How the Small Object Detection via Machine Learning and UAS-Based Remote-Sensing Imagery Can Support the Achievement of SDG2: A Case Study of Vole Burrows

Abstract

Small mammals, and particularly rodents, are common inhabitants of farmlands, where they play key roles in the ecosystem, but when overabundant, they can be major pests, able to reduce crop production and farmers’ incomes, with tangible effects on the achievement of Sustainable Development Goals no 2 (SDG2, Zero Hunger) of the United Nations. Farmers do not currently have a standardized, accurate method of detecting the presence, abundance, and locations of rodents in their fields, and hence do not have environmentally efficient methods of rodent control able to promote sustainable agriculture oriented to reduce the environmental impacts of cultivation. New developments in unmanned aerial system (UAS) platforms and sensor technology facilitate cost-effective data collection through simultaneous multimodal data collection approaches at very high spatial resolutions in environmental and agricultural contexts. Object detection from remote-sensing images has been an active research topic over the last decade. With recent increases in computational resources and data availability, deep learning-based object detection methods are beginning to play an important role in advancing remote-sensing commercial and scientific applications. However, the performance of current detectors on various UAS-based datasets, including multimodal spatial and physical datasets, remains limited in terms of small object detection. In particular, the ability to quickly detect small objects from a large observed scene (at field scale) is still an open question. In this paper, we compare the efficiencies of applying one- and two-stage detector models to a single UAS-based image and a processed (via Pix4D mapper photogrammetric program) UAS-based orthophoto product to detect rodent burrows, for agriculture/environmental applications as to support farmer activities in the achievements of SDG2. Our results indicate that the use of multimodal data from low-cost UASs within a self-training YOLOv3 model can provide relatively accurate and robust detection for small objects (mAP of 0.86 and an F1-score of 93.39%), and can deliver valuable insights for field management with high spatial precision able to reduce the environmental costs of crop production in the direction of precision agriculture management.

Detection of spatial avoidance between sousliks and moles by combining field observations, remote sensing and deep learning techniques

“…The model training was stopped when the model was no longer improving. The following metrics were used to evaluate the classification model: precision (TP/(TP + FP)), recall (TP/(TP + FN) and F1-score 29 with “Compute Accuracy For Object Detection” tool in ArcGIS Pro. True Positives [TP] is the number of detections with intersection over union (IoU) > 0.5.…”

Rodent hole detection in a typical steppe ecosystem using UAS and deep learning

“…Wan et al (2021) successfully detected grassland rat holes (unspecified species) using R-CNN and improved Single Shot MultiBox Detector (SSD). Ezzy et al (2021) detected Levant voles burrows (2.5-7.5 cm in diameter) in farmlands and found that YOLOv3 provided relatively accurate and robust results. Previous studies explored various algorithms to detect different rodent holes under various environments.…”

Classification of airborne multispectral imagery to quantify common vole impacts on an agricultural field

“…In regards to rodent populations, two main areas have been explored in the literature: (1) survey and recognition of rodent infestation; 11,18,19 and (2) automatic identification of rodent evidence, such as burrows, by following a featureextraction schema. 12,20 A third group of investigations explores objective assessments of the damaged areas caused by the rodents, but in this case, the research relies on remote sensing, mainly by the use of Moderate Resolution Imaging Spectroradiometer (MODIS) imagery. 21 Currently, very little scientific research exists on the combined use of UAS and multispectral imagery for assessments of rodents' impact on agricultural fields, and more specifically in relation to the common vole.…”

“…Presently, there is scientific literature related to the use of drones and red/green/blue (RGB)‐based imagery for rodent infestation assessment and monitoring. In regards to rodent populations, two main areas have been explored in the literature: (1) survey and recognition of rodent infestation; 11,18,19 and (2) automatic identification of rodent evidence, such as burrows, by following a feature‐extraction schema 12,20 . A third group of investigations explores objective assessments of the damaged areas caused by the rodents, but in this case, the research relies on remote sensing, mainly by the use of Moderate Resolution Imaging Spectroradiometer (MODIS) imagery 21 .…”

“…However, little effort has been devoted to distinguishing pest‐damaged areas from healthy vegetation at a fine scale. For example, classification methods can help differentiate fully vegetated areas from ‘bald’ areas created by voles 20 and also identify intermediate stages of damage. In pest damage evaluations, different classification approaches have been used, such as vegetation index thresholds, 21 automatic classification methods 26 and supervised classifications based on machine learning, and either pixel‐oriented (each pixel is classified independently) 27 or object‐oriented methods (all pixels within defined objects are included to define spectral behavior through an iterative classification process) 28 .…”