Book review: Tilling with IoT and machine learning
By Shalin Hai-Jew, Kansas State University
Nova Science Publishers
2021
243 pp.
Vishal Jain and Jyotir Moy Chatterjee’s Internet of Things and Machine Learning in Agriculture (2021) may be conceptualized as a Venn Diagram, with one circle as computational advances and the other agriculture, and a thin slice of overlap between the two circles. The Internet of Things (IoT) refers to interconnected devices that form various “networks” that enable awareness and controls for certain complex processes. The machine learning generally refers to various algorithms that enable the identification of information and patterns from various types of data, visual and textual and numerical, among others.
Some initial questions in starting this work may be the following:
Agriculture is a widespread and historic human endeavor that touches everyone’s lives. With a growing human population, “…by some estimates, worldwide food production will need to increase 70% by 2050 to keep up with global demand” (Jain & Chatterjee, 2021, p. vii), while also mitigating the stresses and damage to the environment (harms to land, fertilizer run-off to water, methane releases in the air, energy consumption, and others).
In the introduction, the co-editors explain that this work is designed as a reference work. As such, it offers a light introduction to IoT in smart agriculture, with many of the above questions still open ones.
Himani Mittal’s “Smart Farming Enabling Technologies: A Systematic Review” (Ch. 1) reads as two separate works, with one section highlighting recent innovations in artificial intelligence, machine learning, and the Internet of Things (IoT) generally speaking, and then lighting touching on some of the research literature on smart agriculture. The summary includes a superficial introduction to artificial neural networks, machine learning, search engine functions, image processing, handwriting recognition, facial recognition, object recognition, service robots, drug identification, and other technology tools without grounding them in the agricultural space. Some of the examples seem to be general observations without the necessary selectivity to focus on the applied technologies in agriculture. [Actual harnessing of tools for practical engineered purposes is difficult work.] IoT has been applied to smart cities, smart manufacturing, smart mobility, smart health, smart grid and energy, and smart homes (Mittal, 2021, pp. 9 – 10).
With technologies goes the argument, various agricultural tasks may be less labor-intensive and more “technology-intensive” to raise crop yield and quality (Mittal, 2021, p. 1). Some technologies in this space include technologies used in labs, fields and growing spaces, food processing facilities, animal processing facilities, packaging factories, transportation spaces, and others. The technologies used include sensors, cameras, machine vision, image analysis, unmanned aerial vehicles or drones, robots, automated machines (tractors), GPS technologies, and various analytics platforms.
There is the hope that smart farming helps “optimize costs, time and resources” and improve “the quality and variety of products” (Mittal, 2021, p. 2). Some components of smart farming include the following: “soil management, conditions for sowing, enhancing soil fertility, rainfall prediction, irrigation system and water management, pesticide control, weed management, crop diseases, harvesting, storage management and transportation of crops” (Mittal, 2021, p. 3). In this space, there is a need for sensors “for measuring temperature, luminosity, humidity, pressure, ground chemical concentration and soil moisture level” (Mittal, 2021, p. 12). The author also mentions devices, including “drones, video cameras, MIS software, global positioning systems and communication networks” (Mittal, 2021, p. 12). Computer vision involves using drones to manage crops, identify weeds, capture crop disease imagery,
Identify pests, and engage other tasks (Mittal, 2021). Remote sensing enables weather monitoring. There have to be technologies that enable clear communications and analytics with the collected information and perhaps decision supports for the farmers, ranchers, harvesters, food processors, and other workers in this space. This and other works would be stronger with the voices and experiences of those who work in agriculture and some real-world examples of how the technologies may be applied.
Figure 1. “Smart Agriculture” (by Lxz2208180358)
Mukta Sharma and Neha Aggarwal’s “Internet of Things Platform for Smart Farming” (Ch. 2) begins with a premise that human farming will have to change in order to achieve higher production and efficiencies to ensure sufficient food stocks into the future. The shift to different farming methods will require new capabilities, such as sampling the environment and making necessary changes. In the 12,000 years since plants and animals have been domesticated, various innovations have occurred, and the research and experimentation and innovations continue through the present (with biotechnology exploration, test plots, and other advances).
