Agricultural Greenhouses Detection in High-Resolution Satellite Images Based on Convolutional Neural Networks: Comparison of Faster R-CNN, YOLO v3 and SSD.
Agricultural Greenhouses Detection in High-Resolution Satellite Images Based on Convolutional Neural Networks: Comparison of Faster R-CNN, YOLO v3 and SSD.
复制标题
DOI：
10.3390/s20174938
复制
发表时间：
2020-08-31
期刊：
影响因子：
--
通讯作者：
Guo X
中科院分区：
文献类型：
--
作者：
Li M;Zhang Z;Lei L;Wang X;Guo X
农业大棚是现代农业发展的重要设施。准确有效地检测AGs是现代农业战略规划的必要条件。随着深度学习算法的出现，各种基于卷积神经网络（CNN）的模型被提出用于高空间分辨率图像的目标检测。在本文中，我们对三种成熟的基于cnn的模型进行了比较评估，即Faster R-CNN， You Look Only one -v3 （YOLO v3）和Single Shot Multi-Box Detector (SSD)，用于检测AGs。采用迁移学习和微调方法对模型进行训练。从平均精度（mAP）、每秒帧数（FPS）和视觉检测等方面对YOLO v3的精度和效率进行了评价。SSD在检测速度方面表现出优势，FPS比Faster R-CNN高两倍，尽管他们的mAP在测试集上接近。将训练好的模型应用于两个独立的测试集，证明了这些模型具有一定的可移植性，高分辨率的图像对精度的提高是显著的。我们的研究表明，YOLO v3在精度和计算效率上都具有优势，可以应用于高分辨率卫星图像的AGs检测。
Agricultural greenhouses (AGs) are an important facility for the development of modern agriculture. Accurately and effectively detecting AGs is a necessity for the strategic planning of modern agriculture. With the advent of deep learning algorithms, various convolutional neural network (CNN)-based models have been proposed for object detection with high spatial resolution images. In this paper, we conducted a comparative assessment of the three well-established CNN-based models, which are Faster R-CNN, You Look Only Once-v3 (YOLO v3), and Single Shot Multi-Box Detector (SSD) for detecting AGs. The transfer learning and fine-tuning approaches were implemented to train models. Accuracy and efficiency evaluation results show that YOLO v3 achieved the best performance according to the average precision (mAP), frames per second (FPS) metrics and visual inspection. The SSD demonstrated an advantage in detection speed with an FPS twice higher than Faster R-CNN, although their mAP is close on the test set. The trained models were also applied to two independent test sets, which proved that these models have a certain transability and the higher resolution images are significant for accuracy improvement. Our study suggests YOLO v3 with superiorities in both accuracy and computational efficiency can be applied to detect AGs using high-resolution satellite images operationally.
登录
查看更多内容
DOI：
10.1109/tgrs.2016.2612821
发表时间：
2017-02-01
影响因子：
8.2
作者：
Maggiori, Emmanuel;Tarabalka, Yuliya;Alliez, Pierre
通讯作者：
Alliez, Pierre
使用深度图和姿势进行人类动作识别的深度卷积神经网络
DOI：
10.1109/tsmc.2018.2850149
发表时间：
2019-09-01
影响因子：
8.7
作者：
Kamel, Aouaidjia;Sheng, Bin;Feng, David Dagan
通讯作者：
Feng, David Dagan
影响因子：
2.3
作者：
Chen, Chaoyue;Gong, Weiguo;Li, Weihong
通讯作者：
Li, Weihong
学习用于 VHR 光学遥感图像中目标检测的旋转不变卷积神经网络
DOI：
10.1109/tgrs.2016.2601622
发表时间：
2016-12-01
影响因子：
8.2
作者：
Cheng, Gong;Zhou, Peicheng;Han, Junwei
通讯作者：
Han, Junwei
DOI：
10.1016/j.isprsjprs.2017.05.002
发表时间：
2017-08-01
影响因子：
12.7
作者：
Alshehhi, Rasha;Marpu, Prashanth Reddy;Dalla Mura, Mauro
通讯作者：
Dalla Mura, Mauro