基于深度学习的小目标受灾树木检测方法

doi:10.11707/j.1001-7488.20210310

Abstract

Abstract:

Objective: The recognition of forest images taken with the unmanned aerial vehicle (UAV) It is an the important technology using the unmanned aerial vehicle (UAV) images for pest automatic monitoring and early warning tree recognition for the automatic pest-control warning system in a forestry. An approach on object detection based on deep learning was used to accurately recognize and locates small-scale infected trees in high-resolution forest UAV images are accurately recognized and located through deep learning object detection technology. Method: The images of pine forest infected by Dendroctonus valens from Liaoning Province were chosen used as the experimental data. The UAV forest image dataset was constructed by images taken from the a height of 180-240 m and annotated by open source software LabelImg. Because the instance size of UAV forest images dataset in this paper is far smaller than general datasets, a small infected trees detector (SITD) is designed to improve the ability of recognition and localization for small object, since the instance size of UAV forest images dataset in this paper is far smaller than general datasets. Firstly, based on the framework of single shot multi-box detector (SSD), the prediction module is built from conv3_3 and tailored according to the size of the target tree in the image, then the generation method of default boxes is optimized. Secondly, the base feature map is transformed into the enhanced feature map by the feature enhance module, thus generating the base and enhanced prediction module are generated by using base feature maps and enhanced feature maps which is transferred by the feature enhance module. Finally, dual loss functions are used to facilitate the feature learning. Result: The method of small object detection based on deep learning can realize UAV remote monitoring of for forest pests, and accurately detect small and dense infected trees in UAV images accurately. In this paper, five SITD models with different structures and other object detection frameworks as SSD, FSSD and RFBNet are used to trained and tested with on UAV forest image dataset, were compared by and average precision (AP) and Precision-Recall curve are selected as evaluation indexes. The proposed method achieves the average precision of 92.62% using the UAV images captured in our study, which gets a gain of 69.71% compared to the original SSD300, and higher than the other three frameworks. Conclusion: Experimental results demonstrate that the method of small object detection based on deep learning can realize automatic detection of infected trees, and simplify the monitoring process for forest pests, and which can improve the warning ability of forest pests.

Key words: unmanned aerial vehicle imagery, forest pest, deep learning, object detection

CLC Number:

TP391.41

Yan Zhou,Wenping Liu,Youqing Luo,Shixiang Zong. Small Object Detection for Infected Trees Based on the Deep Learning Method[J]. Scientia Silvae Sinicae, 2021, 57(3): 98-107.

