ReviewA review on the practice of big data analysis in agriculture
Introduction
Population growth, along with socioeconomic factors have historically been associated to food shortage (Slavin, 2016). In the last 50 years, the world’s population has grown from three billion to more than six, creating a high demand for food (Kitzes et al., 2008). As the (Food and Agriculture Organization of the United Nations, 2009) estimates, the global population would increase by more than 30% until 2050, which means that a 70% increase on food production must be achieved. Land degradation and water contamination, climate change, sociocultural development (e.g. dietary preference of meat protein), governmental policies and market fluctuations add uncertainties to food security (Gebbers and Adamchuk, 2010), defined as access to sufficient, safe and nutritious food by all people on the planet. These uncertainties challenge agriculture to improve productivity, lowering at the same time its environmental footprint, which currently accounts for the 20% of the anthropogenic Greenhouses Gas (GHG) emissions (Sayer and Cassman, 2013).
To satisfy these increasing demands, several studies and initiatives have been launched since the 1990s. Advancements in crop growth modeling and yield monitoring (Basso et al., 2001), together with global navigation satellite systems (e.g. GPS) (Aqeel ur et al., 2014) have enabled precise localization of point measurements in the field, so that spatial variability maps can be created (Pierce and N., 1999), a concept known as “precision agriculture” (Bell et al., 1995).
Nowadays, agricultural practices are being supported by biotechnology (Rahman et al., 2013) and emerging digital technologies such as remote sensing (Bastiaanssen et al., 2000), cloud computing (Hashem et al., 2015) and Internet of Things (IoT) (Weber and Weber, 2010), leading to the notion of “smart farming” (Tyagi, 2016, Babinet,Gilles et al., 2015). The deployment of new information and communication technologies (ICT) for field-level crop/farm management extend the precision agriculture concept (Lokers et al., 2016), enhancing the existing tasks of management and decision making by context (Kamilaris et al., 2016), situation and location awareness (Karmas et al., 2016).
Smart farming is important for tackling the challenges of agricultural production in terms of productivity, environmental impact, food security and sustainability. Sustainable agriculture (Senanayake, 1991) is very relevant and directly linked to smart farming (Bongiovanni and Lowenberg-DeBoer, 2004), as it enhances the environmental quality and resource base in which agriculture depends, providing basic human food needs (Pretty, 2008). It can be understood as an ecosystem-based approach to agriculture, which integrates biological, chemical, physical, ecological, economic and social sciences in a comprehensive way, in order to develop safe smart farming practices that do not degrade our environment.
To address the challenges of smart farming and sustainable agriculture, as (McQueen et al., 1995) and (Gebbers and Adamchuk, 2010) point out, the complex, multivariate and unpredictable agricultural ecosystems need to be better analyzed and understood. The aforementioned emerging digital technologies contribute to this understanding by monitoring and measuring continuously various aspects of the physical environment (Sonka, 2016), producing large quantities of data in an unprecedented pace (Chi et al., 2016). This implies, as (Hashem et al., 2015) note, the need for large-scale collection, storage, pre-processing, modeling and analysis of huge amounts of data coming from various heterogeneous sources.
Agricultural “big data” creates the necessity for large investments in infrastructures for data storage and processing (Nandyala and Kim, 2016, Hashem et al., 2015), which need to operate almost in real-time for some applications (e.g. weather forecasting, monitoring for crops’ pests and animals’ diseases). Hence, “big data analysis” is the term used to describe a new generation of practices (Kempenaar et al., 2016, Sonka, 2016), designed so that farmers and related organizations can extract economic value from very large volumes of a wide variety of data by enabling high-velocity capture, discovery, and/or analysis (Waga and Rabah, 2014, Lokers et al., 2016).
Big data analysis is successfully being used in various industries, such as banking, insurance, online user behavior understanding and personalization, as well as in environmental studies (Waga and Rabah, 2014, Cooper et al., 2013). As (Kim et al., 2014) show, governmental organizations use big data analysis to enhance their ability to serve their citizens addressing national challenges related to economy, health care, job creation, natural disasters and terrorism.
Although big data analysis seems to be successful and popular in many domains, it started being applied to agriculture only recently (Lokers et al., 2016), when stakeholders started to perceive its potential benefits (Bunge, 2014, Sonka, 2016). According to some of the largest agricultural corporations, tailoring advice to farmers based on analyzing big data could increase annual global profits from crops by about US $20 billion (Bunge, 2014).
The motivation for preparing this survey stems from the fact that big data analysis in agriculture is a modern technique with growing popularity, while recent advancements and applications of big data in other domains indicate its large potential (Kim et al., 2014, Cooper et al., 2013). Current relevant surveys (Wolfert et al., 2017, Nandyala and Kim, 2016, Waga and Rabah, 2014, Wu et al., 2016) cover mostly theoretical aspects of this technique (e.g. conceptual framework, socioeconomics, business processes, stakeholders’ network) or focus on particular sub-domains such as remote sensing (Chi et al., 2016, Liaghat and Balasundram, 2010, Teke et al., 2013, Ozdogan et al., 2010, Karmas et al., 2014) and geospatial analysis (Karmas et al., 2016). Thus, the main contribution of this survey is that it presents a more focused overview of the particular problems encountered in agriculture, compared to existing surveys, where data analysis is a key aspect and solutions are found inside the big data realm. Our survey highlights the (big) data used, the methods and techniques employed, giving specific insights from a technical perspective on the potential and opportunities of big data analysis, open issues, barriers and ways to overcome them.
