Early detection of basal stem rot (BSR) disease in oil palm trees is important for the sustainable production of palm oil in the limited land for plantation in Southeast Asia. However, previous studies based on satellite and aircraft hyperspectral remote sensing could not discriminate oil palm trees in the early-stage of the BSR disease from healthy or late-stage trees. In this study, hyperspectral imaging of oil palm trees from an unmanned aerial vehicle (UAV) and machine learning using a random forest algorithm were employed for the classification of four infection categories of the BSR disease: healthy, early-stage, late-stage, and dead trees. A concentric disk segmentation was applied to tree crown segmentation at the sub-plant scale, and recursive feature elimination was used for feature selection. The results revealed that the classification performance for the early-stage trees is maximum at the specific tree crown segments, and only a few spectral bands in the red-edge region are sufficient to classify the infection categories. These findings will be useful for future UAV-based multispectral imaging to efficiently cover a wide area of oil palm plantations for the early detection of BSR disease.

Agricultural drought is crucial in understanding the relationship to crop production functions which can be monitored using satellite remote sensors. The aim of this research is to combine temperature vegetation dryness index (TVDI) and normalized difference water index (NDWI) classifications for identifying drought areas in Chuping, Malaysia which has regularly recorded high temperatures. TVDI and NDWI are assessed using three images of the dry spell period in March for the years 2015, 2016 and 2017. NDWI value representing water content in vegetation decreases numerically to −0.39, −0.37 and −0.36 for the year 2015, 2016 and 2017. Normalized difference vegetation indices (NDVI) values representing vegetation health status in the given area for images of years 2015 to 2017 decreases significantly (p ≤ 0.05) from 0.50 to 0.35 respectively. Overall, TVDI in the Chuping area showed agricultural drought with an average value of 0.46. However, Kilang Gula Chuping area in Chuping showed a significant increase in dryness for all of the three years assessed with an average value of 0.70. When both TVDI and NDWI were assessed, significant clustering of spots in Chuping, Perlis for all the 3 years was identified where geographical local regressions of 0.84, 0.70 and 0.70 for the years 2015, 2016 and 2017 was determined. Furthermore, Moran’s I values revealed that the research area had a high I value of 0.63, 0.30 and 0.23 with respective Z scores of 17.80, 8.63 and 6.77 for the years 2015, 2016 and 2017, indicating that the cluster relationship is significant in the 95–99 percent confidence interval. Using both indices alone was sufficient to understand the drier spots of Chuping over 3 years. The findings of this research will be of interest to local agriculture authorities, like plantation and meteorology departments to understand drier areas in the state to evaluate water deficits severity and cloud seeding points during drought.

Accurate oil palm yield prediction is necessary to sustain oil palm production for food security and economic return. However, there are limited studies on comprehensive mapping and accurate oil palm yield prediction using advanced machine learning algorithms. Using multi-temporal remote sensing data, this paper proposed a new approach to predict oil palm yield based on the normalized difference vegetation index (NDVI) and ensemble machine learning algorithm. ReliefF algorithm with linear projection was employed to select the best combination of spectral indices in oil palm discrimination. Oil palm land cover was classified using random forest (RF) and modified AdaBoost algorithms. A time-series approach known as walk-forward validation was firstly introduced to train the model using the 2016-2019 data and the one-step prediction was performed for 2020 using RF and AdaBoost. Result of the study revealed that the RF model (RMSE = 0.384; MSE = 0.148; MAE = 0.147) outperformed the AdaBoost model (RMSE = 0.410; MSE = 0.168; MAE = 0.176). Our research has demonstrated the value of detailed mapping and subsequent yield prediction by developing a novel approach utilising time-series satellite imagery, ensemble machine learning, and NDVI, which will assist decision-makers in managing their practices related to oil palm.

