High-resolution drone aerial surveys combined with object-based image analysis are transforming our capacity to monitor and manage aquatic vegetation in an era of invasive species. To better exploit the potential of these technologies, there is a need to develop more efficient and accessible analysis workflows and focus more efforts on the distinct challenge of mapping submerged vegetation. We present a straightforward workflow developed to monitor emergent and submerged invasive water soldier (Stratiotes aloides) in shallow waters of the Trent-Severn Waterway in Ontario, Canada. The main elements of the workflow are: (1) collection of radiometrically calibrated multispectral imagery including a near-infrared band; (2) multistage segmentation of the imagery involving an initial separation of above-water from submerged features; and (3) automated classification of features with a supervised machine-learning classifier. The approach yielded excellent classification accuracy for emergent features (overall accuracy = 92%; kappa = 88%; water soldier producer’s accuracy = 92%; user’s accuracy = 91%) and good accuracy for submerged features (overall accuracy = 84%; kappa = 75%; water soldier producer’s accuracy = 71%; user’s accuracy = 84%). The workflow employs off-the-shelf graphical software tools requiring no programming or coding, and could therefore be used by anyone with basic GIS and image analysis skills for a potentially wide variety of aquatic vegetation monitoring operations.

ABSTRACT. Computer-automated image analysis techniques can save time and resources for detecting and counting birds in aerial imagery. Sophisticated object-based image analysis (OBIA) software is now widely available and has proven effective for various challenging detection tasks, but there is a need to develop accessible and readily adaptable procedures that can be implemented in an operational context. We developed a systematic, repeatable approach using commercial off-the-shelf OBIA software, and tested its effectiveness and efficiency to detect and count Lesser Snow Geese (Chen caerulescens caerulescens) in large numbers of images of breeding colonies across the Canadian Arctic that present a variety of landscapes, numerous confounding features, and varying illumination conditions and exposure levels. Coarse-scale review of analysis results was necessary to remove conspicuous clusters of commission errors, thus rendering the technique semiautomated. It was effective for imagery with spatial resolutions of 4-5 cm, producing overall accurate estimates of goose numbers compared to manual counts (R2 = 0.998, regression coefficient = 0.974) in 41 test images drawn from several breeding colonies. The total automated count (19,920) across all test images exceeded the manual count (19,836) by just 0.4%. We estimate the typical time required to review images for errors to be only 5-10% of that required to count birds manually. This could reduce the person-time required to analyze aerial photos of the major Arctic colonies of Snow Geese from several months to several days. Our approach could be adapted to many other bird detection tasks in aerial imagery by anyone possessing at least basic skills in image analysis and geographic information systems.

Small unmanned aircraft systems (UAS) combined with automated image analysis may provide an efficient alternative or complement to labour-intensive boat-based monitoring of invasive aquatic vegetation. A small mapping drone was assessed for collecting high-resolution (≤5 cm/pixel) true-colour and near-infrared imagery revealing the distribution of invasive water soldier (Stratiotes aloides) in the Trent–Severn Waterway, Ontario (Canada). We further evaluated the capacity of an object-based image analysis approach based on the Random Forests classification algorithm to map features in the imagery, chiefly emergent and submerged water soldier colonies. The imagery contained flaws and inconsistencies resulting from data collection in suboptimal weather conditions that likely negatively impacted classification performance. Nevertheless, our best-performing classification had a producer’s and user’s accuracy for water soldier of 81% and 74%, respectively, an overall accuracy of 78%, and a kappa value of 61%, indicating “substantial” accuracy. This trial provides an instructive case study on results achieved in a “real-world” application of a UAS for environmental monitoring, notably characterized by time constraints for data collection and analysis. Beyond avoiding data collection in unfavourable weather conditions, adaptations of the image segmentation process and use of a true discrete-band multispectral camera may help to improve classification accuracy, particularly of submerged vegetation.

No abstract

No abstract

No abstract