Skip to main content

Table 3 Part 2: an analytical study on examples of crop prediction methods; highlights: the applied learning algorithms, the crop type, data type and pre-processing, the other studied and considered parameters in each proposed approach and the predictor variables for each used algorithm

From: Data analytics for crop management: a big data view

References

Algorithm

Plant

Data type

Data pre/processing

Other parameters

Predictor variables

[152]

CNN, LSTM Gaussian

Soybean

Sensor MODIS data

Transform multi-spectral images to individual histograms

/

Histogramsgeographic location year

[62]

Regression Rulequest Cubist

Corn, soybean

Satellite data

8-days periods data-points composed then averaged

Pre and within season

NDVI, precipitation day and night land surface temperature

[67]

DT (Extremely randomised) SVM(RBF);DNN

Corn

Satellite images climate, yield

/

Seasonal sensitivities

NDVI and many other features

[27]

Deep learning semi-parametric NN for training: bayesian hyper-parameter optimisation and early stopping

Corn

Historical yield weather

/

Climate change impact and semi-parametric prediction model

Precipitation temperature humidity, wind radiation Latitude and longitude Growing degree, soil County, irrigation

[68]

DNN; ANN; RF; multivariate adaptive regression splines SVM; extremely randomised trees

Corn and soybean

Satellite images MODIS historical yield meteorological and crop landhydrological

/

Effect of phenology

EVI Leaf Area Index Gross Primary Production precipitation; min, max air temperature soil Moisture, NDVI

[50]

SVM

Rice

Climate and geographical data

/

Effect of phenology and climate pre-season

Mean, max, mintemperature Daily Sunshine hours precipitation mean relative humidity min relative humidity mean wind speed maximum wind speed

[60]

LR

Corn

MODIS remote sensing

Savitzky-Golay filter for smoothing NDVI time series

Effect of phenology

Max correlation NDVI crop growth rate crop growth days

[111]

RF

Chickpea

Modis images weather data yield statistics remote sensing

/

Drylands sensitivity to data time

EVI, NDVI, Leaf Area Index precipitation and 5 other features

[4]

Bidirectional LSTM

Tomato, Potato

Climate datairrigation scheduling soil water content

Moving average method for data imputation multi-collinear parametersremoval

Effect of irrigation scheduling

Min, max, mean temperature min, max, average relative humidity average solar radiation min and average wind speed precipitation

[91]

3D-CNN

Wheat, Barley, Oats

Weather data UAV RGB image yield data

Images resizing

Effects of time:efficiency of using time series data vs point-in-time data

RGB Images, cumulative temperature

[19]

LSTM

Winter, wheat

Climate satellite data soil surveys

/

/

Min and max temperatureprecipitation, EVI, soil depth and texture, pH geographic properties