НАУКА О ДАННЫХ В СЕЛЬСКОМ ХОЗЯЙСТВЕ: ВОЗМОЖНОСТИ И ВЫЗОВЫ

Научная статья
DOI:
https://doi.org/10.23670/IRJ.2018.71.023
Выпуск: № 5 (71), 2018
Опубликована:
2018/05/19
PDF

Кошкаров А.В.

ORCID: 0000-0002-3630-2911, Кандидат технических наук,

Астраханский государственный университет, Астрахань, Россия

НАУКА О ДАННЫХ В СЕЛЬСКОМ ХОЗЯЙСТВЕ: ВОЗМОЖНОСТИ И ВЫЗОВЫ

Аннотация

В современном мире с растущим населением проблемы продовольственной безопасности стран становятся наиболее актуальными. Одним из решений является повышение эффективности сельского хозяйства. Это может быть достигнуто с использованием методов науки о данных применительно к сельскохозяйственной отрасли. Задачей исследования было выявить и описать возможности и проблемы использования методов интеллектуального анализа данных в сельском хозяйстве. Были проанализированы практические и теоретические аргументы и отличительные особенности применения данных в сельскохозяйственной отрасли. На основе этого анализа были отобраны и описаны несколько возможностей и проблем с использованием методов науки о данных в сельском хозяйстве.

Ключевые слова: наука о данных, цифровое сельское хозяйство, анализ данных, точное земледелие.

Koshkarov A.V.

ORCID: 0000-0002-3630-2911, PhD in Engineering,

Astrakhan State University, Astastrakhan, Russia

DATA SCIENCE IN AGRICULTURE: OPPORTUNITIES AND CHALLENGES

Abstract

In the modern world with a growing population, the issues of food safety are very important. Improving the efficiency of agriculture is one of the solutions. This can be achieved using data science methods. The primary objective of this study was to identify and describe opportunities and challenges by utilizing data science techniques in agricultural sector. This study analyzed the practical and theoretical arguments and distinctions in agricultural technologies and assessed the available evidence relating to the utilization of data science approaches in agricultural industry. Based on this analysis, several opportunities and challenges with utilising data science methods in the agricultural sector were selected and described.

Keywords: data science, digital agriculture, data analysis, precision agriculture.

Introduction

In modern times, huge amounts of data can be generated, and the data are often regarded as a major strategic asset of organizations. Appropriate methods and algorithms are needed to extract useful knowledge from raw data, and data science deals with it. Data science can be described as a multidisciplinary field to study how data can be turned into a useful resource for various business reasons [10, P. 10].

A key data science concept is precisely defined by Loukides [4]. The author addresses the issue why there are so many expectations associated with the data and the results of the analysis. An explanation of the increasing interest to data is connected with data products [5, P. 2].

At the current stage, the data are the important asset, and data science is a generalized name of the sum of technologies for the development of data products [9, P. 9]. Today, the sale of information becomes a big business. There are a huge number of various data-driven applications, but it tends to involve the passive use of data.

Several economic sectors have gained a new stimulus as the result of the development of data science methods and the Internet of Things. These include agriculture, energy, health, logistics, mining industry, telecommunications, and education.

Agriculture

Agriculture is the economic sector, aimed to provide the population with food and to obtain raw materials for a number of industries. This sector is one of the most important parts of the economy, represented in almost all countries. Food safety of the state depends on the situation in the sector [1, P. 689].

The development of agriculture affected by ‘green revolution’ that originated in the middle of the XX century. Agriculture began to use advances in science, fertilizers, new technologies, genetic engineering, electronics, robotics and biotechnology, and this led to an increase in productivity [2, P. 758].

Due to the fact that agriculture uses huge areas of land, it has a significant impact on the environment. Impact factors are as follows: land plowing, tillage, use of mineral fertilizers and pesticides, land reclamation [11, P. 73]. There are certain techniques and technologies of agriculture, which mitigate or completely eliminate the negative factors. An example is precision agriculture.

