ROS下使用摄像头进行机器人巡线功能（附带功能包链接）

功能包地址：https://download.csdn.net/download/qq_42145185/12256533     下面代码是HSV格式的，参考图：   主要功能代码(线条检测功能代码):

#pragma once
#include <cv_bridge/cv_bridge.h>
#include <vector>
#include <opencv2/highgui/highgui.hpp>
#include "opencv2/opencv.hpp"
#include "ros/ros.h"
#include "line_follower_turtlebot/pos.h"
#include <geometry_msgs/Twist.h>
#include <cstdlib>
#include <string>
#include "ros/console.h"
/**
* @简短的Line Detect类包含用于图像处理和方向发布的所有功能
*/
class LineDetect {
 public:
    cv::Mat img;  /// 以OpenCV矩阵格式输入图像
    cv::Mat img_filt;  /// OpenCV矩阵格式的过滤图像
    int dir;  /// 要发布的指导消息
/**
* @ brief回调用于从Turtlebot订阅图像主题并将其转换为opencv图像格式
* @ param msg是ROS的图像消息
* @不返回
*/
    void imageCallback(const sensor_msgs::ImageConstPtr& msg);
/**
*@brief Function that applies Gaussian filter in the input image 
*@param input is the image from the turtlebot in opencv matrix format
*@return Mat of Gaussian filtered image in opencv matrix format
* @ brief函数在输入图像中应用高斯滤波器
* @ param输入是来自turtlebot的图像，格式为opencv
* @ opencv矩阵格式的高斯滤波图像的返回垫
*/
    cv::Mat Gauss(cv::Mat input);
/**
*@brief Function to perform line detection using color thresholding,image masking and centroid detection to publish direction 
*@param input is the Filtered input image in opencv matrix format
*@return int direction which returns the direction the turtlebot should head in
* @ brief功能使用颜色阈值，图像蒙版和质心检测执行发布线检测以发布方向
* @ param input是opencv矩阵格式的过滤后的输入图像
* @ return int direction返回乌龟应该进入的方向
*/
    int colorthresh(cv::Mat input);
 
 private:
    cv::Scalar LowerYellow;
    cv::Scalar UpperYellow;
    cv::Mat img_hsv;
    cv::Mat img_mask;
};
 
 
void LineDetect::imageCallback(const sensor_msgs::ImageConstPtr& msg) {
  cv_bridge::CvImagePtr cv_ptr;
  try {
    cv_ptr = cv_bridge::toCvCopy(msg, sensor_msgs::image_encodings::BGR8);
    img = cv_ptr->image;
    cv::waitKey(30);
  }
  catch (cv_bridge::Exception& e) {
    ROS_ERROR("Could not convert from '%s' to 'bgr8'.", msg->encoding.c_str());
  }
}
 
cv::Mat LineDetect::Gauss(cv::Mat input) {
  cv::Mat output;
// Applying Gaussian Filter应用高斯滤波器
  cv::GaussianBlur(input, output, cv::Size(3, 3), 0.1, 0.1);
  return output;
}
 