A generalized cycle of agriculture goes as follows: soil preparation, seed sowing, fertilizers and additives, irrigation, protection of crop against weeds, harvesting, and storage (Sharma & Aggarwal, 2021, pp. 21-22), with a range of complexities within each phase. Those who labor in agriculture are at the mercy of the climate in this Anthropocene age, geopolitics, Nature, but also global market pricing of agricultural commodities. Perhaps, precision agriculture or “digital agriculture” may be the difference in some contexts. Right now, the main focus has been on agribusiness scale.
In this work, IoT is referred to as “gadgets that trade data and can work with each other over the web” (Sharma & Aggarwal, 2021, p. 24). They are closely linked to robotization or the uses of machines to replace or supplement human labor. The authors write:
The farmers then are better and more accurately informed, and they can use this information to their advantage. Precision agriculture is practiced “to ensure profitability, productivity, and long-term sustainability by using big data to direct current and future decision-making” (Sharma & Aggarwal, 2021, p. 27).
There are various technologies on the general market for users. This work references particular companies and name brands for certain equipment. There are agri-bots “burrowing, weeding, preparing, picking, and splashing” to support automated farming (Sharma & Aggarwal, 2021, p. 30) and tools for smart greenhouses with microclimates. There are tools to monitor livestock, arrest food spoilage, and enable smart irrigation. There are self-driving combines or driverless tractors. [It is unclear how the current chip shortage may be affecting available supplies.]
This chapter highlights some challenges to deploying smart agriculture. There is a lack of wireless connectivity across swaths of rural lands, in the U.S. and elsewhere. Or the Internet connectivity may be spotty. IoT equipment is costly. Various devices are not interoperable; they may also not report relevant data to various IoT boards or mobile devices, so perhaps these are not yet mature technologies. Farmers and ranchers may lack “innovative knowledge” in this space. IoT devices, broadly, lack security given “assaults by creatures and hunters” and human hacking (Sharma & Aggarwal, 2021, pp. 32 - 34). For IoT systems to work smoothly, the devices have to be effectively tied to mapping services and data collection dashboards.
Whatever the current state of the art, there is high optimism for this space. By 2025, smart agriculture market is thought to be $15.3 billion (Sharma & Aggarwal, 2021, p. 38). However, smaller farms do not seem to bring in much in the way of funding, based on cited numbers: “According to National Geographic, in less than two years, a small farm of 800 acres can earn $11,000 per year, a medium-sized farm of 1600 acres can earn $26,000 per year, and a large farm of 2400 acres can earn $39,000 per year by properly using technology” (Sharma & Aggarwal, 2021, p. 38).
Md. Alimul Haque, Deepa Sonal, Shameemul Haque, and Kailash Kumar’s “Internet of Things for Smart Farming” (Ch. 3) highlights the importance of “data-centric” agriculture. In India, one of the ambitions for the uses of IoT for farming is to save cropland from fire even as it is not clear how IoT would inform that directly. Another ambition is for developing countries to remain up-to-date in using the latest technologies, which may be more about national pride than practical gains from the smart farming. This work offers a helpful summary of various types of sensing enabled by sensors for ascertaining location, optical, electrochemical, mechanical (such as to test soil compaction), dielectric soil moisture sensors, airflow sensors, and weather (Haque, Sonal, Haque, & Kumar, 2021, pp. 54-55) and other various combinations. There have not yet been apparent works showing the actual competitive advantage of intelligent farming, whether the gains are at the margins or something more critical. The hype in this work is that IoT will enable the feeding of billions, but that may read as hyperbolic, without further empirical evidence. The researchers here ask the question of what some practical use cases may be for smart farming and shed light on various practical challenges in operationalizing such an approach.
Historically, in human history, severe epiphytotics (plant diseases that are able to affect a wide range of plants) have led to mass losses of food crops and led to mass starvation. Plant diseases may come from biotic (bacteria, fungi, nematodes, animals, and others) and abiotic sources (non-living elements in the environment, such as light, radiation, temperatures, water, humidity, acidity, atmosphere, soil, and others), and viruses. Ideally, humanity would be able to quickly and reliably detect such onsets of diseases in order to engage in effective preventive measures.
