Pedestrian_detection_dnn_opencv

This code shows the use of Mobile Net Single Shot detector provided with opencv for rosbag data obtained from realsense camera.

The package loads existing Caffe model from (https://docs.opencv.org/3.3.0/d5/de7/tutorial_dnn_googlenet.html)) googlenet dnn page and use it to classify pedestrian in the camera image topic.

Requirements

ROS::Kinect, OpenCV

Steps to follow:

Launch roscore

  • $ roscore

Launch Rosbag node

  • $ rosbag play [Bagfile]

Package

Detected pedestrian is published to /image_converter/output_video