0
0

SLAM算法与工程实践——相机篇：传统相机使用（1）

SLAM算法与工程实践系列文章

下面是SLAM算法与工程实践系列文章的总链接，本人发表这个系列的文章链接均收录于此

SLAM算法与工程实践系列文章链接

下面是专栏地址：

SLAM算法与工程实践系列专栏

文章目录

前言

这个系列的文章是分享SLAM相关技术算法的学习和工程实践

SLAM算法与工程实践——相机篇：传统相机使用（1）

相机相关命令

插上USB相机，使用命令查看USB设备

  1. lsusb

在这里插入图片描述

可以识别相机

使用命令查看识别到几个摄像头

  1. ll /dev/video*

在这里插入图片描述

然后改变其权限

  1. sudochmod777 /dev/video0
  2. sudochmod777 /dev/video1

安装 

  1. v4l-utils

工具包

  1. sudoapt-getinstall v4l-utils

使用下列命令查看相机参数

  1. v4l2-ctl -d /dev/video0 --list-formats-ext

在命令行输出为：

  1. ioctl: VIDIOC_ENUM_FMT
  2.         Type: Video Capture
  4.         [0]: 'MJPG'(Motion-JPEG, compressed)
  5.                 Size: Discrete 2560x800
  6.                         Interval: Discrete 0.017s (60.000 fps)
  7.                         Interval: Discrete 0.033s (30.000 fps)
  8.                         Interval: Discrete 0.067s (15.000 fps)
  9.                 Size: Discrete 2560x720
  10.                         Interval: Discrete 0.017s (60.000 fps)
  11.                         Interval: Discrete 0.033s (30.000 fps)
  12.                 Size: Discrete 1280x800
  13.                         Interval: Discrete 0.008s (120.000 fps)
  14.                         Interval: Discrete 0.017s (60.000 fps)
  15.                         Interval: Discrete 0.033s (30.000 fps)
  16.                 Size: Discrete 1280x792
  17.                         Interval: Discrete 0.008s (120.000 fps)
  18.                         Interval: Discrete 0.017s (60.000 fps)
  19.                         Interval: Discrete 0.033s (30.000 fps)
  20.                 Size: Discrete 1280x400
  21.                         Interval: Discrete 0.008s (120.000 fps)
  22.                         Interval: Discrete 0.017s (60.000 fps)
  23.                         Interval: Discrete 0.033s (30.000 fps)
  24.                 Size: Discrete 640x400
  25.                         Interval: Discrete 0.005s (210.000 fps)
  26.                         Interval: Discrete 0.008s (120.000 fps)
  27.                         Interval: Discrete 0.017s (60.000 fps)
  28.                 Size: Discrete 640x392
  29.                         Interval: Discrete 0.005s (210.000 fps)
  30.                         Interval: Discrete 0.008s (120.000 fps)
  31.                         Interval: Discrete 0.017s (60.000 fps)[1]: 'YUYV'(YUYV 4:2:2)
  32.                 Size: Discrete 2560x800
  33.                         Interval: Discrete 0.500s (2.000 fps)
  34.                 Size: Discrete 2560x720
  35.                         Interval: Discrete 0.500s (2.000 fps)
  36.                 Size: Discrete 1280x800
  37.                         Interval: Discrete 0.200s (5.000 fps)
  38.                 Size: Discrete 1280x792
  39.                         Interval: Discrete 0.200s (5.000 fps)
  40.                 Size: Discrete 1280x400
  41.                         Interval: Discrete 0.067s (15.000 fps)
  42.                         Interval: Discrete 0.100s (10.000 fps)
  43.                 Size: Discrete 640x400
  44.                         Interval: Discrete 0.033s (30.000 fps)
  45.                 Size: Discrete 640x392
  46.                         Interval: Discrete 0.033s (30.000 fps)

可以看出本相机的参数和对应设置下的帧率，最高为 120 帧，分辨率为 2560x800，我这选用 1280x400

在这里插入图片描述

接下来看看你的USB摄像头支持什么附加属性，只需要执行下面的命令即可

  1. v4l2-ctl -d /dev/video0 --list-ctrls

得到下面的结果

  1. User Controls
  3.                      brightness 0x00980900 (int):min=-64 max=64step=1default=0value=0
  4.                        contrast 0x00980901 (int):min=0max=95step=1default=0value=0
  5.                      saturation 0x00980902 (int):min=0max=100step=1default=64value=64
  6.                             hue 0x00980903 (int):min=-2000 max=2000step=1default=0value=0
  7.         white_balance_automatic 0x0098090c (bool):default=1value=1
  8.                           gamma 0x00980910 (int):min=100max=300step=1default=100value=100

