# 目标检测和分类 

Source: [https://docs.qualcomm.com/doc/80-70018-50SC/topic/camera-ai-detection-overlay-composer-display.html](https://docs.qualcomm.com/doc/80-70018-50SC/topic/camera-ai-detection-overlay-composer-display.html)

**gst-camera-two-stream-detection-and-classification-side-by-side.py** 应用程序使用 YOLOv8 LiteRT 模型来检测和分类 AI Overlay Composer 显示的场景中的对象。

Figure : 目标检测和分类的 pipeline
            
            ![](data:image/png;base64,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)

有关该 pipeline 中使用的插件的信息，请参见 [Pipeline 流](https://docs.qualcomm.com/doc/80-70018-50SC/topic/camera-ai-detection-overlay-composer-display.html#camera-ai-detection-overlay-composer-display__section_l1h_cpk_bdc)。

## 模型文件

Table : 用于检测和分类的模型

| 目的 | LiteRT模型 | 说明 |
| :--- | :--- | :--- |
| 目标检测 | YOLOv8 | <ol class="ol" id="camera-ai-detection-overlay-composer-display__ul_cfw_r4k_bdc"><br>                                    <li class="li">从摄像头流中识别场景中的对象。</li><br><br>                                    <li class="li">将边界框覆盖在检测到的对象上。</li><br><br>                                </ol> |
| 图像分类 | Resnet101 | <ol class="ol" id="camera-ai-detection-overlay-composer-display__ol_jll_v4k_bdc"><br>                                    <li class="li">对摄像头流中的场景进行分类。</li><br><br>                                    <li class="li">在屏幕上叠加分类标签。</li><br><br>                                </ol> |

## 运行应用程序

1. 确保您已完成[前提条件](https://docs.qualcomm.com/doc/80-70018-50SC/topic/prerequisites-for-python-sample-applications.html)。
2. 在目标设备上运行检测和分类脚本：

        gst-camera-two-stream-detection-and-classification-side-by-side.pyCopy to clipboard

如需显示可用的帮助选项，可运行以下命令：

gst-camera-two-stream-detection-and-classification-side-by-side.py -hCopy to clipboard

Table : 用于目标检测和分类 python 应用程序的模型和标签文件的默认目录

| 模型和标签文件 | 目录 |
| :--- | :--- |
| 检测模型 | /etc/models/YoloV8N\_Detection\_Quantized.tflite |
| 检测标签 | /etc/labels/yolov8n.labels |
| 分类模型 | /etc/models/Resnet101\_Quantized.tflite |
| 分类标签 | /etc/labels/resnet101.labels |

