A ROS Indigo package implementing the OpenCV HOG pedestrian detector[1] and HAAR face detector. Subscribes to an image being published, detects faces and standing pedestrians in the image, draws the detections on a copy of the image and publishes results as a rostopic. Also displays detections on an OpenCV window.
- $ cd [your indigo catkin src folder]
- $ git clone https://github.com/angusleigh/hog_haar_person_detection.git
- $ cd ..
- $ catkin_make
- $ roslaunch hog_haar_person_detection hog_haar_person_detection.launch
- Launch your camera node and have it publish to the "image_topic" parameter sepecified in hog_haar_person_detection.launch. Check out the open-ni node for the Kinect or USB-Cam node for USB cameras.
[1] Link is for GPU documentation. I'm not aware of any documentation for the CPU implementaion (which is the one used for this repo).