6 Practical applications of multispectral monitoring in precision agriculture

A key objective of multispectral monitoring is its practical application in precision agriculture, where the main principle is to move away from uniform field management towards site-specific management that accounts for spatial variability within agricultural fields [11].

Traditionally, agricultural inputs have often been applied uniformly across entire fields, using standardised application rates without adequately accounting for spatial variability in soil properties, topography, and microclimatic conditions. In contrast, precision agriculture enables field heterogeneity to be taken into account through site-specific management practices. Multispectral sensing plays an important role in this process by helping to identify areas with potentially different crop conditions and supporting the generation of digital prescription maps for subsequent use by agricultural machinery [12].

In practical terms, multispectral monitoring enables a transition from general field observation to spatially detailed crop assessment. This approach supports the evaluation of agroecosystem conditions and addresses a range of challenges associated with efficient and sustainable agricultural production. Its applications include variable-rate fertiliser application, weed detection and monitoring to support targeted herbicide spraying, crop health assessment, yield estimation, and the planning of crop desiccation operations [13].

One of the principal applications of multispectral monitoring is variable-rate fertilisation, particularly the optimisation of nitrogen application. Crop condition is commonly assessed using vegetation indices such as NDVI, NDRE, and GNDVI. These indices reflect variations in vegetation characteristics associated with chlorophyll content, canopy structure, and biomass, and may therefore provide indirect information about crop nitrogen status [13]. However, the relationship between vegetation index values and nitrogen availability depends on crop type, growth stage, canopy structure, and environmental conditions. To support site-specific fertilisation, a vegetation index map is generated from the acquired multispectral imagery. Based on the spatial distribution of index values and, where available, additional agronomic information, the field can be divided into management zones representing areas with different crop conditions or fertilisation requirements. These zones can then be used to develop prescription maps for site-specific field operations. A variable-rate application (VRA) system interprets the prescription map and adjusts fertiliser application rates according to the specified management requirements. In precision seeding applications, compatible equipment can similarly adjust seeding rates in accordance with a prescription map [14].

The second major application of multispectral monitoring is weed detection and the implementation of targeted spot spraying in areas identified as infested. In conventional weed management, herbicides are often applied uniformly across the field, including areas where weed infestation is limited or absent. This can result in unnecessary herbicide use and increased environmental exposure to agrochemicals [15, 16]. By contrast, precision weed management can utilise high-resolution imagery, with a ground sampling distance (GSD) of approximately 1-3 cm per pixel, depending on the sensor and survey configuration [10]. Such imagery can facilitate the identification and differentiation of weeds and cultivated plants, particularly when sufficient spatial detail is available to distinguish individual plants. Spectral information can support this process, although successful discrimination also depends on plant size, canopy structure, illumination conditions, and the degree of overlap between crop and weed canopies. Based on the resulting weed distribution maps, spot spraying can be implemented to treat only the affected areas. This approach can substantially reduce herbicide consumption, with reported savings of approximately 40-60% under certain conditions, while also reducing the environmental burden associated with herbicide application [15, 16].

The third major application is the phytosanitary monitoring of agricultural crops. Multispectral imagery can support the early detection of plant stress and, under suitable conditions, reveal changes before visible symptoms become apparent [17]. This capability is associated with changes in leaf structure, pigment content, and tissue properties caused by certain diseases and other stress factors. These changes can affect the reflectance characteristics of vegetation, particularly in the red-edge and near-infrared (NIR) regions of the electromagnetic spectrum [18, 19]. Analysing these spectral variations can help identify areas requiring further investigation and support the targeted application of plant protection products. However, spectral anomalies are not necessarily disease-specific, as similar responses may be caused by water deficiency, nutrient imbalances, or other environmental factors. Therefore, field inspections and, where necessary, additional diagnostic methods are required to establish the underlying cause before treatment decisions are made.

The fourth important application of multispectral monitoring is yield forecasting and the planning of variable-rate crop desiccation. Yield estimation may incorporate indicators of crop biomass, canopy development, and vegetation condition derived from remote sensing data [20]. These indicators can support the preliminary assessment of expected yields, facilitate logistical planning for harvesting operations, and help optimise the allocation of combine harvesters. However, reliable yield forecasts generally require calibration against field observations, crop-specific information, and, where available, historical yield data. Variable-rate desiccation, in turn, involves adjusting treatment according to spatial differences in crop maturity and moisture conditions. Multispectral data, including information from the NIR and red-edge bands, can contribute to the assessment of vegetation condition and the identification of areas with differing levels of crop maturity [21]. Based on this information, prescription maps can be developed to support differentiated desiccant application, where permitted and agronomically appropriate. Treatment can then be concentrated in areas that require it, while areas that have already reached the desired stage of maturity may be excluded. This approach can help reduce desiccant consumption, improve the consistency of crop maturation at harvest, and support more efficient harvesting operations.

Multispectral monitoring can provide substantial economic benefits by helping to optimise the use of fertilisers and plant protection products. Pesticide use can be reduced by an average of 12.8%, while the average return on investment increases by 22.3% [22]. By enabling the timely identification of crop stress and the targeted implementation of appropriate management measures, this approach can also contribute to maintaining or improving crop yields. Furthermore, it offers environmental benefits by potentially reducing chemical inputs, minimising the risk of soil and groundwater contamination, and lowering the carbon footprint associated with agricultural operations [22, 23]. The targeted application of agricultural inputs may also reduce the need for machinery passes across the field, thereby decreasing fuel consumption and associated emissions.

Thus, multispectral monitoring represents an effective decision-support tool that enables agronomists to identify crop conditions and spatial variations that may not be readily apparent through visual inspection alone. By supporting data-driven management decisions, it helps optimise input use, improve resource efficiency, and promote more sustainable agricultural production. Its integration into precision agriculture creates new opportunities for the development of more efficient, technologically advanced, and environmentally responsible farming practices.