Indu Sharma, Aditi Sharma, Inderjit Singh, Rahul Kumar, Yogesh Kumar, and Ashutosh Sharma’s “Plant Disease Detection Using Image Sensors: A Step Towards Precision Agriculture ” (Ch. 5) explores the potential for plant disease detection through image sensors, given the ambiguity of visual cues and a mix of potential underlying causes for various symptoms. Examples of relevant visuals include “changes in leaf colour, shape, morphology, transpiration rate, and plant density” (Sharma, Sharma, Singh, Kumar, Kumar, & Sharma, 2021, p. 91). The cameras and sensors may be hyperspectral and thermal, to enable multiple streams of information. The distribution of the symptoms may indicate something of disease incidence and spread. They write:
Once anomalies of interest are detected, perhaps samples are collected and tested in labs, records kept by the farmers and ranchers are checked, and plant pathologists and others analyze the data and suggest possible interventions. A variety of causes may underlie a lack of thriving in plants: mineral and other deficiencies, watering regimes, fertilization practices, environment events, biotic actions, and others. Sifting through the data to find actual root causes requires information, deep knowledge, and analytical skills. Then, this also assumes that there are the various potential interventions: chemical, mechanical, and others, that are cost-effective to address the risks in the least harmful ways to the environment and the farmers’ livelihoods.
S. Kannadhasan, R. Nagarajan, and M. Shanmuganantham’s “Recent Trends in Agriculture using IoT, Challenges and Opportunities” (Ch. 6) brings up the harnessing of blockchain technologies (more often linked to cryptocurrencies and non-fungible tokens or “NFTs”) to track “the entire agricultural product lifetime process,” with IoT harnessed alongside that to “track the entire commodity lifecycle to prevent the possibility of bad food safety and expiry” (p. 132).
Various devices used to detect disease in agriculture has to be accurate, with as low rates of false positives as possible (otherwise, it will drive human labor to be wasted at tracking down the false indicators). N. Ambika’s “Early Detection of Infection / Disease in Agriculture” (Ch. 7) describes the uses of an Arduino microcontroller board (with hardware and simple software) and a Raspberry Pi to link farm workers with sensors in the field for an analysis and response. This work describes various test frameworks deployed to fields for different types of monitoring and surveillance and interventions, to lay the groundwork. Then, the researcher proposes a different approach, which reads as theoretical more than applied.
Rohit Rastogi, Sunil Kumar Prajapati, Shiv Kumar, Satyam Verma, and Pardeep Kumar’s “Application of Agriculture Using IoT: Future Perspective for Smart Cities Management 5.0” (Ch. 8) focuses on what is needed for an automated smart irrigation system. Their system…
The world is in a global water crisis, with human uses outstripping supply in many cases, and with issues made worse with anthropocene-based global warming. Making raw water potable at population scale requires fuel energy and expensive infrastructure. With the human population growing at a fast rate, with billions more expected by 2050 and the need for double or triple the amounts of food available today, agriculture is under pressure to provide or risk catastrophic hunger. This work proposes the use of an Arduino microprocessor board linked to various sensors and the Android mobile operating system to create a simple agricultural irrigation system. This work uses effective data flow and other diagrams to explain the functioning. The title’s reference to Smart Cities 5.0 does not seem to fit with the content.
M.S. Sadiq, I.P. Singh, M. M. Ahmad, and N. Karunakaran’s “The Internet of Things (IoT) for Sustainable Agriculture” (Ch. 9) opens with a potent observation: “Most of the food consumed in developed nations is from half a billion small family farmers” (p. 191) and the others from big agriculture. [Various visuals are included in this and other chapters, but many of them are unintelligible without extreme zooming. Many seem to be merely decorative vs. informational.] This work also reads as a survey or overview piece focused on smart agriculture, with a focus on geospatial technologies, like GPS and GIS.
