中国 · 南京 · 栖霞区紫东路2号紫东国际创意园B3-2幢5F
+86-18994094214 (仅工作日:8:30~17:30)
contact@oakchina.cn

OAK 使用不同镜头和本地视频流进行模型推理

OAK 使用不同镜头和本地视频流进行模型推理

使用 oak 的 LEFTRIGHTRGB 相机和视频流VIDEO进行 YOLO 检测

▌RGB

使用 RGB 相机作为输入源

...
camRgb = pipeline.create(dai.node.ColorCamera)
detectionNetwork = pipeline.create(dai.node.YoloDetectionNetwork)
...
camRgb.setPreviewSize(W, H)
...
camRgb.preview.link(detectionNetwork.input)
...

详见:yolov6-rgb.py

▌RGB + DEPTH

使用 RGB 相机作为输入源,并附加深度信息

...
camRgb = pipeline.create(dai.node.ColorCamera)
spatialDetectionNetwork = pipeline.create(dai.node.YoloSpatialDetectionNetwork)
monoLeft = pipeline.create(dai.node.MonoCamera)
monoRight = pipeline.create(dai.node.MonoCamera)
stereo = pipeline.create(dai.node.StereoDepth)
...
camRgb.setPreviewSize(W, H)
...
monoLeft.setBoardSocket(dai.CameraBoardSocket.LEFT)
monoRight.setBoardSocket(dai.CameraBoardSocket.RIGHT)
...
# 将深度图与 RGB 相机的视角对齐,在其上进行推理
stereo.setDepthAlign(dai.CameraBoardSocket.RGB)
stereo.setOutputSize(monoLeft.getResolutionWidth(), monoLeft.getResolutionHeight())
...
monoLeft.out.link(stereo.left)
monoRight.out.link(stereo.right)

camRgb.preview.link(spatialDetectionNetwork.input)
stereo.depth.link(spatialDetectionNetwork.inputDepth)
...

详见:yolov6-rgb-spatial.py


▌RIGHT

使用 RIGHT 相机作为输入源

...
monoRight = pipeline.create(dai.node.MonoCamera)
detectionNetwork = pipeline.create(dai.node.YoloDetectionNetwork)
imageManip = pipeline.create(dai.node.ImageManip)
...
monoRight.setBoardSocket(dai.CameraBoardSocket.RIGHT)
...
# NN 模型需要 BGR 输入。默认情况下 ImageManip 输出类型与输入相同(在本例中为灰色)
imageManip.initialConfig.setFrameType(dai.ImgFrame.Type.BGR888p)
imageManip.initialConfig.setResize(W, H)
imageManip.setMaxOutputFrameSize(W * H * 3)
...
monoRight.out.link(imageManip.inputImage)
imageManip.out.link(detectionNetwork.input)
...

详见:yolov6-right.py

▌RIGHT + DEPTH

使用 RIGHT 相机作为输入源,并附加深度信息

...
monoLeft = pipeline.create(dai.node.MonoCamera)
monoRight = pipeline.create(dai.node.MonoCamera)
spatialDetectionNetwork = pipeline.create(dai.node.YoloSpatialDetectionNetwork)
stereo = pipeline.create(dai.node.StereoDepth)
imageManip = pipeline.create(dai.node.ImageManip)
...
monoLeft.setBoardSocket(dai.CameraBoardSocket.LEFT)
monoRight.setBoardSocket(dai.CameraBoardSocket.RIGHT)
...
# NN 模型需要 BGR 输入。默认情况下 ImageManip 输出类型与输入相同(在本例中为灰色)
imageManip.initialConfig.setFrameType(dai.ImgFrame.Type.BGR888p)
imageManip.initialConfig.setResize(W, H)
imageManip.setMaxOutputFrameSize(W * H * 3)
...
# 将深度图与 RIGHT 相机的视角对齐,在其上进行推理
stereo.setDepthAlign(
    dai.RawStereoDepthConfig.AlgorithmControl.DepthAlign.RECTIFIED_RIGHT
)
stereo.setOutputSize(monoLeft.getResolutionWidth(), monoLeft.getResolutionHeight())
...
monoLeft.out.link(stereo.left)
monoRight.out.link(stereo.right)

imageManip.out.link(spatialDetectionNetwork.input)

stereo.rectifiedRight.link(imageManip.inputImage)
stereo.depth.link(spatialDetectionNetwork.inputDepth)
...

详见:yolov6-right-spatial.py


▌LEFT

使用 LEFT 相机作为输入源

...
monoLeft = pipeline.create(dai.node.MonoCamera)
detectionNetwork = pipeline.create(dai.node.YoloDetectionNetwork)
imageManip = pipeline.create(dai.node.ImageManip)
...
monoLeft.setBoardSocket(dai.CameraBoardSocket.LEFT)
...
# NN 模型需要 BGR 输入。默认情况下 ImageManip 输出类型与输入相同(在本例中为灰色)
imageManip.initialConfig.setFrameType(dai.ImgFrame.Type.BGR888p)
imageManip.initialConfig.setResize(W, H)
imageManip.setMaxOutputFrameSize(W * H * 3)
...
monoLeft.out.link(imageManip.inputImage)
imageManip.out.link(detectionNetwork.input)
...

详见:yolov6-left.py

▌LEFT + DEPTH

使用 LEFT 相机作为输入源,并附加深度信息

...
monoLeft = pipeline.create(dai.node.MonoCamera)
monoRight = pipeline.create(dai.node.MonoCamera)
spatialDetectionNetwork = pipeline.create(dai.node.YoloSpatialDetectionNetwork)
stereo = pipeline.create(dai.node.StereoDepth)
imageManip = pipeline.create(dai.node.ImageManip)
...
monoLeft.setBoardSocket(dai.CameraBoardSocket.LEFT)
monoRight.setBoardSocket(dai.CameraBoardSocket.RIGHT)
...
# NN 模型需要 BGR 输入。默认情况下 ImageManip 输出类型与输入相同(在本例中为灰色)
imageManip.initialConfig.setFrameType(dai.ImgFrame.Type.BGR888p)
imageManip.initialConfig.setResize(W, H)
imageManip.setMaxOutputFrameSize(W * H * 3)
...
# 将深度图与 LEFT 相机的视角对齐,在其上进行推理
stereo.setDepthAlign(
    dai.RawStereoDepthConfig.AlgorithmControl.DepthAlign.RECTIFIED_LEFT
)
stereo.setOutputSize(monoLeft.getResolutionWidth(), monoLeft.getResolutionHeight())
...
monoLeft.out.link(stereo.left)
monoRight.out.link(stereo.right)
​
imageManip.out.link(spatialDetectionNetwork.input)
​
stereo.rectifiedLeft.link(imageManip.inputImage)
stereo.depth.link(spatialDetectionNetwork.inputDepth)
...

详见:yolov6-left-spatial.py

▌VIDEO

使用 VIDEO 作为输入源

...
xinFrame = pipeline.create(dai.node.XLinkIn)
detectionNetwork = pipeline.create(dai.node.YoloDetectionNetwork)
...
xinFrame.setStreamName("inFrame")
...
xinFrame.out.link(detectionNetwork.input)
...
# 输入队列将用于将视频帧发送到设备。
inFrameQueue = device.getInputQueue(name="inFrame")
...
img = dai.ImgFrame()
img.setData(to_planar(frame, (W, H)))
img.setTimestamp(monotonic())
img.setWidth(W)
img.setHeight(H)
inFrameQueue.send(img)
...

详见:yolov6-video.py

Tags: ,

索引