-
Notifications
You must be signed in to change notification settings - Fork 0
/
realsense_bytestream.py
263 lines (223 loc) · 9.05 KB
/
realsense_bytestream.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
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
## License: Apache 2.0. See LICENSE file in root directory.
## Copyright(c) 2015-2017 Intel Corporation. All Rights Reserved.
###############################################
## Open CV and Numpy integration ##
###############################################
# From Connor
# - line 205 & 206 define a "border" that gets cropped off. Currently, 128px are chopped off the top and bottom of the image, and 96 are chopped off from the sides. You can reduce that if you are closer to your camera than I am (although I think Realsense suggest a minimum of 1m -- this could also explain some of the detail cutoff you see in the above image if you are close to it).
# - line 237 calls a function called process_frame. The last parameter is the maximum number of points it will take. If there are more than that many available points to send, it randomly samples to get down to the parameter. It's currently 10k.
# - line 131 has a "maxVal" param that is the maximum distance (in mm) from the camera, anything further than that will get discarded to get rid of background noise. Currently set to 2000 = 2m.
import pyrealsense2 as rs
import numpy as np
import cv2
import sys
import socket
import pickle
import json
import struct
import threading
import time
import random
from math import ceil, floor
IP = '127.0.0.1'
# IP = "128.122.215.23"
portTCP = 20010
#PORT = 20011
# toggle this to either stream through Corelink or write to a file.
WRITE_TO_FILE = False
username = "Testuser"
password = "Testpassword"
BUFFERSIZE = 1024
try:
s = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
except socket.error as e:
print("Error creating socket: %s" % e)
sys.exit(1)
try:
s.connect((IP, portTCP))
except socket.gaierror as e:
print("Address-related error connecting to server: %s" % e)
sys.exit(1)
except socket.error as e:
print("Connection error: %s" % e)
sys.exit(1)
try:
s.sendall(('{"function":"auth","username":"'+username+'","password":"'+password+'"}').encode())
except socket.error as e:
print("Error sending data %s" % e)
sys.exit(1)
try:
data = s.recv(BUFFERSIZE)
token = json.loads(data.decode("utf-8"))['token']
sourceIP = json.loads(data.decode("utf-8"))['IP']
s.send(('{"function":"sender","workspace":"Chalktalk","proto":"tcp","IP":"'+str(sourceIP)+'","port":0,"type":"sync","token":"'+token+'"}').encode())
data = s.recv(BUFFERSIZE)
streamID = json.loads(data.decode("utf-8"))['streamID']
port = json.loads(data.decode("utf-8"))['port']
except socket.error as e:
print("Error receiving data: %s" % e)
sys.exit(1)
tcp = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
tcp.connect((IP, port))
time.sleep(3)
def sendMessage(message, pacNum, totalPac, frame):
pac = str(pacNum)+'/'+str(totalPac)
header = {
"packet" : pac,
"frame" : frame
}
header = json.dumps(header).encode("utf-8")
header_size = len(header)
header_size = struct.pack('<H', header_size)
data_size = struct.pack('<H', len(message))
streamID_send = struct.pack('<L', streamID)
tcp.sendall(header_size+data_size+streamID_send+header+message)
# tcp.sendall(message)
#print(f"Sent: {header_size+data_size+streamID_send+header}, Length: {len(header_size+data_size+streamID_send+header+message)}")
'''
def sendString():
try:
i=1
while i < 16:
sendMessage(bytes(str(i).encode("utf-8")), i)
i=i+1
except KeyboardInterrupt:
print("interrupt received")
while True:
t = threading.Thread(target = sendString)
t.start()
time.sleep(1)
'''
def rescale(frame, percent=75):
width = int(frame.shape[1] * percent/ 100)
height = int(frame.shape[0] * percent/ 100)
dim = (width, height)
return cv2.resize(frame, dim, interpolation =cv2.INTER_AREA)
def sendVideo(data, frame):
bytes_sent = 0
packet = 1
bytes_per_packet = 57642
totalPac = ceil(len(data) / bytes_per_packet)
while bytes_sent < len(data):
start = bytes_sent
bytes_sent += bytes_per_packet
if bytes_sent > len(data):
bytes_sent = len(data)
sendMessage(data[start:bytes_sent], packet, totalPac, frame)
packet += 1
# time.sleep(1) # debugging purposes
sys.exit(1)
def process_frame(depth_image, color_image, maxPoints):
height = depth_image.shape[0]
width = depth_image.shape[1]
# max depth ~2m
maxVal = 2000
arr = []
flatdepth = depth_image.flatten()
vals = np.where(flatdepth < maxVal, flatdepth, 0)
