# 使用 LiteRT 进行目标检测和编码 Source: [https://docs.qualcomm.com/doc/80-70018-50SC/topic/single-camera-stream-with-object-detection-and-encode.html](https://docs.qualcomm.com/doc/80-70018-50SC/topic/single-camera-stream-with-object-detection-and-encode.html) 该用例使用 YOLOv5 LiteRT 模型来识别场景中的对象。该用例在检测到的对象上叠加或合成边界框，然后将此流编码为 H.264 码流。 ## 使用 qtivoverlay 插件应用边框叠加 运行用例： gst-launch-1.0 -e qtiqmmfsrc name=camsrc ! video/x-raw,format=NV12_Q08C,width=1280,height=720,framerate=30/1 ! queue ! tee name=split split. ! \ queue ! qtimetamux name=metamux ! queue ! qtivoverlay ! queue ! v4l2h264enc capture-io-mode=4 output-io-mode=5 ! h264parse ! queue ! mp4mux ! \ queue ! filesink location=/etc/media/video.mp4 split. ! queue ! qtimlvconverter ! queue ! qtimltflite delegate=external \ external-delegate-path=libQnnTFLiteDelegate.so external-delegate-options="QNNExternalDelegate,backend_type=htp;" model=/etc/models/yolov5.tflite ! queue ! \ qtimlvdetection threshold=75.0 results=10 module=yolov5 labels=/etc/labels/yolov5.labels \ constants="YoloV5,q-offsets=<3.0>,q-scales=<0.005047998391091824>;" ! text/x-raw ! queue ! metamux.Copy to clipboard 要停止用例，请按 CTRL + C。 Figure : 边框叠加和编码的 pipeline ![](data:image/png;base64,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) 下图显示了用例执行流程： 1. 从摄像头源传来的视频流中识别目标对象场景。 2. 使用 overlaylib 将边框叠加在检测到的目标对象上。 3. 将数据流编码为 H.264 码流。 4. 多路复用 MP4 容器中的数据流并将其存储为 MP4 文件。 pipeline 执行的顺序处理阶段如下表所示： | 处理过程 | 说明 | | --- | --- | | [qtiqmmfsrc](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtiqmmfsrc.html) |

  1. 采集视频流（源）并创建源的两个副本：


| | **预处理** | **预处理** | | [qtimlvconverter](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtimlvconverter.html) |

  1. 在其接收端口上接收视频流。


  2. 执行预处理：


  3. 在其发送端口上将视频流转换为张量数据。

    目标检测模型使用此张量数据进行推理。




| | **推理** | **推理** | | [qtimltflite](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtimltflite.html) |

  1. 加载目标检测模型。


  2. 为选择的 delegate 修改图。


  3. 在其接收端口上接收张量数据。


  4. 运行推理并在其发送端口上生成带有目标检测结果的张量数据。


| | **后处理** | **后处理** | | [qtimlvdetection](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtimlvdetection.html) |

  1. 接收来自目标检测模型的推理张量。


  2. 将接收端口上的推理张量转换为视频或文本等格式，稍后可由多媒体插件进行处理。


  3. 将阈值应用于所选的结果数。


  4. 加载检测模型的相应模块。

    在此用例中，qtimlvdetection 执行以下操作：


    1. 加载 YOLOv5 子模块。


    2. 将结果生成为文本结构。


    3. 接着发送到 qtimetamux 的接收端口。





| | [qtimetamux](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtimetamux.html) |

  1. 在接收端口上接收视频流和文本流，以及与视频流相对应的边框结果。


  2. 使用接收端口中的视频流内容生成 GST 缓存。


  3. 将边框作为 GstVideoRegionOfInterest 从数据接收端添加到其发送端口上的 GST 缓存元数据（元复用）。


| | [qtivoverlay](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtioverlay.html) |

  1. 接收多路复用流。


  2. 使用 CL 将边框叠加在 VideoFrame 上。


  3. 在其发送端口上生成带有叠加层的 GST 缓存。


| | [v4l2h264enc](https://docs.qualcomm.com/doc/80-70018-50SC/topic/v4l2h264enc.html) |

