReal)
self.frameImag = cv2.filter2D(self.frameImag, cv2.CV_32F, self.matImag)
self.frameRealSq = np.multiply(self.frameReal, self.frameReal)
self.frameImagSq = np.multiply(self.frameImag, self.frameImag)
self.frameSumSq = self.frameRealSq + self.frameImagSq
# Calculate convolution of third harmonics for quality estimation.
self.frameRealThirdHarmonics = cv2.filter2D(self.frameRealThirdHarmonics, cv2.CV_32F, self.matRealThirdHarmonics)
self.frameImagThirdHarmonics = cv2.filter2D(self.frameImagThirdHarmonics, cv2.CV_32F, self.matImagThirdHarmonics)
min_val, max_val, min_loc, max_loc = cv2.minMaxLoc(self.frameSumSq)
self.lastMarkerLocation = max_loc
(xm, ym) = max_loc
self.determineMarkerOrientation(frame)
self.determineMarkerQuality()
return max_loc
def determineMarkerOrientation(self, frame):
(xm, ym) = self.lastMarkerLocation
realval=self.frameReal[ym, xm]
imagval = self.frameImag[ym, xm]
self.orientation = (math.atan2(-realval, imagval) - math.pi / 2) / self.order
maxValue = 0
maxOrient = 0
searchDist = self.kernelSize / 3
for k in range(self.order):
orient = self.orientation + 2 * k * math.pi / self.order
xm2 = int(xm + searchDist*math.cos(orient))
ym2 = int(ym + searchDist*math.sin(orient))
if(xm2 > 0 and ym2 > 0 and xm2 < frame.shape[1] and ym2 < frame.shape[0]):
try:
intensity = frame[ym2,xm2]
if(intensity[0] > maxValue):
maxValue = intensity[0]
maxOrient = orient
except:
print("determineMarkerOrientation: error: %d %d %d %d" % (ym2, xm2, frame.shape[1], frame.shape[0]))
pass
self.orientation = self.limitAngleToRange(maxOrient)
def determineMarkerQuality(self):
(xm, ym) = self.lastMarkerLocation
realval=self.frameReal[ym, xm]
imagval = self.frameImag[ym, xm]
realvalThirdHarmonics = self.frameRealThirdHarmonics[ym, xm]
imagvalThirdHarmonics = self.frameImagThirdHarmonics[ym, xm]
argumentPredicted = 3*math.atan2(-realval, imagval)
argumentThirdHarmonics = math.atan2(-realvalThirdHarmonics, imagvalThirdHarmonics)
argumentPredicted = self.limitAngleToRange(argumentPredicted)
argumentThirdHarmonics = self.limitAngleToRange(argumentThirdHarmonics)
difference = self.limitAngleToRange(argumentPredicted - argumentThirdHarmonics)
strength = math.sqrt(realval*realval + imagval*imagval)
strengthThirdHarmonics = math.sqrt(realvalThirdHarmonics*realvalThirdHarmonics + imagvalThirdHarmonics*imagvalThirdHarmonics)
#print("Arg predicted: %5.2f Arg found: %5.2f Difference: %5.2f" % (argumentPredicted, argumentThirdHarmonics, difference))
#print("angdifferenge: %5.2f strengthRatio: %8.5f" % (difference, strengthThirdHarmonics / strength))
# angdifference \in [-0.2; 0.2]
# strengthRatio \in [0.03; 0.055]
self.quality = math.exp(-math.pow(difference/0.3, 2))
#self.printMarkerQuality(self.quality)
def printMarkerQuality(self, quality):
stars = ""
if(quality > 0.5):
stars = "**"
if(quality > 0.7):
stars = "***"
if(quality > 0.9):
stars = "****"
print("quality = %d): %5.2f %s" % (self.order, quality, stars))
def limitAngleToRange(self, angle):
while(angle < math.pi):
angle += 2*math.pi
while(angle > math.pi):
angle -= 2*math.pi
return angle