Laser & Optoelectronics Progress, Volume. 61, Issue 15, 1530002(2024)
Identification of Types of Tobacco Leaf Diseases Using Near-Infrared Spectroscopy and Random Forest Algorithm
Figures & Tables(10)
Fig. 1. Spectral scanning position of tobacco leaf sample. (a) Healthy; (b) powdery mildew; (c) deep green; (d) mosaic virus; (e) brown-spot
Fig. 2. Process of forming a random forest (N is the number of samples and K is the number of sampling sets)
Fig. 3. Original spectra and the spectra after pretreatment of different disease species of tobacco leaf. (a) Original spectra; (b) spectra after pretreatment
Fig. 4. ROC curves of different disease species based on random forest algorithm.(a) Healthy; (b) powdery mildew; (c) deep green; (d) mosaic virus; (e) brown-spot
Table 1. Samples of different types of tobacco leaf
View tableTable 1. Samples of different types of tobacco leaf
Disease type Year Regional source Spectral dimension /nm Total sample Training sample Test sample Healthy 2022 Longtankou village，Fumin county，Yunnan province 908‒1676 80 44 36 Powdery mildew 2022 Longtankou village，Fumin county，Yunnan province 908‒1676 100 64 36 Deep green 2022 Longtankou village，Fumin county，Yunnan province 908‒1676 80 48 28 Mosaic virus 2022 Longtankou village，Fumin county，Yunnan province 908‒1676 120 76 44 Brown-spot 2022 Longtankou village，Fumin county，Yunnan province 908‒1676 80 44 36
Table 2. Comparison of SG+D1 pre-processing methods based on SVM under different parameters
View tableTable 2. Comparison of SG+D1 pre-processing methods based on SVM under different parameters
Algorithm Preprocessing method Smoothing point Polynomial order Training set accuracy /% Test set accuracy /% SVM SG+D1 1 2 71.37 73.18 SG+D1 3 2 70.40 68.42 SG+D1 5 2 67.52 71.24 SG+D1 7 2 66.38 69，87 SG+D1 9 2 70.46 69.63
Table 3. Different pre-processing methods accuracies based on SVM algorithm
View tableTable 3. Different pre-processing methods accuracies based on SVM algorithm
Algorithm Pre-processing method Training set accuracy /% Test set accuracy /% SVM SNV 63.76 64.13 MSC 69.92 67.93 SG 70.12 71.30 SG+D1 71.37 73.18
Table 4. Comparison model effects of training models for different species of tobacco leaf disease test
View tableTable 4. Comparison model effects of training models for different species of tobacco leaf disease test
Algorithm Type Training sample Training correct number Number of false positive Precision /% Sensitivity /% Specificity /% F1-score OAC /% SVM Healthy 51 50 1 98.1 100.0 99.6 0.99 74.28 Powdery mildew 35 28 7 81.4 55.6 96.2 0.66 Deep green 23 7 6 71.9 53.5 96.1 0.62 Mosaic virus 43 22 21 51.8 66.2 85.8 0.58 Brown-spot 53 43 10 80.3 98.1 94.1 0.88 BP neural network Healthy 43 39 4 91.3 97.7 98.3 0.94 91.67 Powdery mildew 59 57 2 96.4 91.5 99.1 0.94 Deep green 52 45 7 85.7 80.8 96.9 0.83 Mosaic virus 70 67 3 95.6 92.9 98.5 0.94 Brown-spot 52 45 7 87.7 96.2 96.9 0.92 PLS-DA Healthy 48 48 0 100.0 100.0 100.0 1.00 99.63 Powdery mildew 60 60 0 100.0 100.0 100.0 1.00 Deep green 48 47 1 98.0 100.0 99.5 0.98 Mosaic virus 71 71 0 100.0 98.6 100.0 0.99 Brown-spot 48 48 0 100.0 100.0 100.0 1.00 RF Healthy 44 44 0 100.0 100.0 100.0 1.00 100.00 Powdery mildew 64 64 0 100.0 100.0 100.0 1.00 Deep green 48 48 0 100.0 100.0 100.0 1.00 Mosaic virus 76 76 0 100.0 100.0 100.0 1.00 Brown-spot 44 44 0 100.0 100.0 100.0 1.00
Table 5. Comparison model effects of test models for different species of tobacco leaf disease
View tableTable 5. Comparison model effects of test models for different species of tobacco leaf disease
Algorithm Type Test sample Number of correct test Number of incorrect test Precision /% Sensitivity /% Specificity /% F1-score OAC /% SVM Healthy 29 29 0 100.0 100.0 100.0 1.00 67.93 Powdery mildew 16 12 4 76.2 43.2 96.6 0.55 Deep green 18 11 5 72.0 48.6 95.2 0.58 Mosaic virus 38 21 17 55.9 69.1 76.7 0.62 Brown-spot 24 14 10 58.5 92.3 89.2 0.72 BP neural network Healthy 37 32 5 88.1 100.0 96.6 0.94 85.87 Powdery mildew 41 38 3 91.9 82.9 97.9 0.87 Deep green 28 22 6 80.0 57.1 97.4 0.67 Mosaic virus 50 45 5 89.6 86.0 96.3 0.88 Brown-spot 28 21 7 75.7 100.0 94.2 0.87 PLS-DA Healthy 33 32 1 97.0 100.0 99.3 0.98 94.00 Powdery mildew 40 30 1 97.5 97.5 99.3 0.97 Deep green 37 30 7 81.2 93.8 95.3 0.87 Mosaic virus 41 40 1 97.6 83.8 99.2 0.90 Brown-spot 33 32 1 97.0 100.0 99.3 0.98 RF Healthy 36 36 0 100.0 100.0 100.0 1.00 98.10 Powdery mildew 36 35 1 97.3 100.0 100.0 0.99 Deep green 28 28 0 100.0 87.5 97.4 0.93 Mosaic virus 44 42 2 95.7 100.0 100.0 0.98 Brown-spot 36 35 1 97.3 100.0 100.0 0.99
Table 6. AUC values of different classification algorithms
View tableTable 6. AUC values of different classification algorithms
Algorithm Healthy Powdery mildew Deep green Mosaic virus Brown-spot SVM 1.00 0.75 0.74 0.69 0.73 BP neural network 0.97 0.87 0.62 0.89 0.96 PLS-DA 0.98 0.90 0.94 0.89 0.95 RF 1.00 0.91 0.96 0.99 0.99
Tools
Get Citation
Copy Citation Text
Ying Liang, Kun Ma, Xinyu Zhang, Qifu Yang, Jiaquan Wu. Identification of Types of Tobacco Leaf Diseases Using Near-Infrared Spectroscopy and Random Forest Algorithm[J]. Laser & Optoelectronics Progress, 2024, 61(15): 1530002
Paper Information
Category: Spectroscopy
Received: Jun. 5, 2023
Accepted: Aug. 8, 2023
Published Online: Aug. 12, 2024
The Author Email: Jiaquan Wu (710866288@qq.com)
CSTR:32186.14.LOP231466
© Copyright 2018-2021 | Chinese Laser Press.
All Rights Reserved 沪ICP备15018463号-20