nach oben

2022 | Buch

Information and Communication Technologies for Agriculture—Theme II: Data

herausgegeben von: Dionysis D. Bochtis, Dimitrios E. Moshou, Giorgos Vasileiadis, Athanasios Balafoutis, Panos M. Pardalos

Verlag: Springer International Publishing

Buchreihe : Springer Optimization and Its Applications

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

This volume is the second (II) of four under the main themes of Digitizing Agriculture and Information and Communication Technologies (ICT). The four volumes cover rapidly developing processes including Sensors (I), Data (II), Decision (III), and Actions (IV). Volumes are related to ‘digital transformation” within agricultural production and provision systems, and in the context of Smart Farming Technology and Knowledge-based Agriculture. Content spans broadly from data mining and visualization to big data analytics and decision making, alongside with the sustainability aspects stemming from the digital transformation of farming. The four volumes comprise the outcome of the 12th EFITA Congress, also incorporating chapters that originated from select presentations of the Congress.
The first part of this book (II) focuses on data technologies in relation to agriculture and presents three key points in data management, namely, data collection, data fusion, and their uses in machine learning and artificial intelligent technologies. Part 2 is devoted to the integration of these technologies in agricultural production processes by presenting specific applications in the domain. Part 3 examines the added value of data management within agricultural products value chain.
The book provides an exceptional reference for those researching and working in or adjacent to agricultural production, including engineers in machine learning and AI, operations management, decision analysis, information analysis, to name just a few.
Specific advances covered in the volume: Big data management from heterogenous sources Data mining within large data setsData fusion and visualizationIoT based management systemsData Knowledge Management for converting data into valuable informationMetadata and data standards for expanding knowledge through different data platformsAI - based image processing for agricultural systemsData - based agricultural businessMachine learning application in agricultural products value chain

Inhaltsverzeichnis

Frontmatter

Data Technologies

Frontmatter

You Got Data‥‥ Now What: Building the Right Solution for the Problem

Abstract

The demands placed upon the agri-food industry are becoming ever greater and ever more urgent, and food producers are turning to technology to provide solutions to maximizing production and productivity. In the last decade, there has been a rapid expansion in Information and Communications Technology that is now capable of answering these challenges and breaking free of dependence upon manual labor and levels of human skill and experience that would take years or even decades to develop. This expansion has taken place at two essential levels. At a first level, it is possible to design and engineer individual sensors that can supply accurate measurements wherever they are placed and whenever they are in place. However, it is only when an array of sensors is deployed in a spatial network over an extended period, does the power of technology become apparent as through these networks remote and automatic control of production processes becomes realized. At a second level, it is now possible to design and engineer computing hardware and software to process the enormous datasets that these sensor networks generate. Furthermore, advancements in machine learning have facilitated the creation of predictive models to divine accurate process control decisions from these datasets.

Patrick Jackman

Data Fusion and Its Applications in Agriculture

Abstract

An information revolution is currently occurring in agriculture resulting in the production of massive datasets at different spatial and temporal scales; therefore, efficient techniques for processing and summarizing data will be crucial for effective precision management. With the profusion and wide diversification of data sources provided by modern technology, such as remote and proximal sensing, sensor datasets could be used as auxiliary information to supplement a sparsely sampled target variable. Remote and proximal sensing data are often massive, taken on different spatial and temporal scales, and subject to measurement error biases. Moreover, differences between the instruments are always present; nevertheless, a data fusion approach could take advantage of their complementary features by combining the sensor datasets in a manner that is statistically robust. It would then be ideal to jointly use (fuse) partial information from the diverse today-available sources so efficiently to achieve a more comprehensive view and knowledge of the processes under study. The chapter investigates the data fusion process in agriculture and its connection to artificial intelligence, neural networks, and IoT in agriculture, and introduces the concepts of data fusion with applications in Remote and Proximal sensing.