Figures/Tables 11

Table 1

Fig.1

Table 2

Fig.2

Fig.3

Fig.4

Table 3

Fig.5

Table 4

Fig.6

Fig.7

References 0

	费运巧, 刘文萍, 陆鹏飞, 等. 基于无人机图像分形特征的油松受灾级别判定. 计算机应用研究, 2017a, 34 (4): 1253- 1256.
	Fei Y Q , Liu W P , Lu P F , et al. Judgment on disaster classification of Chinese pine based on fractal features in UAV image. Applocation Research of Cmputers, 2017a, 34 (4): 1253- 1256.
	费运巧, 刘文萍, 骆有庆, 等. 森林病虫害监测中的无人机图像分割算法比较. 计算机工程与应用, 2017b, 53 (8): 216- 223.
	Fei Y Q , Liu W P , Luo Y Q , et al. Comparison of algorithms for unmanned aerial vehicle image segmentation in monitoring forest diseases and insect pests. Computer Engineering and Applications, 2017b, 53 (8): 216- 223.
	林志玮, 丁启禄, 涂伟豪, 等. 基于多元HoG及无人机航拍图像的植被类型识别. 森林与环境学报, 2018, 38 (4): 444- 450.
	Lin Z W , Ding Q L , Tu W H , et al. Vegetation type recognition based on multivariate HoG and aerial image captured by UAV. Journal of Forest and Environment, 2018, 38 (4): 444- 450.
	吕晓君, 王君, 喻卫国, 等. 无人机监测林业有害生物初探. 湖北林业科技, 2016, 45 (4): 30- 33. doi: 10.3969/j.issn.1004-3020.2016.04.008
	Lü X J , Wang J , Yu W G , et al. Study on monitoring forest pests and diseases by unmanned aerial vehicle. Hubei Forestry Science and Technology, 2016, 45 (4): 30- 33. doi: 10.3969/j.issn.1004-3020.2016.04.008
	吴建功, 赵明梅, 张长明, 等. 红脂大小蠹对油松的危害及越冬前后干、根部分布调查. 中国森林病虫, 2002, 21 (3): 38- 41. doi: 10.3969/j.issn.1671-0886.2002.03.016
	Wu J G , Zhao M M , Zhang C M , et al. Damage of Dendroctonus valens on Pinus tabulaeformis and its distribution on trunk and root before and after overwintering period. Forest Pest and Disease, 2002, 21 (3): 38- 41. doi: 10.3969/j.issn.1671-0886.2002.03.016
	Baeta R , Nogueira K , Menotti D , et al. Learning deep features on multiple scales for coffee crop recognition. 201730th SIBGRAPI Conference on Graphics, Patterns and Images(SIBGRAPI). Niteroi, Brazil, 2017,
	Chen L C , Papandreou G , Kokkinos I , et al. Deeplab: semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected crfs. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018, 40 (4): 834- 848. doi: 10.1109/TPAMI.2017.2699184
	Everingham M , Eslami S M A , Gool L V , et al. The pascal visual object classes challenge: A retrospective. International Journal of Computer Vision, 2015, 111 (1): 98- 136. doi: 10.1007/s11263-014-0733-5
	Gambella F , Sistu L , Piccirilli D , et al. Forest and UAV: a bibliometric review. Contemporany Engineering Sciences, 2016, 9, 1359- 1370. doi: 10.12988/ces.2016.68130
	Guirado E , Tabik S , Alcaraz-Segura D , et al. Deep-learning versus OBIA for scattered shrub detection with Google earth imagery: Ziziphus lotus as case study. Remote Sensing, 2017, 9 (12): 1220. doi: 10.3390/rs9121220
	He K , Zhang X , Ren S , et al. Deep residual learning for image recognition(CVPR). Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, America, 2016,
	Ioffe S , Szegedy C . Batch normalization: accelerating deep network training by reducing internal covariate shift. Proceedings of the 32nd International Conference on International Conference on Machine Learning, 2015, 37, 448- 456.
	Koc-San D , Selim S , Aslan N , et al. Automatic citrus tree extraction from UAV images and digital surface models using circular Hough transform. Computers and Electronics in Agriculture, 2018, 150, 289- 301. doi: 10.1016/j.compag.2018.05.001
	Li W , Fu H , Yu L , et al. Deep learning based oil palm tree detection and counting for high-resolution remote sensing images. Remote Sensing, 2016, 9 (1): 22. doi: 10.3390/rs9010022
	Li Z, Zhou F. 2017. FSSD: feature fusion single shot multibox detector. arXiv preprint arXiv: 1712. 00960. https://arxiv.org/abs/1712.00960.
	Lin T Y , Maire M , Belongie S , et al. Microsoft COCO: common objects in context. European Conference on Computer Vision(ECCV), Zurich, Switzerland, 2014,
	Liu S , Huang D . Receptive field block net for accurate and fast object detection. Proceedings of the European Conference on Computer Vision(ECCV), Munich, German, 2018,
	Liu W , Anguelov D , Erhan D , et al. SSD: Single shot multibox detector. European Conference on Computer Vision(ECCV), Amsterdam, Netherlands, 2016,
	Malek S , Bazi Y , Alajlan N , et al. Efficient framework for palm tree detection in UAV images. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2014, 7 (12): 4692- 4703. doi: 10.1109/JSTARS.2014.2331425
	Ross G . Fast R-CNN. International Conference on Computer Vision(ICCV), Santiago, Chile, 2015,
	Ren S , He K , Ross G , et al. Faster R-CNN: towards real-time object detection with region proposal networks. Los Alamitos. International Conference on Neural Information Processing Systems(NIPS), Montreal, Canada, 2015,
	Tang X , Du D K , He Z , et al. Pyramidbox: A context-assisted single shot face detector. Proceedings of the European Conference on Computer Vision (ECCV), Munich, Germany, 2018,
	Wu Y , He K . Group normalization. Proceedings of the European Conference on Computer Vision(ECCV), Munich, Germany, 2018,
	Yang S , Luo P , Loy C-C , et al. Wider face: A face detection benchmark. Computer Vision and Pattern Recognition(CVPR), Las Vegas, America, 2016,
	Zhang S , Zhu X , Lei Z , et al. S3fd: single shot scale-invariant face detector. Proceedings of the IEEE International Conference on Computer Vision, 2017,
	Zhong T , Liu W , Luo Y , et al. A new type-2 fuzzy algorithm for unmanned aerial vehicle image segmentation. International Journal of Multimedia & Ubiquitous Engineering, 2017, 12 (5): 75- 90.

器材 Equipment	参数 Parameters	数值 Value
DJI Inspire2无人机 DJI Inspire2 UAV	长×宽×高 Length×width×height(mm)	540×595×255
	最大飞行高度 Maximum flight height/m	2 500
	最大水平飞行速度 Maximum horizontal flight speed/(m·s^-1)	26
	最大飞行时间 Maximum flight time/min	≈27
DJI X5S云台相机 DJI X5S camera	长×宽×高 Length×width×height(mm)	140×98×132
	传感器 Sensor	CMOS，4/3″
	图像分辨率 Image resolution/pixels	5 280×3 956

数据集 Dataset	< 50 pixels	50~300 pixels	>300 pixels
PASCAL VOC	0.14	0.61	0.25
MSCOCO	0.43	0.49	0.08
无人机森林图像数据集 UAV forest images dataset	0.78	0.22	0

