當今社會就業人口已從一級產業發展至二、三級產業，農業人口逐漸流失不僅人力資源不足外，更是引發平均每人工時過長、土地面積過大，以及財力資源分配不均等等因素，導致生產力下降，但糧食仍然是維持國家發展的重要因素之一。由於國家發展重點轉移，農業就業人口不足的問題越發仰賴機械和自動化機器來解決，雖然仍有些農業技術是機器無法解決的，必須依靠有農業相關的知識和背景的人員來進行，舉例來說：水稻收割、播種、移栽等工作都可以用機器來完成，但對農作物的監測和病蟲害的防治卻無法用機器來完成的。最常見的防治病蟲害的方法就是噴灑農藥，這種方法能有效快速地防治害蟲、病毒、細菌等入侵農作物，而相對會損害消費者的健康，為了瞭解農作物的生長情況和害蟲的種類和數量，本研究使用無人機飛行載具及手持相機裝置來輔助獲取農地資訊之影像。
本研究將取得的影像透過影像處理後，做深度學習之影像辨識，應用類神經網路檢測出害蟲的位置和種類，來有效控制並分析害蟲種類及分布狀況。由於使用無人機飛行載具在害蟲得影像上很少且檢測不易，因此本研究以特徵金字塔網路（Feature Pyramid Network, FPN）概念做為基礎來建構網路架構，這是一種用於小物體檢測的類神經網路架構。其特徵金字塔利用深度卷積神經網路固有的多尺度金字塔層次結構，透過橫向連接與從上而下的網路架構做結合，達成在多尺度上構建具有豐富語意及高解析度影像層的特徵圖。而後使用候選區域網路(Region Proposal Network, RPN)和全連接層（Fully Connected, FC）取得物體種類及位置。
本研究可分為兩大部分：類神經網路模型架構及影像辨識衡量指標。將類神經網路在不同模型架構組合下做探討，藉由交併比（Intersection over Union, IoU）和混淆矩陣（Confusion matrix）的計算資料集的精確度（Precision）、召回率（Recall）及平均精確度（mean Average Precision, mAP）等等評估值。

Due to the transfer of national development priorities, the problem of insufficient employment in agriculture needs to be solved by machinery and automated machinery. To understand the growth of crops and the type and quantity of pests, image recognition and image processing are proposed to accomplish this purpose.
In this research, object detection of deep learning is performed with image processing. The neural network is applied to detect the location and type of pests to effectively control and analyze the type and distribution of pests. In this research, Feature Pyramid Network (FPN) is utilized to construct the network architecture that performs well for small object detection. Then, region proposal network (RPN) and fully connected layer (FC) are applied to obtain the class and location of the objects.
This research can be divided into two parts: neural network model architecture and object detection performance indicator. The neural networks are discussed with different combinations in model architectures. The performance is indicated with the precision, recall, mean average precision (mAP), and other evaluation values by calculating the intersection over union (IoU).

K. He and J. Sun. Convolutional neural networks at constrained time cost. In CVPR, 2015.
R. K. Srivastava, K. Greff, and J. Schmidhuber. Highway networks. arXiv:1505.00387, 2015.
Sergey Ioffe. Batch Renormalization: Towards Reducing Minibatch Dependence in Batch-Normalized Models. arXiv:1702.03275v2, 2017.
Karen Simonyan, Andrew Zisserman. Very deep convolutional networks for large-scale image recognition. arXiv:1409.1556v6, 2015.
Ross Girshick, Jeff Donahue, Trevor Darrell, Jitendra Malik. Rich feature hierarchies for accurate object detection and semantic aegmention. arXiv:1311.2524v5, 2014.
Ross Girshick. Fast R-CNN. arXiv:1504.08083v2, 2015.
Shaoqing Ren, Kaiming He, Ross Girshick, and Jian Sun. Faster R-CNN: towards real-time object detection with region proposal networks. arXiv:1506.01497v3, 2016.
