Chapter 6

Remote Sensing in Precision Agriculture

Remote Image Processing and Data Analysis

Remote sensing data have to be processed before it can be used. In general, it is not suggested to use raw images directly acquired from remote sensors on satellites, aircraft, and UAVs because the data have to be corrected due to deformations from interactions between sensors, atmospheric conditions, and terrain profiles. Although there might be a slight difference in how datasets from different sensors are handled, the following steps present a broad pipeline that ensures the efficacy of analysis. This pipeline is divided into three main steps: data collection, digital image processing, and data analysis and processing.

Remote Sensing Data Collection

Remote sensing data collection involves gathering information about the Earth’s surface, atmosphere, and oceans from a distance, typically using satellites, aircraft, or UAVs. This process utilizes sensors to detect and record reflected or emitted energy from various sources, providing valuable data for a range of applications. The data collected can be in various formats, such as satellite images, aerial photographs, or even radar data.

Digital Image Processing

Data acquired remotely are affected by several factors, such as sensor characteristics, illumination, geometric alignments, and atmospheric conditions. In order to obtain temporally and spatially consistent field results, data must go through a standard preprocessing pipeline. Throughout preprocessing, raw data (that are mainly in the form of Digital Numbers (DN)) need to be converted to meaningful values and attributed to their real-world correspondents, such as reflectance or temperature.

Data Analysis and Presentation

After the remote sensing data has been acquired, a range of open-access and proprietary software is available for visualization, processing, and analysis. Such programs typically include a Geographic Information System (GIS) component, which can also deal with vector data (points, lines, and polygons). GIS can achieve visualization of remote sensing data and products. Through the use of a GIS, it is possible to examine relationships among layers of georeferenced data within a field (Figure 6.17).

Common Digital Analysis Techniques

A simple way to use remote sensing imagery is to perform a visual inspection (often aided by composite images, such as RGB). Usually, this is enough to identify some features of interest, such as crop health and soil erosion. Additionally, visual inspection of data layers can yield important information about which pre-processing steps might be necessary. However, remote sensing imagery is rich in information that is not easily apparent to ad-hoc human interpretation. In the following, a short overview of three ways to use this information is presented: (1) generating new layers taking into account spatial context (textures), (2) extracting the most important information from many different bands (dimensionality reduction), and (3) image classification techniques.

Click on the following topics for more information on remote rensing in precision agriculture.