Dimitrios E. Moshou, Xanthoula Eirini Pantazi

Machine Learning Technology and Its Current Implementation in Agriculture

Abstract

Humans have always been intrigued by the notion that a machine could simulate their brain and mimic their actions. For that reason, through the last decades, artificial intelligence became the most prominent field of computer science, aiming to the development of intelligent machines, which are able complete tasks that require high level of cognition. Artificial Intelligence (AI) is a broad area comprised of advanced mathematical methods and computational techniques, such as machine learning and deep learning. Machine learning refers to the mathematical and algorithmic approaches that enable computers to automatically improve their efficiency in particular tasks, without being explicit programming. By analyzing large amount of data, and recognizing the patterns and structures within, machine learning is enables computers to iteratively learn and improve their efficiency without any human interaction. This chapter aims to an introduction towards understanding what machine learning is, by highlighting its differences with conventional programming and pointing out some of its fundamental features. Moreover, different types of machine learning algorithms are described, and examples are given in order to underline their importance in our everyday lives. Finally, a preliminary scholarly literature survey is presented, indicating studies that are referred in machine learning algorithms in the agricultural domain for the years 2018–2020. The study reveals that machine learning can undoubtedly expand our capabilities in many fields of expertise that affect our lives. Specifically in agriculture, machine learning solutions can improve quality of products and significantly increase operational productivity and efficiency.

Athanasios Anagnostis, Gabriela Asiminari, Lefteris Benos, Dionysis D. Bochtis

Applications

Frontmatter

Application Possibilities of IoT-based Management Systems in Agriculture

Abstract

The optimization of agricultural production and business processes is a crucial task in order to fulfill the demand of the increasing population, to meet quality requirements, to reduce the environmental impact as well as to improve economic efficiency. The Industry 4.0 concept provides various methods in this regard, including data acquisition based on IoT (Internet of Things), or data analytics based on Big Data, to support the decision-making process of the management and the data requirement of process control methods. During preliminary research, several modular data acquisition systems, as well as management applications have been developed based on a production system to measure various environmental factors at multiple spatial points. Considering the experience gained from the testing sessions, there was a need for further development regarding the end-user perspective in order to substantiate the practical application. A comparative research was required, considering previous experience and the literature of data acquisition systems, used in agriculture. The comparison concerned an own iteration of a production system and other systems, developed by researchers of the field, to examine different options and directions. Considering three important factors, the focus was on the data acquisition systems, data management, and data utilization methods. The comparison begins with a quantitative bibliometric analysis, determining the field and characteristic connections using network and cluster analysis, considering the IoT concept as the central element. Subsequently, the progression of a system and its evaluation is presented, performed in a greenhouse. This iteration highly focuses on data management with the modification of the existing infrastructure by integrating the Hadoop ecosystem to achieve a standardized interface.

Mihály Tóth, János Felföldi, László Várallyai, Róbert Szilágyi

Plant Species Detection Using Image Processing and Deep Learning: A Mobile-Based Application

Abstract

Conservation of biodiversity requires plant species identification skills, and automatic detection is a challenging and fascinating task for both computer/data scientists and botanists alike. This chapter describes a deep learning Convolutional Neural Network (CNN), which is trained to perform mobile imagery classification on plant species found throughout Ireland. The dataset of plant-classified RGB images underwent significant pre-processing, particularly in relation to background removal and data augmentation. Several models of deep learning CNN, with varying amounts of layers and training methods, have been evaluated on this dataset. Several deep learning models were trained and evaluated to document the speed and robustness of the flora identification. The highest performing model was then embedded in a web application, creating an online system to allow for new plant images to be uploaded and classified. This chapter highlights the main research challenges associated with this work, concludes with a mobile-based application, and discusses future research.

