Laser & Optoelectronics Progress, Volume. 55, Issue 1, 11502(2018)
Recognition of Empoasca Flavescens Based on Machine Vision
Figures & Tables(9)
Fig. 4. Segmentation results using different algorithms. (a) Original image; (b) traditional Ostu algorithm; (c) K_means clustering algorithm
Fig. 5. DBSCAN clustering results. (a) Setting threshold to 8; (b) setting threshold to 12; (c) clustering fusion result
Fig. 6. Accuracy versus number of selecting samples for different algorithms. (a) KS algorithm; (b) random selection
Fig. 7. Partial recognition results. (a) Correct recognition; (b) misrecognition; (c) leakage recognition
Table 1. Number of Empoasca flavescens separated by different algorithms
View tableTable 1. Number of Empoasca flavescens separated by different algorithms
Total number Number Ostu K_means SLIC+DBSCAN8+ DBSCAN12 204 164 187 200
Table 2. The test accuracy obtained by different training methods%
View tableTable 2. The test accuracy obtained by different training methods%
Algorithm Percent_test TPR TNR Precision G-mean F-measure LSSVM 98.81 86.0 99.45 91.29 92.47 88.53 LSSVM+SMOTE 99.01 98.2 99.07 87.54 98.63 92.56 LSSVM+SMOTE+KS 99.30 97.5 99.42 91.76 98.46 94.54
Tools
Get Citation
Copy Citation Text
Chen Jing, Zhu Qibing, Huang Min, Zheng Yang. Recognition of Empoasca Flavescens Based on Machine Vision[J]. Laser & Optoelectronics Progress, 2018, 55(1): 11502
Paper Information
Category: Machine Vision
Received: Jul. 17, 2017
Accepted: --
Published Online: Sep. 10, 2018
The Author Email: Zhu Qibing (zhuqib@163.com)
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20