Kinect camera based target detection and mapping system for robotic platforms

In this study, it is aimed to perform object detection and determine the distance of the detected object from the camera by using image processing methods which have been increasingly used in recent years thanks to technological developments. In this context, with the help of a Kinect camera and a GPS receiver placed on a mobile robot platform, target / obstacle detection was performed in open area and the coordinates of the target or obstacle were determined and mapped. It is not a common practice in the literature to identify and coordinate the obstacles or targets encountered during navigation of unmanned land vehicles using the Kinect camera. Related camera applications are mostly carried out in closed areas. The fact that this practice was performed in the open area exposes the authenticity of the study. The system has been designed for vision-based navigation applications in unmanned land vehicles used in agricultural areas and can as well be employed in any other study utilizing unmanned land vehicles. In the project, an obstacle and target detection mechanism has been designed for a robotic platform which has the ability serve in agricultural areas. For this purpose, a mobile robot, a kinect camera, a GPS receiver, a digital compass,a computer and a software developed by us have been used. The GPS receiver and the Kinect camera were positioned in the same direction on the mobile robot. The mobile robot determines the obstacle / target coordinates encountered during the movement using the distance and the coordinate information of the location of the kinect camera and transfers it to the database. In this way, coordinates of all kinds of instant obstacles and targets available in the area were gathered and a database has been created for further studies.