Kang Tong, Yiquan Wu, Fei Zhou. Recent advances in small object detection based on deep learning: A review. Journal of Image and Vision Computing. 103910, 2020.
Kaiming He, Xiangyu Zhang, Shaoqing Ren, Jian Sun. Deep residual learning for image recognition. arXiv:1512.03385v1, 2015.
Abhinav Shrivastava, Rahul Sukthankar, Jitendra Malik, Abhinav Gupta. Beyond Skip Connections: Top-Down Modulation for Object Detection. arXiv:1612.06851v2, 2017.
Tsung-Yi Lin, Piotr Doll´ar, Ross Girshick, Kaiming He, Bharath Hariharan, Serge Belongie. Feature Pyramid Networks for Object Detection. arXiv:1612.03144v2, 2017.
Joseph Redmon, Santosh Divvala, Ross Girshick, Ali Farhadi. You Only Look Once: Unified, Real-Time Object Detection. arXiv:1506.02640v5, 2016.
Joseph Redmon, Ali Farhadi. YOLO9000: Better, Faster, Stronger. arXiv:1612.08242v1, 2016.
Joseph Redmon, Ali Farhadi. YOLOv3: An Incremental Improvement. arXiv:1804.02767,2018.
Wei Liu, Dragomir Anguelov, Dumitru Erhan, Christian Szegedy, Scott Reed, Cheng-Yang Fu, Alexander C. Berg. SSD: Single Shot MultiBox Detector. arXiv:1512.02325v5, 2016.
Z. Cai, Q. Fan, R. S. Feris, and N. Vasconcelos. A unified multi-scale deep convolutional neural network for fast object detection. In ECCV, 2016.
P. Doll´ar, R. Appel, S. Belongie, and P. Perona. Fast feature pyramids for object detection. TPAMI, 2014.
S. Gidaris and N. Komodakis. Object detection via a multiregion & semantic segmentation-aware CNN model. In ICCV, 2015.
G. Ghiasi and C. C. Fowlkes. Laplacian pyramid reconstruction and refinement for semantic segmentation. In ECCV, 2016.
R. Girshick, J. Donahue, T. Darrell, and J. Malik. Rich feature hierarchies for accurate object detection and semantic segmentation. In CVPR, 2014.
R. Girshick. Fast R-CNN. In ICCV, 2015.
T. Kong, A. Yao, Y. Chen, and F. Sun. Hypernet: Towards accurate region proposal generation and joint object detection. In CVPR, 2016.
Krizhevsky, A., Sutskever, I., & Hinton, G.E. Imagenet classification with deep convolution neural networks. Advances in Neural Information Processing Systems 25, 1097-1105, 2012.
David G. Lowe. Distinctive Image Features from Scale-Invariant Keypoints. International Journal of Computer Vision, 2004.
Herbert Bay, Andreas Ess, Tinne Tuytelaars, and Luc Van Gool. Speeded-Up Robust Features (SURF). European conference on Computer Vision, 2008.
Ross Girshick, Jeff Donahue, Trevor Darrell, Jitendra Malik. Rich feature hierarchies for accurate object detection and segmentation. arXiv: 1311.2524v5, 2014.
Christian Szegedy, Alexander Toshev, Dumitru Erhan. Deep Neural Networks for Object Detection. Neural Information Processing Systems, 2013.
Mohamed Esmail Karar, Fahad Alsunaydi, Sultan Albusaymi, Sultan Alotaibi. A new mobile application of agricultural pestsrecognition using deep learning in cloud computingsystem. Alexandria Engineering Journal, 2021.
陳淑佩、高靜華。2018。移動式設施害蟲影像辨識系統。農業試驗所技術服務季刊。
洪嘉鎂。2018。農委會將訂定無人機適用農藥實驗，將有助農藥減量。農傳媒。
甯格致。2016。精準農業生產與小型無人飛行器的應用潮流概述。農業資訊科技應用發展電子報 vol.39。
江彥政。2019年。無人機在農業調查之應用。嘉大農林學報第十六卷第二期。