Improved UAV Opium Poppy Detection Using an Updated YOLOv3 Model

Abstract

Rapid detection of illicit opium poppy plants using UAV (unmanned aerial vehicle) imagery has become an important means to prevent and combat crimes related to drug cultivation. However, current methods rely on time-consuming visual image interpretation. Here, the You Only Look Once version 3 (YOLOv3) network structure was used to assess the influence that different backbone networks have on the average precision and detection speed of an UAV-derived dataset of poppy imagery, with MobileNetv2 (MN) selected as the most suitable backbone network. A Spatial Pyramid Pooling (SPP) unit was introduced and Generalized Intersection over Union (GIoU) was used to calculate the coordinate loss. The resulting SPP-GIoU-YOLOv3-MN model improved the average precision by 1.62% (from 94.75% to 96.37%) without decreasing speed and achieved an average precision of 96.37%, with a detection speed of 29 FPS using an RTX 2080Ti platform. The sliding window method was used for detection in complete UAV images, which took approximately 2.2 sec/image, approximately 10× faster than visual interpretation. The proposed technique significantly improved the efficiency of poppy detection in UAV images while also maintaining a high detection accuracy. The proposed method is thus suitable for the rapid detection of illicit opium poppy cultivation in residential areas and farmland where UAVs with ordinary visible light cameras can be operated at low altitudes (relative height < 200 m).

Detection of sitting posture using hierarchical image composition and deep learning

“…At this stage, we use Long Short Term Memory (LSTM) convolutional block ( Xu et al, 2020 ; Li et al, 2020 ), which is tasked to extract 32 most useful features in the entire sequence. For our main neural network backbone, we use MobileNetV2 , which is the extension of MobileNet , for it has show to achieve great results in predictive capabilities ( Howard et al, 2017 ; Zhou et al, 2019 ), however, the architecture itself is relatively light-weight for it is designed to be used in low power devices such as mobile devices, unlike for example, YOLOV3 , which while having impressive recall results ( Redmon & Farhadi, 2018 ), is much more complex and has a substantially poorer performance. The MobileNetV2 output is then connected to a global average pooling layer in order to reduce dimensionality and improve generalization rate ( Zhou et al, 2016 ).…”

A UAV-Based Framework for Semi-Automated Thermographic Inspection of Belt Conveyors in the Mining Industry

“…Other works linking thermography and UAV are focused on diverse applications, like crop management [29,30], power line [31,32] monitoring, and solar power plants inspection [33][34][35]. The works of [36,37] are examples of applications that use drones and machine learning algorithms to aid inspection processes. Next, we present the details and components of our solution.…”

“…You Only Look Once (YOLO) [51] is a CNN-based method for object detection. It presents excellent results in terms of precision and is widely used in inspection scenarios [36,37], but requires a lot of computing power for real-time execution. The Aggregated Channel Features (ACF) method [52], which can be view as an evolution of the classical method of Boosted Cascade of Simple Features proposed by Viola-Jones [47], is another important machine learning technique for object detection.…”

A Machine Learning Method for Vision-Based Unmanned Aerial Vehicle Systems to Understand Unknown Environments

“…This work focuses more on special object detection for UAVs. Zhou et al [14] used an updated YOLO v3 model to detect the opium poppies in images captured by an UAV. Compared with original YOLO v3 model the model uses the recently proposed Generalized Intersection Over Union (GIOU as the loss function, and a Spatial Pyramid Pooling Unit is added [21], while, the method is run on a RTX2080Ti platform, which means the detection process is offline and could not benefit the automatic control immediately for UAVs in an unknown environment.…”

“…We address the understanding method with this property as onlined in this paper. In contrast, the understanding method to extract the information is offline if the UAV has already finished the aerial mission and the processing result has no influence on the mission (e.g., filming an area [14]). The online method to extract the information from real-time images has wider application prospects than the offline one.…”