Mapping flowering leafy spurge infestations in a heterogeneous landscape using unmanned aerial vehicle Red-Green-Blue images and a hybrid classification method. Issue 23 (2nd December 2021)

• Record Type: Journal Article
• Authors: Yang, Xiaohui; Smith, A. M.; Bourchier, R. S.; Hodge, K.; Ostrander, D.; Houston, B.
• Is Part Of: International journal of remote sensing. Volume 42:Issue 23(2021)
• Page Start: 8930
• Page End: 8951
• Publication Date: 2021-12-02
• Subjects: Remote sensing -- Periodicals; Télédétection -- Périodiques; 621.3678
• DOI: 10.1080/01431161.2021.1973686
• Languages: English
• ISSNs: 0143-1161

Abstract

ABSTRACT: Leafy spurge is a perennial, noxious, and invasive weed. It has been a major threat to the prairie ecosystems in North America due to encroachment and has caused significant biological and economic impacts. Advances in unmanned aerial vehicle (UAV) technology and UAV-based sensors facilitate non-destructive, low cost, flexible and near-real-time vegetation monitoring and provide a promising alternative to field surveys for mapping invasive species. This study investigated the utility of UAV technology for mapping and monitoring flowering leafy spurge in a heterogeneous landscape. The study was conducted in a complex landscape of grasses, shrubs, and trees, near Last Mountain Lake in southern Saskatchewan, Canada. Red-Green-Blue (RGB) images were acquired in 28 June 2015 and 5 July 2017 with partial overlap of geographical coverage during peak leafy spurge flowering period using a fixed-wing UAVs equipped with consumer-grade cameras. The images were classified using a hybrid classification schema, combination of Hue-Intensity-Saturation (HIS) thresholding and object-based post-processing methods. This method had some limitations in separating flowering leafy spurge from the surrounding vegetation however, flowering leafy spurge were effectively detected with an overall accuracy of 78% using 2015 image and 75% using 2017 image. The change in flowering leafy spurge distribution for a common area of coverage overlap was quantified based on maps generated for the two years. Leafy spurge coverage declined 0.2%, approximately 0.27 ha, in 2017. The results revealed that UAV RGB images can be routinely used for operational programmes to monitor leafy spurge in complex landscapes and aid in the management activities.