int LineDetect::colorthresh(cv::Mat input) {
  // Initializaing variables初始化变量
  cv::Size s = input.size();
  std::vector<std::vector<cv::Point> > v;
  auto w = s.width;
  auto h = s.height;
  auto c_x = 0.0;
  // Detect all objects within the HSV range RGBToHSV 检测HSV范围RGBToHSV内的所有对象
  cv::cvtColor(input, LineDetect::img_hsv, CV_BGR2HSV);
  LineDetect::LowerYellow = {0, 252, 255};//Red
  LineDetect::UpperYellow = {0, 255, 255};
  //LineDetect::LowerYellow = {100, 255, 100};//Blue
  //LineDetect::UpperYellow = {100, 255, 255};
  //LineDetect::LowerYellow = {20, 100, 100};//Yellow
  //LineDetect::UpperYellow = {30, 255, 255};
  cv::inRange(LineDetect::img_hsv, LowerYellow,
   UpperYellow, LineDetect::img_mask);
  img_mask(cv::Rect(0, 0, w, 0.8*h)) = 0;
  // Find contours for better visualization 查找轮廓以获得更好的可视化
  cv::findContours(LineDetect::img_mask, v, CV_RETR_LIST, CV_CHAIN_APPROX_NONE);
  // If contours exist add a bounding  如果存在轮廓，则添加边界
  // Choosing contours with maximum area  选择最大面积的轮廓
  if (v.size() != 0) {
  auto area = 0;
  auto idx = 0;
  auto count = 0;
  while (count < v.size()) {
    if (area < v[count].size()) {
       idx = count;
       area = v[count].size();
    }
    count++;
  }
  cv::Rect rect = boundingRect(v[idx]);
  cv::Point pt1, pt2, pt3;
  pt1.x = rect.x;
  pt1.y = rect.y;
  pt2.x = rect.x + rect.width;
  pt2.y = rect.y + rect.height;
  pt3.x = pt1.x+5;
  pt3.y = pt1.y-5;
  // Drawing the rectangle using points obtained  使用获得的点绘制矩形
  rectangle(input, pt1, pt2, CV_RGB(255, 0, 0), 2);
  // Inserting text box 插入文字框
  cv::putText(input, "Line Detected", pt3,
    CV_FONT_HERSHEY_COMPLEX, 1, CV_RGB(255, 0, 0));
  }
  // Mask image to limit the future turns affecting the output 遮罩图像以限制将来的转向影响输出
  img_mask(cv::Rect(0.7*w, 0, 0.3*w, h)) = 0;
  img_mask(cv::Rect(0, 0, 0.3*w, h)) = 0;
  // Perform centroid detection of line  执行线的质心检测
  cv::Moments M = cv::moments(LineDetect::img_mask);
  if (M.m00 > 0) {
    cv::Point p1(M.m10/M.m00, M.m01/M.m00);
    cv::circle(LineDetect::img_mask, p1, 5, cv::Scalar(155, 200, 0), -1);
  }
  c_x = M.m10/M.m00;
  // Tolerance to chooise directions  容忍选择方向
  auto tol = 15;
  auto count = cv::countNonZero(img_mask);
  // Turn left if centroid is to the left of the image center minus tolerance
  // Turn right if centroid is to the right of the image center plus tolerance
  // Go straight if centroid is near image center
  //如果质心在图像中心的左侧减去公差，则向左转
  //如果质心在图像中心的右侧加上公差，则向右转
  //如果质心在图像中心附近，请直走
  if (c_x < w/2-tol) {
    LineDetect::dir = 0;
  } else if (c_x > w/2+tol) {
    LineDetect::dir = 2;
  } else {
    LineDetect::dir = 1;
  }
  // Search if no line detected 搜索是否未检测到行
  if (count == 0) {
    LineDetect::dir = 3;
  }
  // Output images viewed by the turtlebot 输出乌龟机器人查看的图像
  cv::namedWindow("Turtlebot View");
  imshow("Turtlebot View", input);
  return LineDetect::dir;
}
 
 
 
int main(int argc, char **argv) {
    // Initializing node and object初始化节点和对象
    ros::init(argc, argv, "detection");
    ros::NodeHandle n;
    LineDetect det;
    // Creating Publisher and subscriber创建发布者和订阅者
    ros::Subscriber sub = n.subscribe("/camera/rgb/image_raw",
        1, &LineDetect::imageCallback, &det);
    //ros::Subscriber sub = n.subscribe("/usb_cam/image_raw",
    //    1, &LineDetect::imageCallback, &det);
 
    ros::Publisher dirPub = n.advertise<
    line_follower_turtlebot::pos>("direction", 1);
        line_follower_turtlebot::pos msg;
 
    while (ros::ok()) {
        if (!det.img.empty()) {
            // Perform image processing执行图像处理
            det.img_filt = det.Gauss(det.img);
            msg.direction = det.colorthresh(det.img_filt);
            // Publish direction message发布指导消息
            dirPub.publish(msg);
            }
        ros::spinOnce();
    }
    // Closing image viewer关闭图像查看器
    cv::destroyWindow("Turtlebot View");
}

