Agricultural Greenhouses Detection in High-Resolution Satellite Images Based on Convolutional Neural Networks: Comparison of Faster R-CNN, YOLO v3 and SSD

Min Li,Xudong Guo,Zhijie Zhang,Xiaofan Wang,Liping Lei

doi:10.3390/s20174938

Abstract

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.

Highlights

Agricultural greenhouses (AGs) are an important technique in modern agriculture to satisfy the human demand for farm products [1,2,3,4]
Precision-recall curve (PRC) is composed of precision (P) and recall (R). It is a more conventional and objective judgment criterion in the field of object detection compared with individual precision or recall metric
The best results were obtained with the You Look Only Once-v3 (YOLO v3) network according to mean average precision (mAP)

Summary

Introduction

Agricultural greenhouses (AGs) are an important technique in modern agriculture to satisfy the human demand for farm products [1,2,3,4]. The vigorous construction and expansion of AGs, has induced many issues in land management, such as occupied high-quality cultivated land [6], damaged soil [7], pollution of plastic wastes [4] etc. To address these problems, we need an effective detection method to monitor the spatial distribution of AGs, in order to reasonably develop AGs and protect the cultivated land [8]. AGs detection was mainly achieved by visual interpretation with remote sensing images in government management This method, depending on the level of expertise, is time-consuming and far from a large-scale automatic AGs detection requirement

Objectives

Results

Conclusion