  1. 将参数应用于在接收端口上接收到的视频流的每一帧。


  2. 将其编码为码流，并通过其发送端口发送。


| | h264parse | 向 GStreamer 缓存元数据添加更多码流信息。 | | mp4mux | 接收这些缓存并创建具有格式规范缓存的容器。 | | **输出** | **输出** | | Filesink | 将生成的数据流存储在 /opt/video.mp4 文件中。 | | 播放 | 从主机拉取 video.mp4 并在媒体播放器上播放：
`scp root@ target device>:/opt/ directory>` | ## 使用 qtivcomposer 混合原始帧与边框掩码 运行用例： gst-launch-1.0 -e \ qtiqmmfsrc name=camsrc ! video/x-raw,format=NV12_Q08C,width=1280,height=720,framerate=30/1 ! queue ! tee name=split \ split. ! queue ! qtivcomposer name=mixer ! queue ! video/x-raw,format=NV12,width=1920,height=1080,interlace-mode=progressive,colorimetry=bt601 ! \ v4l2h264enc capture-io-mode=4 output-io-mode=5 ! h264parse ! queue ! mp4mux ! queue ! filesink location=/etc/media/video.mp4 \ split. ! queue ! qtimlvconverter ! queue ! qtimltflite delegate=external external-delegate-path=libQnnTFLiteDelegate.so \ \external-delegate-options="QNNExternalDelegate,backend_type=htp;" model=/etc/models/yolov5.tflite ! queue ! \ qtimlvdetection threshold=75.0 results=10 module=yolov5 labels=/etc/labels/yolov5.labels \ constants="YoloV5,q-offsets=<3.0>,q-scales=<0.005047998391091824>;" ! video/x-raw,format=BGRA,width=640,height=360 ! queue ! mixer.Copy to clipboard 要停止用例，请按 CTRL + C。 Figure : 使用 qtivcomposer 进行边框掩码和编码的 pipeline ![](data:image/png;base64,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) 下图显示了用例执行流程： 1. 从摄像头源传来的视频流中识别目标对象场景。 2. 使用 qtivcomposer，在检测到的对象和原始视频流上组合边界框。 3. 将此流编码为 H.264 码流。 4. 多路复用 MP4 容器中的数据流并将其存储为 MP4 文件。 pipeline 执行的顺序处理阶段如下表所示： | 处理过程 | 说明 | | --- | --- | | [qtiqmmfsrc](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtiqmmfsrc.html) |

  1. 采集视频流（源）并创建源的两个副本：

    • 一个视频流被发送到 qtimetamux 插件以保留视频流。


    • 另一个视频流被发送到 ML 推理 pipeline。





| | **预处理** | **预处理** | | [qtimlvconverter](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtimlvconverter.html) |

  1. 在其接收端口上接收视频流。


  2. 执行预处理：

    • 颜色转换


    • 缩小/放大


    • 在模型需要浮点值作为输入时，对流数据进行归一化





  3. 在其发送端口上将视频流转换为张量数据。

    目标检测模型使用此张量数据进行推理。




| | **推理** | **推理** | | [qtimltflite](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtimltflite.html) |

  1. 加载目标检测模型。


  2. 为选择的 delegate 修改图。


  3. 在其接收端口上接收张量数据。


  4. 运行推理并在其发送端口上生成带有目标检测结果的张量数据。


| | **后处理** | **后处理** | | [qtimlvdetection](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtimlvdetection.html) |

  1. 接收来自目标检测模型的推理张量。


  2. 将接收端口上的推理张量转换为视频或文本等格式，稍后可由多媒体插件进行处理。


  3. 将阈值应用于所选的结果数。


  4. 加载检测模型的相应模块。

    在此用例中，qtimlvdetection 执行以下操作：


    1. 加载 YOLOv5 子模块。


    2. 生成包含可叠加在目标对象上的边框的视频帧。


    3. 将它们发送至 qtivcomposer 的接收端口。





| | [qtivcomposer](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtivcomposer.html) |

  1. 在其接收端上接收原始视频流和带有边框的视频流


  2. 在其发送端口上，生成在其接收端口处理的视频流所组成的内容。


| | [v4l2h264enc](https://docs.qualcomm.com/doc/80-70018-50SC/topic/v4l2h264enc.html) |

  1. 将参数应用于在接收端口上接收到的视频流的每一帧。


  2. 将其编码为码流，并通过其发送端口发送。


| | h264parse | 向 GStreamer 缓存元数据添加更多码流信息。 | | mp4mux | 接收这些缓存并创建具有格式规范缓存的容器。 | | **输出** | **输出** | | Filesink | 将生成的数据流存储在 /etc/media/video.mp4 文件中。 | | Playback | 从主机拉取 video.mp4 并在媒体播放器上播放：
`scp root@ target device>:/etc/ directory>` | **Parent Topic:** [LiteRT 用例](https://docs.qualcomm.com/doc/80-70018-50SC/topic/tensorflow-lite-use-cases.html) Last Published: Nov 12, 2025 [Previous Topic 使用 LiteRT 进行目标检测和显示](https://docs.qualcomm.com/bundle/publicresource/80-70018-50SC/topics/single-camera-stream-with-object-detection-and-display.md) [Next Topic 使用 LiteRT 进行图像分割和显示](https://docs.qualcomm.com/bundle/publicresource/80-70018-50SC/topics/single-camera-stream-with-image-segmentation-and-display.md)