Eleni Mangina, Elizabeth Burke, Ronan Matson, Rossa O’Briain, Joe M. Caffrey, Mohammad Saffari

Computer Vision-based Detection and Tracking in the Olive Sorting Pipeline

Abstract

A procedure that is currently mostly handled by manual labor is olive fruit sorting by ripeness, as indicated by the color of each individual olive. Small and medium olive tree owners in olive-producing countries have to employee a significant number of workers for several days to perform this tedious task. Big industrial machines, capable of sorting do exist, but their cost is prohibitively high. These machines consist of two main components (a) the computer vision algorithm that is responsible for the detection of the olive fruit that are moving on a conveyor belt in lines and (b) a mechanical part that is responsible to sort the olives. However, advancements in computer vision allow for the implementation of low-cost solutions for the detection of olives. In this work, we present an automated solution that, by using computer vision, can perform the task of detecting the moving olive fruit effectively. Centroids of moving olive fruit are extracted through the Watershed Transform and by employing an Unscented Kalman filter, their position is estimated and tracked in consecutive video frames. Simulation experiments are designed, and the method is tested in various conditions to study its performance. Results suggest that the proposed approach tracks individual olives in synthetic videos created by scrolling images of olives with high accuracy. Even in the presence of induced noise, that resembles motion traces in images, the procedure remains capable of detecting and tracking olive fruit that are not trivially detected by the human eye.

George Georgiou, Petros Karvelis, Christos Gogos

Integrating Spatial with Qualitative Data to Monitor Land Use Intensity: Evidence from Arable Land – Animal Husbandry Systems

Abstract

This chapter investigates changes in land use intensity in a crop-livestock farming system on Lemnos Island through the combination of land use/land cover (LULC) types extracted from black and white (B&W) aerial images, statistics, and qualitative data. Combining quantitative and qualitative data, different insights of land use and landscape changes are assessed. The steps are: first, the timeline of historical changes was compiled; then, remote sensing was used to assess land cover changes and conversions that represent changes in intensity, and this information was complemented by participatory mapping with local farmers. Land use trajectories revealed that extensification is the basic trend from 1960 to 2002. Intensification coexists in at the same time, as grasslands convert to crops. There is a distinct pattern between periods: extensification seems to be the main process of change from 1960 to 1980 affecting mainly the hilly uplands as more remote and marginal areas are being converted from crops to grasslands or are abandoned, whereas intensification is the main trend for the next 20 years mainly in the lowlands as modernized agriculture (irrigated fields, land aggregation, machinery use) replaces more extensive land uses and traditional landscape elements such as tree hedges. The role of complementarity is very important. Conclusively, this case study shows that in some farming systems land use intensity changes cannot be represented through simple dichotomist differentiations.

Thymios Dimopoulos, Christos Vasilakos, Thanasis Kizos

Air drill Seeder Distributor Head Evaluation: A Comparison between Laboratory Tests and Computational Fluid Dynamics Simulations

Abstract

In this work, a commercial distributor head is evaluated. In parallel, both numerical simulations and laboratory tests, in a bench test belonging to the National University of Rosario, are carried out. This test bench has been built to evaluate components of air drill seeder’s pneumatic transport and distribution system. Soybean (Glycine max) seeds are used in the experimental tests. In Computational Fluid Dynamics (CFD) simulations, soybean seeds are modeled as spherical, rigid, and uniform size particles. The CFD simulations of the air-seed mixture are carried out with the commercial software ANSYS Fluent, and particle trajectories are numerically computed using a Lagrangian approach. A two-way coupling method is used, named Discrete Phase Model (DPM). Results show that numerical simulations are consistent with the laboratory tests, obtained in controlled trials. In both cases, the highest flow rates of seeds are produced in frontal outlets, while rear outlets present the lowest flow.

Ignacio Rubio Scola, Sebastián Rossi, Gastón Bourges

Value Chain

Frontmatter

Data-Based Agricultural Business Continuity Management Policies

Abstract

Data-driven decisions are crucial for modern enterprises regardless of the sector in which they operate. In agriculture, data processing, storage, and manipulation are crucial for boosting agricultural productivity. Nevertheless, the reliance of modern agriculture on information technologies has triggered a great concern regarding the exposure of agricultural processes to various threats that can cause unexpected interruptions. Business continuity deals with these types of threats. Data collection, storage, and processing which can be effectively implemented by modern business intelligence systems can undoubtedly help modern agricultural enterprises implement standard business continuity policies. The present chapter introduces a novel multidimensional approach for facilitating effective data-based business continuity management policies in agriculture. The approach relies on realistic business continuity data from two agrarian industries that are used for the design of two business intelligence multidimensional schemas which facilitate decisions based on descriptive data and for conducting data mining predictions. Examples of descriptive data-based decision-making processes are depicted using business process modeling notation tools and the predictive decisions are conducted via machine learning classifiers. In this way, agricultural business continuity experts in collaboration with agronomists, researchers, and farmers can be motivated to apply fully data-driven agricultural business continuity policies in specific agricultural companies.

