Skip to main content

Table 2 Summary of techniques successfully used to detect drought and diseases in plants

From: Hyperspectral image analysis techniques for the detection and classification of the early onset of plant disease and stress

Technique Plant (stress) References Accuracy
Quadratic discriminant analysis (QDA) Wheat (yellow rust)
Avacado (laurel wilt)
[26]
[33]
92%
94%
Decision tree (DT) Avacado (laurel wilt)
Sugarbeet (cerospora leaf spot)
Sugarbeet (powdery mildew)
Sugarbeet (leaf rust)
[33]
[41]
95%
95%
86%
92%
Multilayer perceptron (MLP) Wheat (yellow rust) [27] 98.9/99.4%
H/D
Partial least square regression (PLSR)
 Raw
 Savitsky-Golay 1st derivative
 Savitsky-Golay 2nd derviative
Celery (sclerotinia rot) [35] 88.92%
88.18%
86.38%
Partial least square regression (PLSR)
 Fishers linear determinant analysis
Wheat (yellow rust)
Wheat (aphid)
Wheat (powdery mildew)
Wheat (powdery mildew)
[36]
[37]
92%
60%
90%
Fishers linear determinant analysis (FLDA) Wheat (yellow rust)
Wheat (powdery mildew)
[30] 93%
Erosion and dilation Cucumber (downey mildew) [31] 90%
Spectral angle mapper (SAM) Sugarbeet (cerospora leaf spot)
Sugarbeet (powdery mildew)
Sugarbeet (leaf rust)
Wheat (head blight)
[40]
[39]
89.01–98.90%
90.18–97.23%
61.7%
87%
Artificial neural network (ANN) Sugarbeet (cerospora leaf spot)
Sugarbeet (powdery mildew)
Sugarbeet (leaf rust)
[41] 96%
91%
95%
Support vector machine (SVM) Sugarbeet (cerospora leaf spot)
Sugarbeet (powdery mildew)
Sugarbeet (leaf rust)
Barley (drought)
[41]
[45]
97%
93%
93%
10 days before visible signs
Spectral information divergence (SID) Grapefruit (cankerous, normal, greasy spot. Insect damage, melanose, scab, wind scar) [38] 95.2%
Simplex volume maximisation
SiVM with DAR
Barley (drought)
Barley (drought)
[44]
[47]
4 days before Vegetation Indices
1.5wk Before  visible signs
LSSVM Wheat (drought) [46] 86.6%(H)/76.3%(S)
  1. H healthy, S stressed, D diseased