8
7

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

More than 5 years have passed since last update.

ROS講座80 compressedイメージをOpenCVで読む

Last updated at Posted at 2018-11-11

環境

この記事は以下の環境で動いています。

項目
CPU Core i5-8250U
Ubuntu 16.04
ROS Kinetic
Gazebo 7.0.0
python 2.7.12
OpenCV 3.3.1-dev

インストールについてはROS講座02 インストールを参照してください。
またこの記事のプログラムはgithubにアップロードされています。ROS講座11 gitリポジトリを参照してください。

概要

これまでOpenCVで取り込む画像はimage_raw[sensor_msgs/Image]を使っていました。ただ伝送路の帯域の都合で圧縮画像を使いたいことがあります。image_transportのROSノードを使ってもよいのですが面倒なのでOpenCVで変換してみます。

ソースコード

cam_lecture/scripts/edge_filter_compressed.py
#!/usr/bin/env python
import rospy
import sys
import cv2
from sensor_msgs.msg import Image, CompressedImage, CameraInfo
from cv_bridge import CvBridge, CvBridgeError
import numpy as np

class cvBridgeDemo:
    def __init__(self):
        self.node_name = "cv_bridge_demo_compressed"
        rospy.init_node(self.node_name)
        rospy.on_shutdown(self.cleanup)
        self.bridge = CvBridge()
        self.image_sub = rospy.Subscriber("input_image", CompressedImage, self.image_callback, queue_size=1)
        self.image_pub = rospy.Publisher('output_image', Image, queue_size=1)

    def image_callback(self, ros_image_compressed):
        try:
            np_arr = np.fromstring(ros_image_compressed.data, np.uint8)
            input_image = cv2.imdecode(np_arr, cv2.IMREAD_COLOR)
        except CvBridgeError, e:
            print e
        
        output_image = self.process_image(input_image)
        self.image_pub.publish(self.bridge.cv2_to_imgmsg(output_image, "mono8"))
        
        cv2.imshow(self.node_name, output_image)   
        cv2.waitKey(1)
                          
    def process_image(self, frame):
        grey = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
        grey = cv2.blur(grey, (7, 7))
        edges = cv2.Canny(grey, 15.0, 30.0)
        return edges
        
    def cleanup(self):
        cv2.destroyAllWindows()   
    
if __name__ == '__main__':
    cvBridgeDemo()
    rospy.spin()

edge_filter.pyとの違いを説明します。
self.image_sub = rospy.Subscriber("input_image", CompressedImage, self.image_callback, queue_size=1)def image_callback(self, ros_image_compressed):のようにsubする型をImage->CompressedImageに変更します。
ROSメッセージからOpenCV形式への変換はnp_arr = np.fromstring(ros_image_compressed.data, np.uint8)input_image = cv2.imdecode(np_arr, cv2.IMREAD_COLOR)で行います。

以下が起動用のlaunchです。

cam_lecture/launch/sim_edge_filter_compressed.launch
<?xml version="1.0" encoding="UTF-8"?>
<launch>
  <arg name="model" default="$(find cam_lecture)/urdf/camera_sim.urdf" />
  <arg name="rvizconfig" default="$(find cam_lecture)/rviz/camera_sim.rviz" />
  <param name="robot_description" type="str" textfile="$(arg model)"/>

  <include file="$(find gazebo_ros)/launch/empty_world.launch">
    <arg name="world_name" value="$(find cam_lecture)/world/camera_sim.world" />
    <arg name="paused" value="false"/>
    <arg name="use_sim_time" value="true"/>
    <arg name="gui" value="true"/>
    <arg name="headless" value="false"/>
    <arg name="debug" value="false"/>
  </include>

  <node name="spawn_urdf" pkg="gazebo_ros" type="spawn_model" args="-param robot_description -urdf -model my_robot" />

  <node name="edge_filter_compressed" pkg="cam_lecture" type="edge_filter_compressed.py" >
    <remap from="input_image" to="/camera/image_raw/compressed" />
  </node>

  <node name="robot_state_publisher" pkg="robot_state_publisher" type="robot_state_publisher" />   
  <node name="rviz" pkg="rviz" type="rviz" args="-d $(arg rvizconfig)" required="true" />
</launch>

実行

roslaunch cam_lecture sim_edge_filter_compressed.launch 

参考

ROS wiki

目次ページへのリンク

ROS講座の目次へのリンク

8
7
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
8
7

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?