Precision agriculture - is an optimal management of crop production per square meter of the field in order to maximize profits with savings of economic and natural resources [12, P. 28]. This requires a modern agricultural equipment controlled by an onboard computer, precise positioning devices, technical systems that detect inhomogeneity of the field, the system of automatic crop accounting, fertilizer application system, computer programs for data analysis and mapping. Real-time data collection became possible due to the Internet of things technology.

There is an ongoing research on the improvement of forms of agriculture with the use of various technologies [7, P. 24]. This is very important because of the growing world population. According to UN experts, consumption growth may require a doubling of food production by 2050 [3]. Such growth can be achieved with the use of technologies based on the data. In the near future data specialists will manage agriculture, robots will cultivate the land, and humans will program these machines and make decisions [8].

Opportunities

By measuring the characteristics of agricultural fields and soils and having information about soil types and predicted level of precipitation, it is possible to make forecasts of productivity and provide recommendations that will help farmers to grow a larger crop with the same field sizes. This is the most significant opportunity that data science would bring to agriculture.

Data for processing and analysis can be collected from two main sources (but not limited to). The first source is the wireless soil sensor network (see Figure). A special base station automatically collects data from these sensors and sends it to the server for storage and processing. The second source is an unmanned aerial vehicle (drone) for aerial shooting with a special camera (image data can be processed and analyzed). It is worth noting that these data can also be obtained with the help of satellites (satellite images).

25-05-2018 11-35-12

Fig. 1 – Data collection in agriculture

 

Agricultural lands can be supplied with digital sensors that produce measurements of soil and air. Sensors send the data into a single repository where data are structured and prepared for post-processing. A comprehensive picture of the treated area at any given time can be obtained using of such data flow. Gathering data continuously, it is possible not only to monitor the overall situation but also promptly adjust and predict it [14, P. 5]. The data analysis can tell the farmer how to optimize operations, save water and chemicals.

Data from the fields can be collected using the drone with a camera that shoots in the visible and thermal range. A complete set of digital information is obtained for each image: the geographical coordinates, the height of the shooting, the full range of telemetry data for the transfer and use in GIS systems. The data allow judging about the status of crops, the soil moisture, the volume of fertilizers, and help to predict the harvest. A key role is played by NDVI (Normalized Difference Vegetation Index), which is calculated on the basis of spectral photography. The NDVI is an indicator of the amount of photosynthetically active biomass; it is intended to measure the ecological and climatic characteristics of the vegetation [13, P. 10].

Data analysis in agriculture by using the drones provides additional opportunities such as:

  • creating electronic maps of fields,
  • conducting operational monitoring crop conditions,
  • evaluation of germinating ability of crops,
  • prediction of crop yields, and
  • conducting environmental monitoring of agricultural lands.

It is possible to trace the dynamics of changes within the same field during the regular shooting of agricultural lands and processing (analysis) of data collected. These data can be compared with the soil sensors data and the productivity of land. On the other hand, not the data itself, but the specific information for decision-making will become more popular and in demand.

Challenges

One of the main challenges of the implementation of data science methods in agriculture is the high cost of equipment for data collection. Implementation of new technology often has a certain risk. Introduction of modified agricultural machines and wireless sensor systems require significant financial investments. Furthermore, it is important to understand the payback period of the investment.

The second challenge is legislation. Political will often is necessary to implement technology in the agricultural industry at the state level. At the same time, the solution of this question is the interest of the state. Knowledge of the real areas and the structure of agricultural lands will allow to learn how many used and abandoned lands in the country and to set a policy on its effective use. In addition, the criteria for improper use of agricultural lands and violations of land legislation should be clearly defined.

An additional aspect is the legal restrictions of drone flights. Driving a drone requires some training. In some countries, the drones should be registered, and a special license must be obtained. The main problems of using drones are safety, the protection of information collected, and the nuances of insurance [6].