A. Firos’ “IoT Based Data Collection and Data Analytics Decision Making for Precision Farming” (Ch. 10) connects the possibilities of IoT in farming to farmer adaptation to “sudden changes in situation like atmospheric changes, monsoon, pest attacks” and perhaps to dealing with the “Green Revolution problem” (or sustainability) (p. 215), which may be debatable. Will local sensors be indicative of a monsoon sooner than global weather systems? The author highlights some practical points: “In some of the countries where we want to get aerial imagery and wanted to use a drone, we needed to get permission from the Ministry of Defense,” writes one researcher (Firos, 2021, p. 218). What happens is people go to more creative solutions, like tethered helium balloons, and they work through the problems they face, such as getting cameras to face the right way and other adjustments. There are discussions of various spaces where deeper oversight may benefit for plant and animal health, sustainability, grain storage, business efficiencies, and others. A general sequence of IoT in smart agriculture relies at heart on a “decision model” that optimally assesses the context accurately and informs decision-making in constructive ways.
Shalin Hai-Jew works as an instructional designer / researcher at Kansas State University. Her email is shalin@ksu.edu.
Internet of Things and Machine Learning in Agriculture
Vishal Jain and Jyotir Moy Chatterjee, Editors Nova Science Publishers
2021
243 pp.
When innovations are achieved in computer science and information technologies, these may be both theoretical and applied advances. They may offer proof of concepts, and they may be used in devices and practical applications in the world. If individual farmers only have 40 chances to get it right in a typical lifetime, each year, the more important thing is to get it right, with all available capabilities.
Introduction
Some initial questions in starting this work may be the following:
- What is IoT? What are the various device capabilities? How can these be cobbled for various practical applications, given the needs of agriculture in its various forms and phases?
- What do IoT and machine learning solve for those who work in agriculture (in any of a wide range of endeavors)?
- How well do the respective technologies work? How effectively are the technologies cobbled for agricultural purposes?
- How compelling is the business case for the various technologies? For packages of technologies?
- How mature is the field? In the hype cycle for emerging technologies, where is this space? Innovation Trigger? Peak of Inflated Expectations? Trough of Disillusionment? Slope of Enlightenment? Plateau of Productivity?
- In terms of technology adoption, who is involved? Innovators? Early adopters? Early majority? Late majority? Laggards?
- What are open problems in the agricultural space that may still be addressed with computation and engineering?
- Can technologies be used to identify “select agents” in crops? In animals raised for food?
Agriculture is a widespread and historic human endeavor that touches everyone’s lives. With a growing human population, “…by some estimates, worldwide food production will need to increase 70% by 2050 to keep up with global demand” (Jain & Chatterjee, 2021, p. vii), while also mitigating the stresses and damage to the environment (harms to land, fertilizer run-off to water, methane releases in the air, energy consumption, and others).
In the introduction, the co-editors explain that this work is designed as a reference work. As such, it offers a light introduction to IoT in smart agriculture, with many of the above questions still open ones.
Smart Farming
Himani Mittal’s “Smart Farming Enabling Technologies: A Systematic Review” (Ch. 1) reads as two separate works, with one section highlighting recent innovations in artificial intelligence, machine learning, and the Internet of Things (IoT) generally speaking, and then lighting touching on some of the research literature on smart agriculture. The summary includes a superficial introduction to artificial neural networks, machine learning, search engine functions, image processing, handwriting recognition, facial recognition, object recognition, service robots, drug identification, and other technology tools without grounding them in the agricultural space. Some of the examples seem to be general observations without the necessary selectivity to focus on the applied technologies in agriculture. [Actual harnessing of tools for practical engineered purposes is difficult work.] IoT has been applied to smart cities, smart manufacturing, smart mobility, smart health, smart grid and energy, and smart homes (Mittal, 2021, pp. 9 – 10).
With technologies goes the argument, various agricultural tasks may be less labor-intensive and more “technology-intensive” to raise crop yield and quality (Mittal, 2021, p. 1). Some technologies in this space include technologies used in labs, fields and growing spaces, food processing facilities, animal processing facilities, packaging factories, transportation spaces, and others. The technologies used include sensors, cameras, machine vision, image analysis, unmanned aerial vehicles or drones, robots, automated machines (tractors), GPS technologies, and various analytics platforms.