# remove all zero-depth pixels (shadows, artifacts, etc.)
# then, convert this into indices of a 1D sparse vector
indices = vals.nonzero()[0]
d = depth_image.flatten()[indices]
f_x = lambda x: x % width
f_y = lambda x: height - floor(x / width)
x = np.array([f_x(idx) for idx in indices])
y = np.array([f_y(idx) for idx in indices])
r = (color_image[:,:,2].flatten())[indices - 20]
g = (color_image[:,:,1].flatten())[indices - 20]
b = (color_image[:,:,0].flatten())[indices - 20]
frame = np.matrix([x, y, d, r, g, b], dtype='<u2')
frame = np.transpose(frame)
# if there are more than this many valid points,
# then we randomly sample to not send/write too much data
if len(frame) > maxPoints:
frame = frame[np.random.choice(frame.shape[0], maxPoints, replace=False), :]
frame = frame.flatten()
return frame[0].tobytes()
# cap = cv2.VideoCapture(0)
# cap.set(3, 640)
# cap.set(4, 480)
# while cv2.waitKey(1) & 0xFF != ord('q'):
# success, img = cap.read()
# cv2.imshow("Sender", img)
# data = pickle.dumps(rescale(img, 50))
# t = threading.Thread(target = sendVideo, args = [data])
# t.start()
# # time.sleep(5) # debugging purposes
# cap.release()
# cv2.destroyAllWindows()
# Configure depth and color streams
pipeline = rs.pipeline()
config = rs.config()
# Get device product line for setting a supporting resolution
pipeline_wrapper = rs.pipeline_wrapper(pipeline)
pipeline_profile = config.resolve(pipeline_wrapper)
device = pipeline_profile.get_device()
device_product_line = str(device.get_info(rs.camera_info.product_line))
config.enable_stream(rs.stream.depth, 640, 480, rs.format.z16, 30)
if device_product_line == 'L500':
config.enable_stream(rs.stream.color, 960, 540, rs.format.bgr8, 30)
else:
config.enable_stream(rs.stream.color, 640, 480, rs.format.bgr8, 30)
# Start streaming
pipeline.start(config)
color_images = []
depth_images = []
frameNum = 0
try:
while True:
print("frame",frameNum,"\n")
# Wait for a coherent pair of frames: depth and color
frames = pipeline.wait_for_frames()
depth_frame = frames.get_depth_frame()
color_frame = frames.get_color_frame()
if not depth_frame or not color_frame:
continue
hborder = 128
vborder = 96
# Convert images to numpy arrays
depth_image = np.asanyarray(depth_frame.get_data())[vborder:-vborder,hborder:-hborder]
color_image = np.asanyarray(color_frame.get_data())[vborder:-vborder,hborder:-hborder,:]
# Apply colormap on depth image (image must be converted to 8-bit per pixel first)
depth_colormap = cv2.applyColorMap(cv2.convertScaleAbs(depth_image, alpha=0.03), cv2.COLORMAP_JET)
depth_colormap_dim = depth_colormap.shape
color_colormap_dim = color_image.shape
resized_color_image = color_image
# If depth and color resolutions are different, resize color image to match depth image for display
if depth_colormap_dim != color_colormap_dim:
resized_color_image = cv2.resize(color_image, dsize=(depth_colormap_dim[1], depth_colormap_dim[0]), interpolation=cv2.INTER_AREA)
images = np.hstack((resized_color_image, depth_colormap))
else:
images = np.hstack((color_image, depth_colormap))
# sendMessage("testtest", 1)
if WRITE_TO_FILE:
color_images.append(color_image.tolist())
depth_images.append(depth_image.tolist())
else:
# data = pickle.dumps(rescale(depth_frame, 50))
# data = {}
# data["type"] = "realsense"
# data["username"] = username
# data["frame"] = process_frame(depth_image, color_image, 10000)
# dataJson = json.dumps(data).encode("utf-8")
#t = threading.Thread(target = sendVideo, args = [dataJson, frameNum])
data = process_frame(depth_image, color_image, 10000)
t = threading.Thread(target = sendVideo, args = [data, frameNum])
t.start()
# Show images
cv2.namedWindow('RealSense', cv2.WINDOW_AUTOSIZE)
cv2.imshow('RealSense', images)
frameNum += 1
k = cv2.waitKey(1)
if k == 27:
break
finally:
# Stop streaming
pipeline.stop()
if WRITE_TO_FILE:
allframes = []
for i in range(len(color_images)):
allframes.append(process_frame(np.asarray(depth_images[i]), np.asarray(color_images[i]), 10000))
f = open("realsensedata.js", "w")
f.write(json.dumps(allframes))
f.close()