Section snippets
Methodology
The bibliographic analysis in the domain under study involved three steps: (a) collection of related work, (b) filtering of relevant work, and (c) detailed review and analysis of state of the art related work. In the first step, a keyword-based search for conference papers and articles was performed from the scientific databases IEEE Xplore and ScienceDirect, as well as from the web scientific indexing services Web of Science (Thomson Reuters, 2017) and Google Scholar. As search keywords, we
Big data in agriculture
Chi et al. (2016) characterize big data according to the following five dimensions:
- •
Volume (V1): The size of data collected for analysis.
- •
Velocity (V2): The time window in which data is useful and relevant. For example, some data should be analyzed in a reasonable time to achieve a given task, e.g. to identify pests (PEAT UG, 2016) and animal diseases (Chedad et al., 2001).
- •
Variety (V3): Multi-source (e.g. images, videos, remote and field-based sensing data), multi-temporal (e.g. collected on
Discussion
As Section 3 illustrated, a large variety of agricultural issues are currently approximated by the use of big data analysis, employing a variety of different algorithms, approaches and techniques (see Table 3). Table 4 presents the specific, most common software used in the revised papers, mapped according to the type of analysis employed. A wide variety of software available for big data analysis exist, in all different types of analysis.
A recent practice is to approximate agricultural
Conclusion
This paper performed a review of big data analysis in agriculture, mostly from a technical perspective. Thirty-four research papers were identified and analyzed, examining the problem they addressed, the solution proposed, tools/techniques employed as well as data used. Based on these projects, the reader can be informed about which types of agricultural applications currently use big data analysis, which characteristics of big data are being used in these different scenarios, as well as which
Acknowledgments
We would like to thank the reviewers for their valuable feedback, which helped us to reorganize the structure of the survey more appropriately, as well as to improve its overall quality. This research has been supported by the P-SPHERE project, which has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skodowska-Curie grant agreement No 665919.
References (108)
- et al.
Assessment of multi-temporal, multi-sensor radar and ancillary spatial data for grasslands monitoring in Ireland using machine learning approaches
Remote Sens. Environ.
(2014) Spatial validation of crop models for precision agriculture
Agric. Syst.
(2001)- et al.
Remote sensing for irrigated agriculture: examples from research and possible applications
Agric. Water Manage.
(2000) AP – animal production technology: recognition system for pig cough based on probabilistic neural networks
J. Agric. Eng. Res.
(2001)- et al.
Comparison of genetic algorithm and neural network approaches for the drying process of carrot
J. Food Eng.
(2007) - et al.
Phenomics – technologies to relieve the phenotyping bottleneck
Trends Plant Sci.
(2011) Wavelet analysis of MODIS time series to detect expansion and intensification of row-crop agriculture in Brazil
Remote Sens. Environ.
(2008)Communicating complexity: integrated assessment of trade-offs concerning soil fertility management within African farming systems to support innovation and development
Agric. Syst.
(2011)Logistic regression product-unit neural networks for mapping Ridolfia segetum infestations in sunflower crop using multitemporal remote sensed data
Comput. Electron. Agric.
(2008)The rise of “big data” on cloud computing: review and open research issues
Inform. Syst.
(2015)
Analysis of Big Data technologies for use in agro-environmental science
Environ. Model. Software
Machine learning for Big Data analytics in plants
Trends Plant Sci.
Using Bayesian belief networks to evaluate fish and wildlife population viability under land management alternatives from an environmental impact statement
For. Ecol. Manage.
Intensified fuzzy clusters for classifying plant, soil, and residue regions of interest from color images
Comput. Electron. Agric.
Big data challenges in building the global earth observation system of systems
Environ. Model. Software
To mulch or to munch? Big modelling of big data
Agric. Syst.
A crop phenology detection method using time-series MODIS data
Remote Sens. Environ.
Ecological views of big data: perspectives and issues
Telematics Inform.
Downscaling of precipitation for climate change scenarios: a support vector machine approach
J. Hydrol.
Using data mining techniques to predict industrial wine problem fermentations
Food Control
Web technologies for environmental Big Data
Environ. Model. Software
A review of wireless sensors and networks' applications in agriculture
Comput. Stand. Interfaces
Advances in remote sensing of agriculture: context description, existing operational monitoring systems and major information needs
Remote Sens.
Monitoring global croplands with coarse resolution earth observations: the Global Agriculture Monitoring (GLAM) project
Remote Sens.
Precision agriculture and sustainability
Precision Agric.
Next generation data systems and knowledge products to support agricultural producers and science-based policy decision making
Agric. Syst.
The ethics of big data in big agriculture
Internet Policy Rev.
Publicising food: big data, precision agriculture, and co-experimental techniques of addition
Soc. Ruralis
Big data for remote sensing: challenges and opportunities
Proc. IEEE
Big data in life cycle assessment
J. Ind. Ecol.
The potential and uptake of remote sensing in insurance: a review
Remote Sens.
Drivers of household food availability in sub-Saharan Africa based on big data from small farms
Proc. Natl. Acad. Sci.
Precision agriculture and food security
Science
Cited by (607)
Influence of methodological choices in farm sustainability assessments: A word of caution from a case study analysis of European dairy farms
2024, Environmental Science and PolicyThe power of on-farm data for improved agronomy
2024, Global Food SecurityMulti-task learning model for agricultural pest detection from crop-plant imagery: A Bayesian approach
2024, Computers and Electronics in AgricultureIoT smart farming adoption intention under climate change: The gain and loss perspective
2024, Technological Forecasting and Social ChangeDos and Don'ts of using drone technology in the crop fields
2024, Technology in SocietyFoundations of Programmable Process Structures for the unified modeling and simulation of agricultural and aquacultural systems
2024, Information Processing in Agriculture