## 预期输出

图像并排显示在显示器上。

## Pipeline 流

| 处理过程 | 说明 |
| --- | --- |
| [qtiqmmfsrc](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtiqmmfsrc.html) | 从摄像头采集两个视频流：<ul class="ul"><br>                                    <li class="li">用于检测的流使用 tee 分割并发送到以下位置：<ul class="ul" id="camera-ai-detection-overlay-composer-display__ul_elf_dtk_bdc"><br>                                            <li class="li">qtimetamux 保留视频流。</li><br><br>                                            <li class="li">qtimlvconverter 将视频流转换为输入张量以进行检测推理。</li><br><br>                                        </ul><br></li><br><br>                                    <li class="li">用于分类的流使用 tee 进行分割并发送到以下位置：<ul class="ul" id="camera-ai-detection-overlay-composer-display__ul_nsj_htk_bdc"><br>                                            <li class="li">qtimetamux 保留视频流。</li><br><br>                                            <li class="li">qtimlvconverter 将视频流转换为输入张量，用于分类推理。</li><br><br>                                        </ul><br><br>                                    </li><br><br>                                </ul> |
| **预处理** | **预处理** |
| [qtimlvconverter](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtimlvconverter.html) | <ol class="ol" id="camera-ai-detection-overlay-composer-display__ol_i5w_4wl_vbc"><br>                                    <li class="li">在其接收端口上接收视频流。</li><br><br>                                    <li class="li">执行预处理：<ul class="ul" id="camera-ai-detection-overlay-composer-display__ol_zdw_qwl_vbc"><br>                                            <li class="li">颜色转换</li><br><br>                                            <li class="li">缩小/放大</li><br><br>                                            <li class="li">当模型期望浮点值作为输入时对流数据进行标准化</li><br><br>                                        </ul><br></li><br><br>                                    <li class="li">在其发送端口上将视频流转换为张量数据。<p class="p">分类模型使用此张量数据进行推理。</p><br></li><br><br>                                </ol> |
| **推理** | **推理** |
| [qtimltflite](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtimltflite.html) | <ol class="ol" id="camera-ai-detection-overlay-composer-display__ol_u1l_cxl_vbc"><br>                                    <li class="li">加载模型。</li><br><br>                                    <li class="li">为选择的 delegate 修改图。</li><br><br>                                    <li class="li">在其接收端口上接收张量数据。</li><br><br>                                    <li class="li">执行推理并在其发送端口上生成包含推理结果的张量数据。</li><br><br>                                </ol> |
| **后处理** | **后处理** |
| [qtimlvdetection](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtimlvdetection.html) | <ol class="ol" id="camera-ai-detection-overlay-composer-display__ol_ky5_grn_vbc"><br>                                    <li class="li"> 接收来自目标检测模型的推理张量。 </li><br><br>                                    <li class="li">将其接收端口上的推理张量转换为多媒体插件稍后可以处理的视频或文本等格式。</li><br><br>                                    <li class="li">将阈值应用于所选的结果数。 </li><br><br>                                    <li class="li">加载检测模型的相应模块。 <p class="p">在此用例中，qtimlvdetection 执行以下操作：<br>                                            </p><ol class="ol" type="a" id="camera-ai-detection-overlay-composer-display__ol_jcd_wnk_5bc"><br>                                            <li class="li">加载 YOLOv8 子模块。 </li><br><br>                                            <li class="li">将结果生成为文本结构。</li><br><br>                                            <li class="li">接着发送到 qtimetamux 的接收端口。</li><br><br>                                        </ol><br></li><br><br>                                </ol> |
| [qtimlvclassification](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtimlvclassification.html) | <ol class="ol" id="camera-ai-detection-overlay-composer-display__ol_o3v_2xl_vbc"><br>                                    <li class="li">从其接收端口上的分类模型接收推理结果。 </li><br><br>                                    <li class="li">将推理张量转换为多媒体插件稍后可以处理的视频或文本等格式。 </li><br><br>                                    <li class="li">将阈值应用于所选的结果数。 </li><br><br>                                    <li class="li">加载分类模型的相应模块。 <p class="p">在此用例中，qtimlvclassification 执行以下操作： </p><ol class="ol" type="a" id="camera-ai-detection-overlay-composer-display__ol_p3v_2xl_vbc"><br>                                            <li class="li">加载模型的子模块。</li><br><br>                                            <li class="li">将结果生成为带有分类标签的视频帧。</li><br><br>                                            <li class="li">将它们发送至 qtivcomposer 的接收端口。</li><br><br>                                        </ol><br></li><br><br>                                </ol> |
| [qtimetamux](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtimetamux.html) | <ol class="ol" id="camera-ai-detection-overlay-composer-display__ol_ll3_x5l_vbc"><br>                                    <li class="li">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。</li><br><br>                                    <li class="li">使用接收端口中的视频流内容生成 GST 缓存。</li><br><br>                                    <li class="li">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）。</li><br><br>                                </ol> |
| [qtivoverlay](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtioverlay.html) | <ol class="ol" id="camera-ai-detection-overlay-composer-display__ol_wst_y5l_vbc"><br>                                    <li class="li">接收多路复用流。</li><br><br>                                    <li class="li">使用 CL 将边框叠加在 VideoFrame 上。</li><br><br>                                    <li class="li">在其发送端口上生成带有叠加层的 GST 缓存。</li><br><br>                                </ol> |
| [qtivcomposer](https://docs.qualcomm.com/doc/80-70018-50SC/topic/qtivcomposer.html) | <ol class="ol" id="camera-ai-detection-overlay-composer-display__ol_nmc_lxl_vbc"><br>                                    <li class="li">在接收端口上接收原始视频流和分类结果。 </li><br><br>                                    <li class="li">在其发送端口上生成 GST 缓存，其内容由来自其接收端的视频流组成。</li><br><br>                                </ol> |
| **输出** | **输出** |
| [Waylandsink](https://docs.qualcomm.com/doc/80-70018-50SC/topic/waylandsink.html) | <ol class="ol" id="camera-ai-detection-overlay-composer-display__ol_cgt_mwl_vbc"><br>                                    <li class="li">在其接收端口上接收视频</li><br><br>                                    <li class="li">将视频流提交到 Weston。 </li><br><br>                                    <li class="li">Weston 在本地显示器设备上呈现视频流。</li><br><br>                                </ol> |

**Parent Topic:** [Python 应用程序](https://docs.qualcomm.com/doc/80-70018-50SC/topic/python-sample-applications.html)

**Related Resources**  

- [目标检测](https://docs.qualcomm.com/doc/80-70018-50SC/topic/gst-ai-object-detection.html)
- [图像分类](https://docs.qualcomm.com/doc/80-70018-50SC/topic/gst-ai-classification.html)

Last Published: Nov 12, 2025

[Previous Topic
使用 Python 解码 JPEG 图像](https://docs.qualcomm.com/bundle/publicresource/80-70018-50SC/topics/decode-jpeg-images-using-python.md) [Next Topic
转换和编码摄像头流](https://docs.qualcomm.com/bundle/publicresource/80-70018-50SC/topics/camera-transform-downscale-and-rotate-encode.md)