-
Notifications
You must be signed in to change notification settings - Fork 0
/
utils.py
161 lines (136 loc) · 4.58 KB
/
utils.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
157
158
159
160
161
try:
import qi
from naoqi import ALProxy
except:
print('Not on real Robot')
import argparse
import sys
import time
from PIL import Image
import numpy as np
import cv2 as cv
import copy
import ffmpeg
import math
from PIL import ImageFont, ImageDraw, Image
import zmq
import json, ast
context = zmq.Context()
socket = context.socket(zmq.REQ)
socket.connect("tcp://localhost:5555")
def send_array(socket, A, flags=0, copy=True, track=False):
"""send a numpy array with metadata"""
md = dict(
dtype = str(A.dtype),
shape = A.shape,
)
socket.send_json(md, flags|zmq.SNDMORE)
return socket.send(A, flags, copy=copy, track=track)
def recv_array(socket, flags=0, copy=True, track=False):
"""recv a numpy array"""
md = socket.recv_json(flags=flags)
msg = socket.recv(flags=flags, copy=copy, track=track)
#import ipdb;ipdb.set_trace()
#buf = memoryview(msg)
#something wrong with md key name -> u'key1'
A = np.frombuffer(msg, dtype=np.float32)
return A.reshape(4)
def detect_keypoint_yolo(frame):
global socket
send_array(socket, frame)
keypoint= recv_array(socket)
if keypoint[0]==-100:
return frame, None, None
confidence = keypoint
center = copy.deepcopy(keypoint[:2])
center[1]=center[1] * frame.shape[0]
center[0]=center[0] * frame.shape[1]
#print(center)
r = int(abs(keypoint[-1])*frame.shape[0])
return frame,center.astype(np.int), r
def detect_keypoint(frame):
frame = cv.cvtColor(frame, cv.COLOR_RGB2BGR)
l_red_lower = np.array([2,150,71])
l_red_upper = np.array([23,255,1715])
#frame = cv.cvtColor(frame, cv.COLOR_BGR2RGB)
#import ipdb;ipdb.set_trace()
#u_red_lower = np.array([0,170,120])
#u_red_upper = np.array([35,255,245])
erosion_size=1
erosion_shape=cv.MORPH_RECT
kernel_size = 3
kernel = np.ones((kernel_size, kernel_size)).astype(np.uint8)
floodflags = 8
floodflags |= cv.FLOODFILL_FIXED_RANGE
previous_track = None
blurred = cv.GaussianBlur(frame, (3,3), 0)
img_hsv = cv.cvtColor(blurred, cv.COLOR_BGR2HSV)
img_gray = cv.cvtColor(img_hsv, cv.COLOR_BGR2GRAY)
#img_gray = cv.GaussianBlur(img_gray, (kernel_size, kernel_size), 0)
ret,thresh1 = cv.threshold(img_gray,131,255,cv.THRESH_BINARY)
#green_mask = cv.inRange(img_hsv, low_green, high_green)
#green_mask = 255-green_mask
mask=cv.inRange(img_hsv, l_red_lower, l_red_upper)
#if i==9:
#import ipdb;ipdb.set_trace()
#mask_u=cv.inRange(img_hsv, u_red_lower, u_red_upper)
#mask = cv.Canny(blurred,100,200)#mask_l #+ mask_u
tmp = mask.copy()
#mask = cv.erode(mask.astype(np.uint8), kernel, iterations=2)
mask_ = np.zeros((frame.shape[0], frame.shape[1]),np.uint8)
contour = []
contours,hierarchy = cv.findContours(mask, cv.RETR_TREE, cv.CHAIN_APPROX_SIMPLE)
radius=0
if len(contours)>0:
contours = contours[0]
#print('contour',contour)
#print('ctr', contours)
(x,y), radius = cv.minEnclosingCircle(contours)
center = (int(x), int(y))
radius = int(radius)
else:
center=None
radius = None
cv.imwrite('test.jpg', img_gray)
cv.imwrite('mask.jpg', mask)
return frame, center, radius
def load_imgs(int1=1, int2=100, parent_folder='./datasets/rgb', name='farbe'):
loaded = []
for i in range(int1, int2):
#import ipdb;ipdb.set_trace()
loaded.append(cv.imread(parent_folder+'/'+name+str(i)+'.jpg'))
return loaded
def convert_centers2out(c, tmp_c):
#see where at which pixel the ball will roll out
#d = c-tmp_c -> movement vector
#scalar where the movement vecot hits image border S= (image_height - C.x)/d.x
#scale d.y and add to current ball location C.y + d.y *scale
dy = c[0]-tmp_c[0]
dx = c[1]-tmp_c[1]
#x is down
if dx==0:
dx+=1e-4
scalar = (600-c[1])/( dx)
print(scalar)
if scalar<0:
scalar=1e9
out_pixel = c[0]+(dy * scalar)
#print('s',scalar,' cutofy', cut_of_y,'d ', (dx,dy), ' C', c )
d = np.array([dx,dy])
d = d / np.linalg.norm(d)
angle = np.arctan2(d[1],d[0])
angle = angle*180/math.pi
left_right = angle
norm = np.sqrt((dx)**2 + (dy)**2)
return out_pixel, angle, dx, dy, norm
def output_pixel2decision(cut_of_y, r):
#cut_of_y = output pixel
#r radius
angle_ = '------'
if cut_of_y>=320 and cut_of_y<=470:
angle_='middle'
elif cut_of_y>470 and cut_of_y<2000:
angle_='right'
elif cut_of_y<320 and cut_of_y>-2000:
angle_='left'
return angle_