  机器人控制功能代码：

#include <cv_bridge/cv_bridge.h>
#include <cstdlib>
#include <string>
#include <opencv2/highgui/highgui.hpp>
#include "opencv2/opencv.hpp"
#include "ros/ros.h"
#include "ros/console.h"
#include "line_follower_turtlebot/pos.h"
#include <geometry_msgs/Twist.h>
#include <vector>
 
 
/**
*@brief Class turtlebot subscribes to the directions published and publishes velocity commands
* @ brief类turtlebot订阅发布的指令并发布速度命令
*/
class turtlebot {
 public:
    int dir;  /// Direction message to read published directions指示消息以读取已发布的指示
/**
*@brief Callback used to subscribe to the direction message published by the Line detection node
*@param msg is the custom message pos which publishes a direction int between 0 and 3
*@return none
* @ brief回调用于订阅线路检测节点发布的方向消息
* @ param msg是自定义消息pos，它发布0到3之间的int方向
* @不返回
*/
    void dir_sub(line_follower_turtlebot::pos msg);
/**
*@brief Function to publish velocity commands based on direction
*@param velocity is the twist 
*@param pub is used to publish the velocity commands to the turtlebot
*@param rate is the ros loop rate for publishing the commands
*@return none
* @ brief函数可根据方向发布速度命令
* @参数速度是扭曲
* @ param pub用于将速度命令发布到turtlebot
* @ param rate是发布命令的ros循环速率
* @不返回
*/
    void vel_cmd(geometry_msgs::Twist &velocity,
     ros::Publisher &pub, ros::Rate &rate);
};
 
 
 
void turtlebot::dir_sub(line_follower_turtlebot::pos msg) {
    turtlebot::dir = msg.direction;
}
void turtlebot::vel_cmd(geometry_msgs::Twist &velocity,
 ros::Publisher &pub, ros::Rate &rate) {
    // If direction is left
    if (turtlebot::dir == 0) {
        velocity.linear.x = 0.3;
        velocity.angular.z = 0.3;
        pub.publish(velocity);
        rate.sleep();
        ROS_INFO_STREAM("Turning Left");
    }
    // If direction is straight
    if (turtlebot::dir == 1) {
        velocity.linear.x = 0.3;
        velocity.angular.z = 0;
        pub.publish(velocity);
        rate.sleep();
        ROS_INFO_STREAM("Straight");
    }
    // If direction is right
    if (turtlebot::dir == 2) {
        velocity.linear.x = 0.3;
        velocity.angular.z = -0.3;
        pub.publish(velocity);
        rate.sleep();
        ROS_INFO_STREAM("Turning Right");
    }
    // If robot has to search
    if (turtlebot::dir == 3) {
        velocity.linear.x = 0;
        velocity.angular.z = 0.25;
        pub.publish(velocity);
        rate.sleep();
        ROS_INFO_STREAM("Searching");
    }
}
 
 
int main(int argc, char **argv) {
    // Initializing node and object
    ros::init(argc, argv, "Velocity");
    ros::NodeHandle n;
    turtlebot bot;
    geometry_msgs::Twist velocity;
    // Creating subscriber and publisher
    ros::Subscriber sub = n.subscribe("/direction",
        1, &turtlebot::dir_sub, &bot);
    ros::Publisher pub = n.advertise<geometry_msgs::Twist>
        ("/cmd_vel_mux/input/teleop", 10);
    ros::Rate rate(10);
    while (ros::ok()) {
        ros::spinOnce();
        // Publish velocity commands to turtlebot
        bot.vel_cmd(velocity, pub, rate);
        rate.sleep();
    }
    return 0;
}