There is the hope that smart farming helps “optimize costs, time and resources” and improve “the quality and variety of products” (Mittal, 2021, p. 2). Some components of smart farming include the following: “soil management, conditions for sowing, enhancing soil fertility, rainfall prediction, irrigation system and water management, pesticide control, weed management, crop diseases, harvesting, storage management and transportation of crops” (Mittal, 2021, p. 3). In this space, there is a need for sensors “for measuring temperature, luminosity, humidity, pressure, ground chemical concentration and soil moisture level” (Mittal, 2021, p. 12). The author also mentions devices, including “drones, video cameras, MIS software, global positioning systems and communication networks” (Mittal, 2021, p. 12). Computer vision involves using drones to manage crops, identify weeds, capture crop disease imagery,
Identify pests, and engage other tasks (Mittal, 2021). Remote sensing enables weather monitoring. There have to be technologies that enable clear communications and analytics with the collected information and perhaps decision supports for the farmers, ranchers, harvesters, food processors, and other workers in this space. This and other works would be stronger with the voices and experiences of those who work in agriculture and some real-world examples of how the technologies may be applied.
Conceptualizing IoT Platforms for Smart Farming
Mukta Sharma and Neha Aggarwal’s “Internet of Things Platform for Smart Farming” (Ch. 2) begins with a premise that human farming will have to change in order to achieve higher production and efficiencies to ensure sufficient food stocks into the future. The shift to different farming methods will require new capabilities, such as sampling the environment and making necessary changes. In the 12,000 years since plants and animals have been domesticated, various innovations have occurred, and the research and experimentation and innovations continue through the present (with biotechnology exploration, test plots, and other advances).
A generalized cycle of agriculture goes as follows: soil preparation, seed sowing, fertilizers and additives, irrigation, protection of crop against weeds, harvesting, and storage (Sharma & Aggarwal, 2021, pp. 21-22), with a range of complexities within each phase. Those who labor in agriculture are at the mercy of the climate in this Anthropocene age, geopolitics, Nature, but also global market pricing of agricultural commodities. Perhaps, precision agriculture or “digital agriculture” may be the difference in some contexts. Right now, the main focus has been on agribusiness scale.
In this work, IoT is referred to as “gadgets that trade data and can work with each other over the web” (Sharma & Aggarwal, 2021, p. 24). They are closely linked to robotization or the uses of machines to replace or supplement human labor. The authors write:
The key to advanced rural development is an increment of inefficiency for every unit of land. Precision cultivating, otherwise called exactness horticulture, can be portrayed as whatever makes cultivating more controlled and exact, particularly with regards to developing crops and raising domesticated animals. It can help ranchers get more exact information about, say, the kind of seeds utilized, the measure of seeds utilized, supplements, compost, and water system per unit of land. Ongoing information from sensors, gear, climate, control frameworks, advanced mechanics, self-governing vehicles, mechanized equipment, variable rate innovation, water, GIS, and GPS can assist with land investigation…dampness, water system or waste necessities, and arranged versus real harvest yields, in addition to other things. (Sharma & Aggarwal, 2021, pp. 25 - 26).
The farmers then are better and more accurately informed, and they can use this information to their advantage. Precision agriculture is practiced “to ensure profitability, productivity, and long-term sustainability by using big data to direct current and future decision-making” (Sharma & Aggarwal, 2021, p. 27).
There are various technologies on the general market for users. This work references particular companies and name brands for certain equipment. There are agri-bots “burrowing, weeding, preparing, picking, and splashing” to support automated farming (Sharma & Aggarwal, 2021, p. 30) and tools for smart greenhouses with microclimates. There are tools to monitor livestock, arrest food spoilage, and enable smart irrigation. There are self-driving combines or driverless tractors. [It is unclear how the current chip shortage may be affecting available supplies.]