Athanasios Podaras

Soybean Price Trend Forecast Using Deep Learning Techniques Based on Prices and Text Sentiments

Abstract

Predicting product prices is an essential activity in agricultural value chains. It can improve decision making and revenues for all agents. This chapter explores the use of deep learning techniques for predicting soybeans price trends in Brazil. A long short-term memory neural network (LSTM) forecasts the price signal. A convolutional neural network (CNN) generates a sentiment signal based on the sentiment analysis of news headlines. A multi-layer perceptron (MLP) is also evaluated to generate the sentiment signal, and an ensemble model, composed of both signals, prices and sentiment, is implemented. The four models (LSTM, CNN, and two ensembles with different weights for each signal) are evaluated in terms of their ability to predict the daily price trend. A hyperparameter analysis is conducted for all models, using the mean squared error (MSE) as a metric. Three models obtained the best result (0.60): (i) the LSTM alone; (ii) an ensemble model composed of a simple averaging of the signals; and (iii) an ensemble model composed of 90% price and 10% sentiment. The main findings are: (i) the analysis of the impact of hyperparameters on the models; (ii) the use of dictionaries has not significantly improved the sentiment prediction; (iii) the use of more than 50% of weight in the sentiment signal leads to worse predictions; and (iv) the CNN model provided a better sentiment signal than the MLP model. The benefits and possible uses of the models are discussed. The methodology used can be implemented for other products. Future work is related to improving data sets and implementing econometric models, unsupervised learning, and deep reinforcement learning.

Roberto F. Silva, Angel F. M. Paula, Gustavo M. Mostaço, Anna H. R. Costa, Carlos E. Cugnasca

Use of Unsupervised Machine Learning for Agricultural Supply Chain Data Labeling

Abstract

The heterogeneous data produced in agricultural supply chains can be divided into three main systems: (i) product identification and traceability, related to identifying production batches and locations of the product throughout the supply chain; (ii) environmental monitoring, considering environmental variables in production, storage and transportation; and (iii) processes monitoring, related to the data describing the production processes and inputs used. Data labeling on the different systems can improve decision-making, traceability, and coordination in the chains. Nevertheless, this is a labor-intensive task. The objective of this Chapter was to evaluate if unsupervised machine learning techniques could be used to identify patterns in the data, clusters of data, and generate labels for an unlabeled agricultural supply chain dataset. A dataset was generated through merging seven datasets that contained information from the three systems, and the k-means and self-organizing maps (SOM) models were evaluated on clustering the data and generating labels. The use of principal component analysis (PCA) was also evaluated together with the k-means model. Several supervised and unsupervised learning metrics were evaluated. The SOM model with the Gaussian neighborhood function provided the best results, with an F1-score of 0.91 and a more defined clusters map. A series of recommendations for the use of unsupervised learning techniques on supply chain data are discussed. The methodology used in this Chapter can be implemented on other supply chains and unsupervised machine learning research. Future work is related to improving the dataset and implementing other clustering models and dimensionality reduction techniques.

Roberto F. Silva, Gustavo M. Mostaço, Fernando Xavier, Antonio M. Saraiva, Carlos E. Cugnasca

Titel: Information and Communication Technologies for Agriculture—Theme II: Data
herausgegeben von: Dionysis D. Bochtis
Dimitrios E. Moshou
Giorgos Vasileiadis
Athanasios Balafoutis
Panos M. Pardalos
Verlag: Springer International Publishing
Electronic ISBN: 978-3-030-84148-5
Print ISBN: 978-3-030-84147-8
DOI: https://doi.org/10.1007/978-3-030-84148-5