In oil palm crop, soil fertility is less important than the physical soil characteristics. It is important to have a balance and sufficient soil moisture to sustain high yields in oil palm plantations. However, conventional methods of soil moisture determination are laborious and time-consuming with limited coverage and accuracy. In this research, we evaluated synthetic aperture radar (SAR) and in-situ observations at an oil palm plantation to determine SAR signal sensitivity to oil palm crop by means of water cloud model (WCM) inversion for retrieving soil moisture from L-band HH and HV polarized data. The effects of vegetation on backscattering coefficients were evaluated by comparing Leaf Area Index (LAI), Leaf Water Area Index (LWAI) and Normalized Plant Water Content (NPWC). The results showed that HV polarization effectively simulated backscatter coefficient as compared to HH polarization where the best fit was obtained by taking the LAI as a vegetation descriptor. The HV polarization with the LAI indicator was able to retrieve soil moisture content with an accuracy of at least 80%.

Due to environmental threats and weather uncertainty concerns, oil palm yield prediction is crucial for sustaining crop production. This can be achieved through machine learning and utilising remotely sensed data to predict crop yield. However, the comparative studies on remotely sensed data in adopting the machine learning models are still limited due to the data accessibility. Therefore, we compare and evaluate the prediction accuracy between different satellites, namely MODIS and Landsat-7, using machine learning algorithms and the topology of deep neural networks. Random forest and stacking outperformed linear regression, ridge regression, and lasso regression for both Landsat-7 NDVI (R2= 0.78–0.80; RMSE=1.00- 1.26 tonnes per hectare; MAE=0.77- 0.79 tonnes per hectares; MAPE=0.03-0.04 tonnes per hectare) and MODIS NDVI (R2= 0.60–0.65 tonnes per hectares; RMSE= 2.72–2.81 tonne per hectares; MAE= 1.42-1.55, MAPE= 1.01- 1.02 tonnes per hectares). The Landsat-7 NDVI revealed that neural networks with a deeper network topology (R2= 0.85; RMSE= 1.42 tonnes per hectare; MAE=0.57 tonnes per hectares; MAPE=0.06 tonnes per hectare) outperformed neural networks with a baseline and broader network topologies in terms of performance. In contrast, MODIS-NDVI revealed that the neural network with a wider network topology had the highest overall prediction accuracy and the lowest prediction error (R2= 0.75; RMSE= 2.81 tonnes per hectare; MAE=2.27 tonnes per tonnes; MAPE= 0.13). Because of its higher spatial resolution in comparison to MODIS, landsat-7 NDVI used in neural networks with a deep network topology provided the best model performance. Although the use of NDVI as a single input factor may cause uncertainty in some extents, it is an efficient and reliable method for improving yield estimation with the use of medium-resolution satellites, which has important implications for early warning towards the reduction in yield production.

Synthetic-aperture radar’s (SAR’s) capacity to resolve the cloud cover concerns encountered while gathering optical data has tremendous potential for soil moisture data retrieval using SAR data. It is possible to use SAR data to recover soil moisture because the backscatter coefficient is sensitive to both soil and vegetation by penetrating through the vegetation layer. This study investigated the feasibility of employing a SAR-derived radar vegetation index (RVI), the ratios of the backscatter coefficients using polarizations of HH/HV (R_HH/HV) and HV/HH (R_HH/HV) to an oil palm crops as vegetation indicators in the water cloud model (WCM) using phased-array L-band SAR-2 (PALSAR-2). These data were compared to the manual leaf area index (LAI) and a physical soil sampling method for computing soil moisture. The field data included the LAI input parameters and, more importantly, physical soil samples from which to calculate the soil moisture. The fieldwork was carried out in Chuping District, Perlis State, Malaysia. Corresponding PALSAR-2 data were collected on three observation dates in 2019: 17 January, 16 April, and 9 July. The results showed that the WCM modeled using the LAI under HV polarization demonstrated promising accuracy, with the root mean square error recorded as 0.033 m³/m³. This was comparable to the RVI and R_HH/HV under HV polarization, which had accuracies of 0.031 and 0.049 m³/m³, respectively. The findings of this study suggest that SAR-based indicators, R_HH/HV and RVI using PALSAR-2, can be used to reduce field-related input in the retrieval of soil moisture data using the WCM for oil palm crop.