Chapter 34

Precision Viticulture

Remote Sensing

Remote sensing (RS) can be understood as the technique that uses sensors, without direct contact with the object of interest, to obtain information by capturing its reflected and/or emitted electromagnetic radiation. For this, a remote sensing system consists of a source of electromagnetic radiation (e.g., the sun), a terrestrial target (vegetation, water bodies, soil, etc.), a platform that carries sensors (e.g., satellites, aircraft, or drones), and a site of data processing, storage, and distribution. Remote sensing in viticulture utilizes various technologies, like satellites and drones, to gather information about crops, soil, and other agricultural factors from a distance. This data helps growers make informed decisions regarding irrigation, fertilization, pest control, and other aspects of farming. Remote sensing utilizes various types of sensors (e.g., optical, multispectral, hyperspectral, LiDAR) to gather information about grapevines and vineyards from a distance without physical contact.

Sensors

Sensors come in various types, including visible light (RGB) sensors, multispectral sensors, hyperspectral sensors, light detection and ranging (LiDAR) sensors, and thermal sensors, each designed to gather specific information. RGB sensors are commonly used to capture images of visible light, making them suitable for general imaging purposes. Multispectral sensors capture light in multiple bands extending beyond the visible spectrum, typically including the near-infrared (NIR) range. Hyperspectral sensors capture hundreds or even thousands of narrow wavelength bands, providing a more detailed spectral profile.

RGB Sensors

RGB sensors are the most common type of imaging sensor used for plant measurements. They capture plant images in the visible part of the EM spectrum in three broad bands (Red, Green, and Blue). A combination of Blue (≈400–500nm), Green (≈500–600nm), and Red (≈600–700nm) would form an RGB image. RGB images are generally easier to interpret because they resemble how the human eye perceives color.

Multispectral Sensors

The multispectral sensor is one of the most commonly used for remote sensing. Multispectral sensors typically collect data on three to ten relatively wide bands, while hyperspectral cameras for small drones collect data on tens to hundreds of narrow bands. Multispectral sensors have at least one filter beyond the visible spectrum (RGB), usually in the near-infrared (NIR) region. Near-infrared (NIR) measurements are based on specific absorption bands in the electromagnetic spectrum, ranging from 800 to 2,500 nanometers (nm).

Hyperspectral Sensors

Multispectral images have been widely used in agriculture to retrieve various crop and soil attributes, such as crop chlorophyll content, biomass, yield, and soil degradation. However, due to the limitations in spectral resolution, the accuracy of the retrieved variables is often limited, and early signals of crop stresses (e.g., nutrient deficiency, crop disease) cannot be effectively detected in a timely manner. Hyperspectral sensors have emerged as a promising tool for precision viticulture (PV).

Thermal Infrared Sensors

Thermal remote sensing uses information on the radiation emitted in the thermal infrared (TIR) range of the electromagnetic (EM) spectrum. This information is converted into temperature. Two categories can be distinguished within the infrared radiation (IR) region (0.7–100µm), namely, the reflected-IR (0.7–3.0µm) and thermal infrared (TIR) (8–14µm).

LiDAR Sensors

LiDAR stands for Light Detection and Ranging. LiDAR is an active sensor that emits light pulses with a laser, which can be used in conjunction with GPS receivers to provide precise geospatial location data. The reflected light data is processed and stored as spatial coordinates to produce a 3D image, known as a point cloud footprint. LiDAR sensors estimate the distance to the target by measuring the time between the emission of the laser beam and the moment at which the light reflected by the target reaches the photodetector, assuming that laser light travels at the constant speed of light through air.

Remote Sensing Platforms

Remote sensing platforms are defined as vehicles, such as satellites, aircraft, and unmanned aerial vehicles (UAVs), that can carry sensing devices to perform remote measurement operations. These platforms acquire images with different spatial coverage, spatial resolution, temporal resolution, operational complexity, and mission cost. These platforms are continually improving in terms of operational time, reliability, simplicity, and temporal resolution (the time interval between successive remote sensing measurements), which in turn affects the spatial resolution. With diverse options available with specific characteristics, choosing a suitable platform depends on the nature of the problem (Table 34.1). Three critical factors for selecting the best platform are spatial resolution, vineyard size, and operation cost.

Satellite Remote Sensing

Space-borne platforms for remote sensing are considered the most stable among all others. Space-borne platforms are categorized based on their orbits and timing. Unlike UAV platforms, which are a suitable solution when a “micro” view of the land is of interest, satellites provide a relatively low-cost “macro” view of the terrain, making them an efficient method for large-scale mapping, such as desertification, land cover classification, climate change, and inter-field comparisons.

Manned Aircraft Remote Sensing

Aircraft often have a distinct advantage due to their high mobilization flexibility. They can be deployed wherever and whenever weather conditions are favorable. Aircraft on site can respond with a moment’s notice to take advantage of clear conditions, while satellites are locked into a schedule dictated by orbital parameters. Aircraft can also be deployed in small or large numbers, making it possible to collect imagery seamlessly over an entire area in a matter of days by having multiple planes in the air simultaneously.

Unmanned Aerial Vehicle Remote Sensing

Unmanned aerial vehicles (UAVs), commonly referred to as drones, are defined as powered aerial vehicles that can fly autonomously or be piloted remotely over the vineyard to collect data. Unmanned aerial vehicles have become almost ubiquitous in use over the last five years, with affordability and camera payloads ranging from visible, near-infrared, and thermal infrared, as well as 3D LIDAR.

Remote Sensing Applications in Precision Viticulture

Precision viticulture (PV) encompasses the application of a suite of technologies, including geospatial technologies, the Internet of Things (IoT), big data analysis, and artificial intelligence (AI), to optimize agricultural inputs, enhance production, and minimize input losses. The use of remote sensing technologies for PV has increased rapidly during the past few decades. The unprecedented availability of high-resolution (spatial, spectral, and temporal) images has promoted remote sensing in many PV applications, including crop monitoring, irrigation management, nutrient application, disease and pest management, and yield prediction.

Click on the following topics for more information on precision viticulture.