在这里插入图片描述

使用应用程序茄子（cheese）

输入命令：

  1. sudoapt-getinstall cheese

装好后，用命令：

  1. cheese

，即可打开。如果指定打开video1，输入命令：

  1. cheese -d /dev/video1

读取相机配置文件 

  1. HBV-1780-2.yaml
  1. int UseStereoCam::loadCamParam(std::string cfgPath) {
  2.     cv::FileStorage fs(cfgPath, cv::FileStorage::READ);
  3.     if (fs.isOpened()) {
  4.         std::cout << "Stereo Camera Paramter is loading..." << std::endl;
  6.         //内参矩阵
  7.         fs["IntrinsicMatrixLeft"] >> IntrinsicMatrixLeft;      //存放内参矩阵,左相机
  8.         IntrinsicMatrixLeft.at<double>(0, 0) = IntrinsicMatrixLeft.at<double>(0, 0) * resizeScalse;
  9.         IntrinsicMatrixLeft.at<double>(1, 1) = IntrinsicMatrixLeft.at<double>(1, 1) * resizeScalse;
  10.         IntrinsicMatrixLeft.at<double>(0, 2) = IntrinsicMatrixLeft.at<double>(0, 2) * resizeScalse;
  11.         IntrinsicMatrixLeft.at<double>(1, 2) = IntrinsicMatrixLeft.at<double>(1, 2) * resizeScalse;
  13.         fs["IntrinsicMatrixRight"] >> IntrinsicMatrixRight;      //存放内参矩阵,右相机
  14.         IntrinsicMatrixRight.at<double>(0, 0) = IntrinsicMatrixRight.at<double>(0, 0) * resizeScalse;
  15.         IntrinsicMatrixRight.at<double>(1, 1) = IntrinsicMatrixRight.at<double>(1, 1) * resizeScalse;
  16.         IntrinsicMatrixRight.at<double>(0, 2) = IntrinsicMatrixRight.at<double>(0, 2) * resizeScalse;
  17.         IntrinsicMatrixRight.at<double>(1, 2) = IntrinsicMatrixRight.at<double>(1, 2) * resizeScalse;
  19.         //畸变系数
  20.         fs["DistCoeffLeft"] >> DistCoeffLeft;            //畸变矩阵，左相机
  21.         fs["DistCoeffRight"] >> DistCoeffRight;             //畸变矩阵，右相机
  23.         //旋转矩阵 平移向量
  24.         fs["RotationMatrix"] >> R;
  25.         fs["TranslationVector"] >> T;
  26.         T = T * resizeScalse;
  28.         fs.release();
  30.         leftFocalLength =
  31.                 (IntrinsicMatrixLeft.at<double>(0, 0) + IntrinsicMatrixLeft.at<double>(1, 1)) / 2.0;
  32.         rightFocalLength =
  33.                 (IntrinsicMatrixRight.at<double>(0, 0) + IntrinsicMatrixRight.at<double>(1, 1)) / 2.0;
  34.         // 处理焦距（例如取平均值）
  35.         focalLength = (leftFocalLength + rightFocalLength) / 2.0;
  37.         fx = (IntrinsicMatrixLeft.at<double>(0, 0) + IntrinsicMatrixRight.at<double>(0, 0)) / 2.0;
  38.         fy = (IntrinsicMatrixLeft.at<double>(1, 1) + IntrinsicMatrixRight.at<double>(1, 1)) / 2.0;
  39.         cx = (IntrinsicMatrixLeft.at<double>(0, 2) + IntrinsicMatrixRight.at<double>(0, 2)) / 2.0;
  40.         cy = (IntrinsicMatrixLeft.at<double>(1, 2) + IntrinsicMatrixRight.at<double>(1, 2)) / 2.0;
  42.         baseline = cv::norm(T) / 100;     //基线（baseline）单位设置为分米
  44.         std::cout << "camera parameter loaded successfully!" << std::endl;
  46.         return 0;
  48.     } else {
  49.         std::cout << "Failed to load camera parameter!" << std::endl;
  50.         return -1;
  51.     }
  53. }

出现的问题

参考：

解决OpenCV的GStreamer warning警告

在用opencv调用摄像头时报错

在这里插入图片描述

  1. [ WARN:0] global ../modules/videoio/src/cap_gstreamer.cpp (935)open OpenCV | GStreamer warning: Cannot query video position: status=0, value=-1, duration=-1
  2. [ WARN:0] global ../modules/videoio/src/cap_gstreamer.cpp (1758) handleMessage OpenCV | GStreamer warning: Embedded video playback halted; module v4l2src0 reported: Device '/dev/video0' failed during initialization
  3. [ WARN:0] global ../modules/videoio/src/cap_gstreamer.cpp (515) startPipeline OpenCV | GStreamer warning: unable to start pipeline
  4. [ WARN:0] global ../modules/videoio/src/cap_gstreamer.cpp (1057) setProperty OpenCV | GStreamer warning: no pipeline
  5. [ WARN:0] global ../modules/videoio/src/cap_gstreamer.cpp (480) isPipelinePlaying OpenCV | GStreamer warning: GStreamer: pipeline have not been created
  6. 摄像头打开失败!

