Biosensor-Based Drones Anomaly Detection Integration for Sustainable Agriculture Development

Abstract

Recently, biosensor-based drones (BBD) have emerged and have proven to be highly influential in the convergence of modern technology and agriculture. These drones possess the capacity to bring about significant changes in the realm of sustainable agriculture development. This study presents a comprehensive insight into the crucial element of anomaly detection (AD) in integrating BBD and investigates their diverse uses in promoting sustainability in agriculture. With various advanced sensor technologies, BBD collects and transmits real-time data for accurate monitoring of agricultural crops, soil quality, and environmental parameters. The utilization of a diverse range of sensors, including multispectral, hyperspectral, thermal infrared, global positioning system (GPS), light detection and ranging (LiDAR), environmental, chemical, and crop health sensors, provides farmers with the capability to make informed decisions based on data. The management of extensive datasets produced by these sensors presents a considerable obstacle. The utilization of AD techniques is crucial to exploit the capabilities of drones equipped with biosensors fully. Machine learning (ML) algorithms and artificial intelligence (AI) systems significantly impact the processing and interpretation of sensor data. They are essential in detecting deviations from anticipated trends and notifying farmers of abnormalities that could indicate crop stress, illnesses, or pest presence. Detecting issues early enables prompt action, decreasing crop production losses and reducing reliance on chemical treatments. This, in turn, supports the adoption of sustainable farming methods. The utilization of BBD in advancing sustainable agriculture encompasses a wide range of applications. The practices encompass precision irrigation management, targeted fertilization, disease and insect control, land optimization, and minimization of environmental impact. These applications collectively enhance resource efficiency, augment agricultural yields, mitigate environmental impact, and promote sustainable agriculture. Selected studies were obtained from six top academic research databases. The authors used an exhaustive data extraction technique, focusing on the study objectives and type of AD in a smart operation, such as smart agriculture, transportation and smart things settings. According to this analysis, several studies have shown that deep learning (DL) and ML are more employed in preventing point and collective anomalies. Statistical approaches are more applicable in contextual and collaborative AD. The study presents an AD summary of ML, DL, and statistical-based approaches. Conclusively, the study identifies AD's future research directions from a drone operations perspective.