目录
一、depthai ROS驱动
一、depthai ROS驱动
（1）驱动下载地址：2. C++ 开发快速上手 — DepthAI Docs 0.3.0.0 documentation
sudo apt install ./depthai_2.17.1_arm64.deb
//运行
Python3 utilities/cam_test.py -mres 400 -cams rgb,m left,m right,m camd,m（2）或者运行python脚本打开相机
import depthai as dai
import cv2
cam_list = ["CAM_B", "CAM_C", "CAM_D","CAM_A"]
cam_socket_opts = {
"CAM_A": dai.CameraBoardSocket.RGB,
"CAM_B": dai.CameraBoardSocket.LEFT,
"CAM_C": dai.CameraBoardSocket.RIGHT,
"CAM_D": dai.CameraBoardSocket.CAM_D,
}
pipeline = dai.Pipeline()
cam = {}
xout = {}
for c in cam_list:
cam[c] = pipeline.create(dai.node.MonoCamera)
cam[c].setResolution(dai.MonoCameraProperties.SensorResolution.THE_800_P)
if c == "CAM_A":
cam[c].initialControl.setFrameSyncMode(dai.CameraControl.FrameSyncMode.OUTPUT)
else:
cam[c].initialControl.setFrameSyncMode(dai.CameraControl.FrameSyncMode.INPUT)
# cam[c].initialControl.setExternalTrigger(4, 3)
cam[c].setBoardSocket(cam_socket_opts[c])
cam[c].setFps(20)
xout[c] = pipeline.create(dai.node.XLinkOut)
xout[c].setStreamName(c)
cam[c].out.link(xout[c].input)
config = dai.Device.Config()
config.board.gpio[6] = dai.BoardConfig.GPIO(
dai.BoardConfig.GPIO.OUTPUT, dai.BoardConfig.GPIO.Level.HIGH
)
# config.version = dai.OpenVINO.VERSION_UNIVERSAL
with dai.Device(config) as device:
device.startPipeline(pipeline)
print('Connected cameras:')
for p in device.getConnectedCameraFeatures():
print(f' -socket {p.socket.name:6}: {p.sensorName:6} {p.width:4} x {p.height:4} focus:', end='')
print('auto ' if p.hasAutofocus else 'fixed', '- ', end='')
print(*[type.name for type in p.supportedTypes])
q = {}
for c in cam_list:
q[c] = device.getOutputQueue(name=c, maxSize=1, blocking=False)
cv2.namedWindow(c, cv2.WINDOW_NORMAL)
cv2.resizeWindow(c, (640, 480))
while not device.isClosed():git clone https://github.com/ethz-asl/kalibr.git
frame_list = []
for c in cam_list:
pkt = q[c].tryGet()
if pkt is not None:
print(c + ":", pkt.getTimestampDevice())
frame_list.append((c, pkt.getCvFrame()))
# cv2.imshow(c, frame)
if frame_list:
print("-------------------------------")
for c,frame in frame_list:
cv2.imshow(c, frame)
#print("-------------------------------")
key = cv2.waitKey(1)
if key == ord("q"):
break
（3）使用python脚本
'''
func:驱动DepthAi套件的相机
by:2024.7.5
'''
import depthai as dai
import cv2
import queue
class CamerDrive():
cam_list = ['rgb', 'left', 'right', 'camd']
cam_socket_opts = {}
pipeline = dai.Pipeline()
# 初始化两个字典，用于存储摄像头节点和XLinkOut节点
cam = {}
xout = {}
#
frameQueue = {}
device = None
# myFrameQueue = {}
# 构造函数
def __init__(self):
self.cam_socket_opts = {
'rgb': dai.CameraBoardSocket.RGB, # Or CAM_A 板载接口名称可复用
'left': dai.CameraBoardSocket.LEFT, # Or CAM_B
'right': dai.CameraBoardSocket.RIGHT, # Or CAM_C
'camd': dai.CameraBoardSocket.CAM_D,
}
# 配置相机节点
def cameraNodeConfig(self):
for c in self.cam_list:
# 创建MonoCamera节点（注意：这里可能对于RGB摄像头需要改为ColorCamera）
self.cam[c] = self.pipeline.create(dai.node.MonoCamera)
# if c == 'rgb':
# self.cam[c] = self.pipeline.create(dai.node.ColorCamera)
# 设置摄像头分辨率为800P
self.cam[c].setResolution(dai.MonoCameraProperties.SensorResolution.THE_800_P)
# 根据摄像头类型设置帧同步模式
if c == 'rgb':
self.cam[c].initialControl.setFrameSyncMode(dai.CameraControl.FrameSyncMode.OUTPUT) # 输出同步模式
else:
self.cam[c].initialControl.setFrameSyncMode(dai.CameraControl.FrameSyncMode.INPUT) # 输入同步模式
# 设置摄像头连接到板载的哪个接口
self.cam[c].setBoardSocket(self.cam_socket_opts[c])
# 创建XLinkOut节点，用于将摄像头数据输出到主机
self.xout[c] = self.pipeline.create(dai.node.XLinkOut)