解决办法：

修改打开摄像头的代码如下：

  1. cap.open(0,CAP_V4L2);

调用相机

在编写多个主函数共用

  1. src

和

  1. include

定义的类的文件时，CMakeLists.txt的编写应该类似下面的写法，然后包含每个子项目的子文件夹

  1. cmake_minimum_required(VERSION 2.8)
  2. project(DEPTH_COMPUTE)
  4. IF (NOT CMAKE_BUILD_TYPE)
  5.     SET(CMAKE_BUILD_TYPE Release)
  6. ENDIF ()
  8. MESSAGE("Build type: " ${CMAKE_BUILD_TYPE})
  10. # set path
  11. set(LIB_PATH ${PROJECT_SOURCE_DIR}/lib)
  12. set(HEAD_PATH ${PROJECT_SOURCE_DIR}/include)
  13. set(EXEC_PATH ${PROJECT_SOURCE_DIR}/bin)
  15. set(DEPTH_LIB calc_depth)   # set libiary name
  17. find_package(OpenCV 4 REQUIRED)
  18. find_package(Eigen3 3.1.0 NO_MODULE)
  19. find_package(Pangolin REQUIRED)
  21. include_directories(${HEAD_PATH})
  22. aux_source_directory(${PROJECT_SOURCE_DIR}/src LIB_SRC)
  24. SET(LIBRARY_OUTPUT_PATH ${LIB_PATH})
  25. add_library(${DEPTH_LIB} SHARED ${LIB_SRC})
  26. TARGET_LINK_LIBRARIES(${DEPTH_LIB}
  27.         ${OpenCV_LIBS}
  28.         ${EIGEN3_LIBS}
  29.         ${Pangolin_LIBRARIES}
  30. )
  32. # add subdirectories 这里是具体的主函数
  33. add_subdirectory(useCamForShot) 
  34. add_subdirectory(openCamera)

子项目的CMakeLists.txt应该如下

  1. cmake_minimum_required(VERSION 2.8)
  2. project(OPENCAM)
  4. find_package(OpenCV 4 REQUIRED)
  6. include_directories(${HEAD_PATH})
  7. link_directories(${LIB_PATH})
  9. set(CMAKE_BUILD_TYPE "Release")
  10. set(CMAKE_CXX_FLAGS "-O3")
  11. set(EXECUTABLE_OUTPUT_PATH  ${EXEC_PATH})
  13. #add_library(${DEPTH_LIB} SHARED  ${SRC_LIST})
  15. add_executable(open_camera open_camera.cpp)
  16. target_link_libraries(open_camera ${DEPTH_LIB} ${OpenCV_LIBS})

以MJPG格式打开相机

参考：

OpenCV+V4L实现MJPG格式拉取USB摄像头

关于openCV VideoCapture获取摄像头MJPG视频的问题

相机支持的视频流格式

在这里插入图片描述

在这里插入图片描述

在打开相机后，设置参数

  1. capture.open(camIndex, cv::CAP_V4L);    // open camera
  3. //设置编码格式，这里只能写在分辨率和帧率后面，否则无效
  4. capture.set(cv::CAP_PROP_FOURCC, cv::VideoWriter::fourcc('M', 'J', 'P', 'G'));//视频流格式
标签： 算法 计算机视觉 机器人
本文转载自: https://blog.csdn.net/qq_37648371/article/details/134984819
版权归原作者 yangtsejin 所有， 如有侵权，请联系我们删除。
登录后发布评论

“SLAM算法与工程实践——相机篇：传统相机使用（1）”的评论:

还没有评论
关于作者
...
overfit同步小助手
文章同步
相关阅读

网络安全法-网络运行安全

使用selenium/drissionpage时如何阻止chrome自动跳转http到https

docker desktop 里部署的Open WebUI 管理员密码忘记了的处理方法

在ubuntu20.04中搭建onsite比赛运行环境

利用开源的低代码表单设计器FcDesigner高效管理和渲染复杂表单结构

Kafka学习笔记

【前端】浏览器输入url到页面呈现发生了什么？