特征图 Feature maps	特征图大小 Size of feature maps	默认框大小 Size of default boxes	默认框生成步长 Stride of default boxes	默认框长宽比 Length and width ratio of default boxes	默认框数 Number of default boxes
b₁(e₁)	160×160	8(16)	2(4)	1	25 600
b₂(e₂)	80×80	16(32)	4(8)	1	6 400
b₃(e₃)	40×40	32(64)	8(16)	1	1 600
b₄(e₄)	20×20	64(128)	16(32)	1	400

模型 Model	基础特征提取器 Base feature extractor	特征增强模块 Feature enhance module	预测模块 Prediction module	平均查准率 AP(%)
0	VGG16	无 No	b₁—b₆	90.72
1	VGG16	无 No	b₁—b₄	90.42
2	VGG16	有 Yes	e₁—e₄	90.92
3	VGG16	有 Yes	b₁—b₄，e₁—e₄	92.62
4	ResNet101	有 Yes	b₁—b₄，e₁—e₄	92.29

[1]	Sifan Shen,Zhen Zhang,Xiangbo Kong,Fu Liu,Sufang Zhang. Research Techniques of Insect Odour Receptors and Their Application in Forest Insects [J]. Scientia Silvae Sinicae, 2020, 56(5): 150-159.
[2]	Ying Guo,Zengyuan Li,Erxue Chen,Xu Zhang,Lei Zhao,Yan Chen,Yahui Wang. A Deep Learning Method for Forest Fine Classification Based on High Resolution Remote Sensing Images: Two-Branch FCN-8s [J]. Scientia Silvae Sinicae, 2020, 56(3): 48-60.
[3]	Mingwei Shao,Junyu Dong. A New Algorithm for Automatic Optimizing Cross-Cut Saw Based on Deep Learning Algorithm [J]. Scientia Silvae Sinicae, 2020, 56(12): 123-129.
[4]	Song Yining, Liu Wenping, Luo Youqing, Zong Shixiang. Monitoring of Dead Trees in Forest Images Based on Linear Spectral Clustering [J]. Scientia Silvae Sinicae, 2019, 55(4): 187-195.
[5]	Liu Ping, Liang Qianling, Chen Meng, Liu Zhongjun, Sheng Zhaohu, Su Yanle. Construction on Index System of Forest pest Disaster Loss Assessment [J]. Scientia Silvae Sinicae, 2016, 52(6): 101-107.
[6]	Zhan Maokui, Yang Zhongqi, Wang Xiaoyi, Lai Yanxue, Zhang Yifeng. Capacity of Transmitting Bursaphelenchus xylophilus by the Vector Monochamus alternatus Adults [J]. Scientia Silvae Sinicae, 2014, 50(7): 74-81.
[7]	Chang Yuanfei;Wu Honggan;Dong Zhenhui;Cui Fudong;Liu Li;Jian Jun;Zhang Ying;Li Weiguo. A National Monitoring and Warning System for Forest Pest Based on Service-Oriented Architecture [J]. Scientia Silvae Sinicae, 2011, 47(6): 93-100.
[8]	Wu Yaojun;Jiang Xuejian;Li Dewei;Luo Jitong;Zhou Guofu;Chang Mingshan;Yang Zhongqi. Leptocybe invasa, a New Invasive Forest Pest Making Galls on Twigs and Leaves of Eucalyptus Trees in China (Hymenoptera：Eulophidae) [J]. Scientia Silvae Sinicae, 2009, 12(7): 161-163.
[9]	Hu Haiqing;Guo Futao. Influence of Fire on the Occurrence of Forest Pest [J]. Scientia Silvae Sinicae, 2006, 42(12): 77-82.
[10]	Liu Huixiang;Zhang Xingyao. Plant Protease Inhibitors and Their Application on Forest Tree Resisting Pest Genetic Engineering [J]. Scientia Silvae Sinicae, 2005, 41(3): 148-157.
[11]	Zhao Tieliang;Dong Zhenhui;Yu Zhijun;Zhao Qingshan;Jiang Fengyan. STUDY ON ESTABLISHMENT OF FOREST PEST INDEXES [J]. Scientia Silvae Sinicae, 2003, 39(3): 172-176.
[12]	Zhang Zaifu. STUDIES ON THE OPERATION IN CONTROL OF FOREST INSECT PESTS AND DISEASES BY (SUPER) LIGHT PLANE [J]. Scientia Silvae Sinicae, 2000, 36(3): 81-86.
[13]	Zaiping Lan,Hong Cheng,Jinkun Wu,Xianfeng Su. A COMPREHENSIVE ANALYSIS ON FIXED FOREST PEST AND DISEASES, MONITORING FINDINGS IN NEWLY ESTABLISHED HIGH-YIELDING TIMBER PLANATATIONS OF MAIN SPECIES [J]. Scientia Silvae Sinicae, 1999, 35(zk): 128-142.
[14]	Xianqing Xue,Taifeng Chen,Lin Xia,Qingyuan Lin,Lijin Wang. COMMUNICATION THROUGH COMPUTER NETWORK IN PREDICTION OF FOREST PESTS [J]. Scientia Silvae Sinicae, 1996, 32(6): 528-535.

Small Object Detection for Infected Trees Based on the Deep Learning Method

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 11

References 0

Related Articles 14

Recommended Articles

Metrics

Comments