# 设置输出流的名称
self.xout[c].setStreamName(c)
# 将摄像头的输出链接到XLinkOut的输入
self.cam[c].out.link(self.xout[c].input)
def depthAIConfig(self):
config = dai.Device.Config()
# 将GPIO引脚6配置为输出模式，并设置初始电平为高
config.board.gpio[6] = dai.BoardConfig.GPIO(dai.BoardConfig.GPIO.OUTPUT,
dai.BoardConfig.GPIO.Level.HIGH)
# 使用Device类创建一个设备实例，并传入之前配置的配置对象。同时，使用with语句确保设备在使用完毕后正确关闭。
# with dai.Device(config) as device:
self.device = dai.Device(config)
# 启动pipeline对象
self.device.startPipeline(self.pipeline)
# 遍历相机列表，为每个相机设置输出队列、OpenCV窗口，并初始化FPS计算对象。
for c in self.cam_list:
self.frameQueue[c] = self.device.getOutputQueue(name=c, maxSize=1, blocking=False) # 获取相机输出队列，非阻塞模式
cv2.namedWindow(c, cv2.WINDOW_NORMAL) # 为每个相机创建一个OpenCV窗口
cv2.resizeWindow(c, (640, 480)) # 调整窗口大小
def get_Frame(self):
while True:
for c in self.cam_list:
# 尝试从输出队列中获取一帧数据。
try:
pkt = self.frameQueue[c].tryGet()
if pkt is not None:
# 将数据包转换为OpenCV的Mat对象并显示。
frame = pkt.getCvFrame()
cv2.imshow(c, frame)
else:
print('sda')
key = cv2.waitKey(1)
if key == ord('q'):
break
except:
pass
if __name__ == '__main__':
a = CamerDrive()
a.cameraNodeConfig()
a.depthAIConfig()
a.get_Frame()二、获取相机内参
1.安装depthai-ros
mkdir -p dai_ws/src
cd dai_ws/src
git clone https://gitee.com/oakchina/depthai-ros.git
cd ..
source /opt/ros/melodic/setup.bash
rosdep install --from-paths src --ignore-src -r -y
catkin_make
source devel/setup.bash2.执行相机对应的launch文件
cd dai_ws
source devel/setup.bash
roslaunch depthai_examples mobile_publisher.launch camera_model:=OAK-D3.查看相机内参
rostopic echo -n 1 /mobilenet_publisher/color/camera_info4.相机内参结果
| 相机内参Intrinsics: k表示等距畸变系数，mu,mv对应焦距,v,u对应主点坐标。 相机外参Extrinsics(从右目往左目): rotation表示旋转矩阵， translation表示平移矩阵。 D、K、R、P 分别为畸变参数，内参矩阵，矫正矩阵，投影矩阵。 seq: 0 stamp: secs: 1719912958 nsecs: 823117551 frame_id: "oak_rgb_camera_optical_frame" height: 720 width: 1280 distortion_model: "plumb_bob" 畸变参数D: [-1.87286, 16.683033, 0.001053, -0.002063, 61.878521, -2.158907, 18.424637, 57.682858, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0] 内参矩阵 K: [1479.458984, 0.0, 950.694458, 0.0, 1477.587158, 530.697632, 0.0, 0.0, 1.0] R: [0.9997062937394856, -0.023389961528659683, 0.006343181957754751, 0.0233922879452606, 0.999726320114537, -0.0002928052810344399, -0.006334597252184302, 0.0004411008211352551, 0.9999798389506251] P: [1479.458984, 0.0, 950.694458, 0.0, 0.0, 1477.587158, 530.697632, 0.0, 0.0, 0.0, 1.0, 0.0] binning_x: 0 binning_y: 0 roi: x_offset: 0 y_offset: 0 height: 0 width: 0 do_rectify: False --- |
三、OAK相机外参
git clone https://gitee.com/oakchina/depthai.git
python3 install_requirements.py
//运行
python3 calibrate.py -s 2.75 -ms 2.05 -brd OAK-D -cd 0 -c 1 -mst 1-db：表示默认板，表示你正在使用 Charuco 标记。
-
-nx：x 方向上的 Charuco 标记数。
-
-c：每次显示多边形时拍摄的照片数量（可选，建议在你的情况下省略）。
-
-cd：拍摄照片前的倒计时时间（以秒为单位）（可选，建议用于更快的图像校准）。
-
-s：Charuco 标记周围的正方形大小（以厘米为单位）。
-
-ms：标记的大小（以厘米为单位）。
-
-brd：设备的板（在本例中为 OAK-D-SR-POE）
修改其json文件
{
"board_config":
{
"name": "OAK-D",
"revision": "R1M0E1",
"cameras":{
"CAM_A": {
"name": "rgb",
"hfov": 89.5水平视场角
"type": "color"相机类型
},
"CAM_B": {
"name": "left",
"hfov": 71.86,
"type": "mono",
"extrinsics": {
"to_cam": "CAM_C",
"specTranslation": {
"x": -7.5,
"y": 0,
"z": 0
},
"rotation":{
"r": 0,
"p": 0,
"y": 0
}
}
},
"CAM_C": {
"name": "right",
"hfov": 71.86,
"type": "mono",
"extrinsics": {
"to_cam": "CAM_A",
"specTranslation": {
"x": 3.75,
"y": 0,
"z": 0
},
"rotation":{
"r": 0,
"p": 0,
"y": 0
}
}
}
},
"stereo_config":{
"left_cam": "CAM_B",
"right_cam": "CAM_C"
}
}
}标定结果
