Implementing land-based litter surveys through visual inspection of imagery using unmanned aerial vehicles

Abstract

Understanding the location of dumping sites, or where littering occurs, is needed for effective and efficient waste management efforts. South Africa experiences open indiscriminate dumping and faces a challenging social practice of littering. This pilot study used unmanned aerial vehicle imagery, aimed to monitor litter loads (illegal dumps and individual littered items) over a large land area (> 2 km2) to assess potential change in litter through time. Two techniques, the Percentage Cover Technique and the Comparison Technique, were used. The two techniques were trialed at five sites to quantify litter load and monitor change over time. Overall, through the Percentage Cover Technique, 13.5% of the study area was covered in litter in February 2020, and this decreased to 10.3% by July 2021. All five sites individually showed a decrease in litter loads, with significant differences between imagery sets (H(2) = 52.786, P < 0.001) and between surveyors (U = 15 567 399, n1=n2=5 211, P < 0.001 two-tailed). The Comparison Technique revealed that of the 10 422 blocks analysed in this study, litter coverage decreased in 10.3% and increased in 8.3%, of them. The study showed that the Comparison Technique is an easy and inexpensive method to monitor litter loads on large scales and could therefore be replicated across Africa which has limited resources (financial and expertise).