zbw/yiliao2026
1
0
Fork 1
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

vlm和tts

Browse Source
This commit is contained in:
Orange
2026-07-23 14:51:51 +08:00
parent 90761b40b2
commit a5cee5e142
35 changed files with 12694 additions and 168 deletions
Download Patch File Download Diff File Expand all files Collapse all files

80
src/vlm_detect/launch/vlm_detect.launch.py
View File

@@ -1,13 +1,15 @@
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
vlm_detect 联合启动文件
同时启动 vlm_node (图生文) 和 tts_node (语音播报)
vlm_detect 启动文件
同时启动 vlm_node (图生文) + tts_server (语音播报服务) + qr_tts_bridge (二维码播报)
用法:
ros2 launch vlm_detect vlm_detect.launch.py # 默认配置
ros2 launch vlm_detect vlm_detect.launch.py vlm_host:=http://... # 覆盖 VLM 服务地址
ros2 launch vlm_detect vlm_detect.launch.py use_tts:=false # 只启动 vlm_node
ros2 launch vlm_detect vlm_detect.launch.py # 默认全部
ros2 launch vlm_detect vlm_detect.launch.py vlm_host:=http://... # 指定 VLM 服务地址
ros2 launch vlm_detect vlm_detect.launch.py use_tts:=false # 关闭语音播报
ros2 launch vlm_detect vlm_detect.launch.py use_qr_tts:=false # 关闭二维码播报桥接
ros2 launch vlm_detect vlm_detect.launch.py use_vlm:=false # 只启动语音服务
"""
import os
@@ -23,7 +25,9 @@ from launch_ros.actions import Node
def generate_launch_description():
# ==================== Launch 参数 ====================
use_vlm = LaunchConfiguration('use_vlm')
use_tts = LaunchConfiguration('use_tts')
use_qr_tts = LaunchConfiguration('use_qr_tts')
config_file = LaunchConfiguration('config_file')
@@ -37,15 +41,22 @@ def generate_launch_description():
prompt_text = LaunchConfiguration('prompt_text')
max_tokens = LaunchConfiguration('max_tokens')
# tts_node 可覆盖参数
# tts_server 可覆盖参数
audio_sink = LaunchConfiguration('audio_sink')
tts_voice = LaunchConfiguration('tts_voice')
tts_speed = LaunchConfiguration('tts_speed')
# ==================== 声明参数 ====================
# ==================== 参数声明 ====================
declare_use_vlm = DeclareLaunchArgument(
'use_vlm', default_value='true',
description='启动 VLM 图生文节点')
declare_use_tts = DeclareLaunchArgument(
'use_tts', default_value='true',
description='是否同时启动 TTS 语音播报节点')
description='启动 TTS 语音播报服务')
declare_use_qr_tts = DeclareLaunchArgument(
'use_qr_tts', default_value='true',
description='启动二维码 → TTS 桥接节点')
declare_config_file = DeclareLaunchArgument(
'config_file',
@@ -57,47 +68,46 @@ def generate_launch_description():
# vlm_node 参数
declare_vlm_host = DeclareLaunchArgument(
'vlm_host', default_value='http://192.168.10.189:8000',
description='VLM 推理服务地址')
description='VLM 服务器地址')
declare_vlm_model = DeclareLaunchArgument(
'vlm_model', default_value='./OpenGVLab/InternVL3-1B/',
description='VLM 模型名称')
declare_image_topic = DeclareLaunchArgument(
'image_topic', default_value='/image_mjpeg',
description='订阅的压缩图像话题')
description='输入的压缩图像话题')
declare_trigger_topic = DeclareLaunchArgument(
'trigger_topic', default_value='/sign4return',
description='订阅的触发信号话题')
description='输入的触发信号话题')
declare_trigger_sign = DeclareLaunchArgument(
'trigger_sign', default_value='9',
description='触发信号值 (Int32)')
declare_result_topic = DeclareLaunchArgument(
'result_topic', default_value='/vlm_result',
description='发布 VLM 结果的话题')
description='输出 VLM 结果的话题')
declare_prompt_text = DeclareLaunchArgument(
'prompt_text', default_value='描述图片中有一个病人的特征，字数控制在20字以内。',
'prompt_text', default_value='请描述这张图片的内容，用一句简短的话概括，不超过20个字。',
description='发送给 VLM 的提示词')
declare_max_tokens = DeclareLaunchArgument(
'max_tokens', default_value='100',
description='最大输出 token 数')
description='最大生成 token 数')
# tts_node 参数
# tts_server 参数
declare_audio_sink = DeclareLaunchArgument(
'audio_sink',
default_value='alsa_output.usb-C-Media_Electronics_Inc._USB_Audio_Device-00.analog-stereo',
description='音频输出设备 (PulseAudio sink)')
declare_tts_voice = DeclareLaunchArgument(
'tts_voice', default_value='zh-CN-XiaoxiaoNeural',
description='TTS 语音名称 (edge-tts)')
declare_tts_speed = DeclareLaunchArgument(
'tts_speed', default_value='1.5',
description='播放速度倍率 (0.5~2.0)')
description='语速倍率 (0.5~2.0)')
# ==================== 节点 ====================
# VLM 图生文节点 (内部自带 TTS 服务客户端)
vlm_node = Node(
package='vlm_detect',
executable='vlm_node',
name='vlm_detect',
output='screen',
condition=IfCondition(use_vlm),
parameters=[config_file,
{
'vlm_host': vlm_host,
@@ -111,26 +121,35 @@ def generate_launch_description():
}],
)
tts_node = Node(
# TTS 语音播报服务端
tts_server = Node(
package='vlm_detect',
executable='tts_node',
name='tts_node',
executable='tts_server',
name='tts_server',
output='screen',
condition=IfCondition(use_tts),
parameters=[config_file,
{
'vlm_host': vlm_host,
'audio_sink': audio_sink,
'result_topic': result_topic,
'tts_voice': tts_voice,
'tts_speed': tts_speed,
}],
)
# ==================== 启动描述 ====================
# 二维码 → TTS 桥接 (订阅 qr_results，调用 /tts/speak)
qr_tts_bridge = Node(
package='vlm_detect',
executable='qr_tts_bridge',
name='qr_tts_bridge',
output='screen',
condition=IfCondition(use_qr_tts),
)
# ==================== 组装 ====================
return LaunchDescription([
# 参数声明
declare_use_vlm,
declare_use_tts,
declare_use_qr_tts,
declare_config_file,
declare_vlm_host,
declare_vlm_model,
@@ -141,12 +160,13 @@ def generate_launch_description():
declare_prompt_text,
declare_max_tokens,
declare_audio_sink,
declare_tts_voice,
declare_tts_speed,
# 节点
LogInfo(msg=['配置文件: ', config_file]),
LogInfo(msg=['VLM 服务: ', vlm_host]),
LogInfo(msg=['TTS 播报: ', use_tts]),
LogInfo(msg=['TTS 服务: ', use_tts]),
LogInfo(msg=['QR-TTS 桥接: ', use_qr_tts]),
vlm_node,
tts_node,
tts_server,
qr_tts_bridge,
])