An unmanned aerial vehicle for autonomous forestry cutover edge survey.
| dc.contributor.author | Hunt, David | |
| dc.date.accessioned | 2016-08-16T01:32:09Z | |
| dc.date.available | 2016-08-16T01:32:09Z | |
| dc.date.issued | 2016 | en |
| dc.description.abstract | A proof of concept unmanned aerial system capable of navigating the edge defined by the boundary of forest and cutover has been developed. Following gathering of low altitude imagery of forest cutover edges, two image classifiers were developed to classify regions of an image as forest or not-forest from RGB images. The first proposed method, based on selection of hue, saturation, and brightness, achieved a classification accuracy of 83.2%. The second proposed method, based on a SVM classifier, achieved a higher classification of 90.2% by using textural information to avoid false positives. The two methods were combined in order to accurately identify the cutover edge and fit a linear line. A proportional control system was used to generate navigational instructions based on the position and orientation of the line within a frame from a downward facing camera onboard a UAV. Mavlink commands were generated on an external computer onboard the UAV and sent to a Pixhawk flight computer to control a UAV. Successful performance of the proposed combined cutover edge identification algorithm was demonstrated in simulation with an average error of 26.2 pixels, approximately corresponding to a metric value of only 2.4 m. The simulated accuracy of this system compares favourably to a similar automated UAV system (Rathinam et al. [2007]) whose average positional error was 7 m. | en |
| dc.identifier.uri | http://hdl.handle.net/10092/12582 | |
| dc.identifier.uri | http://dx.doi.org/10.26021/3461 | |
| dc.language | English | |
| dc.language.iso | en | |
| dc.publisher | University of Canterbury | en |
| dc.rights | All Right Reserved | en |
| dc.rights.uri | https://canterbury.libguides.com/rights/theses | en |
| dc.title | An unmanned aerial vehicle for autonomous forestry cutover edge survey. | en |
| dc.type | Theses / Dissertations | |
| thesis.degree.discipline | Computer Science | en |
| thesis.degree.discipline | Electrical Engineering | en |
| thesis.degree.grantor | University of Canterbury | en |
| thesis.degree.level | Masters | en |
| thesis.degree.name | Master of Engineering | en |
| uc.bibnumber | 2362367 | |
| uc.college | Faculty of Engineering | en |