Abstract
Effective prediction of crop pest species and quantities can provide a priori knowledge for pest control. However, there are many kinds of pests and the color and texture of different pests are similar, traditional image processing and recognition methods based on time or frequency domain features and classifiers are hard to meet the needs of high-precision recognition. In order to improve the recognition result of agricultural pests, based on the target detection algorithm Faster RCNN, we put forward a method for the classification and identification of pests on the basis of online hard sample mining and residual network. The main work is as follows: (1) According to the different size of pests, the detection ability of the network for small targets is enhanced by changing the size and number of anchors; (2) In the feature extraction stage, we use ResNet-50 with stronger feature extraction ability rather than the original VGG16 network; (3) In order to solve the imbalance of samples, online hard Sample Mining Strategy is introduced in the training stage. The experimental outcome demonstrate that our method is effective, the improved algorithm achieves about 5% improvement in accuracy, and the network has a stronger ability to detect small targets.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Yang, G.G.: Identification and detection of early alfalfa based on machine vision. Zhejiang University (2017)
Han, A.T., Guo, X.H., Liao, Z., et al.: A classification method of agricultural pests based on compressed sensing theory. Trans. Chin. Soc. Agric. Eng. 27(06), 203–207 (2011)
Zhang, M.H., Li, J.H.: Pest identification algorithm based on deep learning and sparse representation. J. Henan Univ. 48(02), 207–221 (2018)
Han, R.Z.: Research on rapid detection and identification of farmland pests based on machine vision. Zhejiang University (2014)
Girshick, R.: Fast R-CNN. In: IEEE International Conference on Computer Vision (2015)
Ren, S., He, K., Girshick, R., et al.: Faster R-CNN: towards real-time object detection with region proposal networks. IEEE Trans. Pattern Anal. Mach. Intell. 39(6), 1137–1149 (2015)
Redmon, J., Divvala, S., Girshick, R., et al.: You only look once: unified, real-time object detection. In: Proceeding of IEEE Conference on Computer Vision and Pattern Recognition (2016)
Liu, W., Anguelov, D., Erhan, D., et al.: Single shot multibox detector. In: Proceeding of European Conference on Computer Vision, Springer, pp. 21–37 (2016)
Xu, S.K., Wang, Y.R., Gu, Y.W., et al.: Research on helmet wearing detection based on improved Faster RCNN. Comput. Appl. Res. 37(3), 1–6 (2019)
Cainiaohudi. https://blog.csdn.net/cainiaohudi/article/details/80995799 (2018)
Shi, W.L.: Research on target detection algorithm based on deep learning. Dalian Maritime University (2018)
Shrivastava, A., Gupta, A., Girshick, R.: Training region-based object detectors with online hard example mining. In: Proceeding of IEEE Conference on Computer Vision and Pattern Recognition, pp. 761–769 (2016)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Zhang, M., Chen, Y., Zhang, B., Pang, K., Lv, B. (2020). Recognition of Pest Based on Faster RCNN. In: Wang, Y., Fu, M., Xu, L., Zou, J. (eds) Signal and Information Processing, Networking and Computers. Lecture Notes in Electrical Engineering, vol 628. Springer, Singapore. https://doi.org/10.1007/978-981-15-4163-6_8
Download citation
DOI: https://doi.org/10.1007/978-981-15-4163-6_8
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-15-4162-9
Online ISBN: 978-981-15-4163-6
eBook Packages: EngineeringEngineering (R0)