Video analysis for the detection of animals using convolutional neural networks and consumer-grade drones
Video analysis for the detection of animals using convolutional neural networks and consumer-grade drones
复制标题
DOI：
10.1139/juvs-2020-0018
复制
发表时间：
2021-06-01
影响因子：
2.3
通讯作者：
Wich, Serge A.
中科院分区：
文献类型：
--
作者：
Chalmers, C.;Fergus, P.;Wich, Serge A.
确定动物的分布和密度在保护中很重要。这一过程既费时又费力。无人机已被用于帮助减轻人力密集型任务，在更短的时间内覆盖大的地理区域。在本文中，我们进一步研究了这个想法，使用概念验证来从无人机镜头中检测犀牛和汽车。概念验证利用了现成的技术和消费级无人机硬件。该研究证明了使用机器学习（ML）自动执行常规保护任务的可行性，例如动物检测和跟踪。该原型是使用DJI Mavic Pro 2开发的，并在全球移动的通信系统（GSM）网络上进行了测试。Faster-RCNN Resnet 101架构用于迁移学习。使用帧采样技术执行推断，以解决精度、处理速度和实时视频馈送同步之间所需的权衡。推理模型托管在Web平台上，来自无人机的视频流（使用OcuSync）传输到实时消息传递协议（RTMP）服务器，用于后续分类。在训练过程中，最佳模型的平均精度（mAP）分别为0.83交集大于并集（@IOU）0.50和0.69@IOU 0.75。在诺斯利Safari中测试系统时，我们的原型能够实现以下目标：灵敏度（Sen），0.91（0.869，0.94）;特异性（Spec），0.78（0.74，0.82）;和准确度（ACC），0.84（0.81，0.87），当检测犀牛时，Sen，1.00（1.00，1.00）; Spec，1.00（1.00，1.00）;当检测汽车时，ACC，1.00（1.00，1.00）。
Determining animal distribution and density is important in conservation. The process is both time-consuming and labour-intensive. Drones have been used to help mitigate human-intensive tasks by covering large geographical areas over a much shorter timescale. In this paper we investigate this idea further using a proof of concept to detect rhinos and cars from drone footage. The proof of concept utilises off-the-shelf technology and consumer-grade drone hardware. The study demonstrates the feasibility of using machine learning (ML) to automate routine conservation tasks, such as animal detection and tracking. The prototype has been developed using a DJI Mavic Pro 2 and tested over a global system for mobile communications (GSM) network. The Faster-RCNN Resnet 101 architecture is used for transfer learning. Inference is performed with a frame sampling technique to address the required trade-off between precision, processing speed, and live video feed synchronisation. Inference models are hosted on a web platform and video streams from the drone (using OcuSync) are transmitted to a real-time messaging protocol (RTMP) server for subsequent classification. During training, the best model achieves a mean average precision (mAP) of 0.83 intersection over union (@IOU) 0.50 and 0.69 @IOU 0.75, respectively. On testing the system in Knowsley Safari our prototype was able to achieve the following: sensitivity (Sen), 0.91 (0.869, 0.94); specificity (Spec), 0.78 (0.74, 0.82); and an accuracy (ACC), 0.84 (0.81, 0.87) when detecting rhinos, and Sen, 1.00 (1.00, 1.00); Spec, 1.00 (1.00, 1.00); and an ACC, 1.00 (1.00, 1.00) when detecting cars.