The third challenge is the lack of specialists in agrarian technologies, capable to implement new technologies in agriculture. The task of improving educational programs is no less important. The programs should be focused on not only acquiring knowledge but also gaining professional skills.

The greatest challenge to making use of data science in agriculture is a quality of the data, a quality of processing methods, and forms of presentation of results. To solve this problem, the industry needs to create more advanced sensors and cameras as well as to develop a highly automated machines and intuitive software that requires minimal training to work with it.

Conclusion

The findings of this study suggest that the methods of data science have considerable advantages for agriculture if these techniques are used properly. There are also barriers to the introduction of innovations in agriculture, but these obstacles can be minimized. It is possible that data specialists will manage agriculture in the future.

Список литературы / References

  1. Carvalho F. P. Agriculture, pesticides, food security and food safety / F. P. Carvalho // Environmental science & policy. – 2006. – Vol. 9. – N 7-8. – pp. 685-692.
  2. Evenson R. E. Assessing the impact of the Green Revolution, 1960 to 2000 / R. E. Evenson, D. Gollin // Science. – 2003. – Vol. 300. – N 5620. – P. 758-762.
  3. Food Production Must Double by 2050 to Meet Demand from World's Growing Population, Innovative Strategies Needed to Combat Hunger, Experts Tell Second Committee / Meetings Coverage and Press Releases [Electronic resource] // United Nations. United Nations. URL: http://www.un.org/press/en/2009/gaef3242.doc.htm (accessed: 22.02.2018).
  4. Loukides M. What is data science? The future belongs to the companies and people that turn data into products [Electronic resource] / M. Loukides // An O’Reilly Radar Report. URL: https://www.oreilly.com/ideas/what-is-data-science (accessed: 10.01.2018).
  5. Loukides M. The evolution of data products / M. Loukides // O'Reilly Media, Inc., 2011. – 14 p.
  6. Mazur M. Six ways drones are revolutionizing agriculture [Electronic resource] / M. Mazur // MIT Technology Review. URL: https://www.technologyreview.com/s/601935/six-ways-drones-are-revolutionizing-agriculture/ (accessed: 02.02.2018).
  7. McCouch S. Agriculture: feeding the future / S. McCouch, G. J. Baute, J. Bradeen and others // Nature. – 2013. – Vol. 499. – N 7456. – P. 23-24.
  8. Melendez S. How Machine Learning Will Change What You Eat [Electronic resource] / S. Melendez // URL: https://www.fastcompany.com/3062262/how-machine-learning-will-change-what-you-eat (accessed: 18.02.2018).
  9. Provost F. Data Science for Business: What you need to know about data mining and data-analytic thinking / F. Provost, T. Fawcett. – O'Reilly Media, Inc., 2013. – 387 p.
  10. O'Neil C. Doing data science: Straight talk from the frontline / C. O'Neil, R. Schutt // O'Reilly Media, Inc., 2014. – 379 p.
  11. Shiva V. The violence of the green revolution: Third world agriculture, ecology, and politics / V. Shiva. – University Press of Kentucky, 2016. – 266 p.
  12. Xia J. An environment monitoring system for precise agriculture based on wireless sensor networks / J. Xia, Z. Tang, X. Shi and others // Mobile Ad-hoc and Sensor Networks (MSN), 2011 Seventh International Conference on. – IEEE, – 2011. – P. 28-35.
  13. Yengoh G. T. Use of the Normalized Difference Vegetation Index (NDVI) to Assess Land Degradation at Multiple Scales: Current Status, Future Trends, and Practical Considerations / G. T. Yengoh, D. Dent, L. Olsson and ohers // Springer, 2015. – 110 p.
  14. Zaks D. P. M. Data and monitoring needs for a more ecological agriculture / D. P. M. Zaks, C. J. Kucharik // Environmental Research Letters. – 2011. – Vol. 6. – N 1. – P. 1-10.