This chapter highlights some challenges to deploying smart agriculture. There is a lack of wireless connectivity across swaths of rural lands, in the U.S. and elsewhere. Or the Internet connectivity may be spotty. IoT equipment is costly. Various devices are not interoperable; they may also not report relevant data to various IoT boards or mobile devices, so perhaps these are not yet mature technologies. Farmers and ranchers may lack “innovative knowledge” in this space. IoT devices, broadly, lack security given “assaults by creatures and hunters” and human hacking (Sharma & Aggarwal, 2021, pp. 32 - 34). For IoT systems to work smoothly, the devices have to be effectively tied to mapping services and data collection dashboards.
Whatever the current state of the art, there is high optimism for this space. By 2025, smart agriculture market is thought to be $15.3 billion (Sharma & Aggarwal, 2021, p. 38). However, smaller farms do not seem to bring in much in the way of funding, based on cited numbers: “According to National Geographic, in less than two years, a small farm of 800 acres can earn $11,000 per year, a medium-sized farm of 1600 acres can earn $26,000 per year, and a large farm of 2400 acres can earn $39,000 per year by properly using technology” (Sharma & Aggarwal, 2021, p. 38).
Another Take on Smart Farming
Md. Alimul Haque, Deepa Sonal, Shameemul Haque, and Kailash Kumar’s “Internet of Things for Smart Farming” (Ch. 3) highlights the importance of “data-centric” agriculture. In India, one of the ambitions for the uses of IoT for farming is to save cropland from fire even as it is not clear how IoT would inform that directly. Another ambition is for developing countries to remain up-to-date in using the latest technologies, which may be more about national pride than practical gains from the smart farming. This work offers a helpful summary of various types of sensing enabled by sensors for ascertaining location, optical, electrochemical, mechanical (such as to test soil compaction), dielectric soil moisture sensors, airflow sensors, and weather (Haque, Sonal, Haque, & Kumar, 2021, pp. 54-55) and other various combinations. There have not yet been apparent works showing the actual competitive advantage of intelligent farming, whether the gains are at the margins or something more critical. The hype in this work is that IoT will enable the feeding of billions, but that may read as hyperbolic, without further empirical evidence. The researchers here ask the question of what some practical use cases may be for smart farming and shed light on various practical challenges in operationalizing such an approach.
Smart Systems to Fight Pests and Plant Diseases
Figure 2. Tilled Fields
G. Rekha and C. Sarada’s “A Comprehensive Review on Intelligent Systems for Mitigating Pests and Diseases in Agriculture” (Ch. 4) is based out of India, “the largest producer of oilseeds, pulses, rice, wheat, sugarcane, and cotton” (p. 63) and a country with a majority of peoples in the country depending on agriculture for their livelihoods. Information and communication technologies may be used to inform farmers of the risks to their crops and perhaps the very earliest phases of possible infestations or diseases, so they may take mitigating actions. What follows are some summaries of general concepts from machine learning and then the IoT. Then, there are summaries of research with available longitudinal data analyzed to understand patterns of threats to general and particular food crops. Such models clearly require rich information to create the data-informed predictive models. In this particular application, various models are used to predict impending disease (and underlying causal factors), perhaps even before any signs show up. The idea is to identify subtle but accurate signals, on which decisions may be made and actions may be taken. Over time, it would be possible to collect the cumulative insights into more complex applied models for in-world decision-making. (Some of the described approaches are for very particular food crops defending against very particular and known threats.) Using digital imagery alone, combinations of machine learning, deep learning, and convolutional neural networks have been found to be quite effective.
Image Sensing for Plant Disease Detection
Historically, in human history, severe epiphytotics (plant diseases that are able to affect a wide range of plants) have led to mass losses of food crops and led to mass starvation. Plant diseases may come from biotic (bacteria, fungi, nematodes, animals, and others) and abiotic sources (non-living elements in the environment, such as light, radiation, temperatures, water, humidity, acidity, atmosphere, soil, and others), and viruses. Ideally, humanity would be able to quickly and reliably detect such onsets of diseases in order to engage in effective preventive measures.
Indu Sharma, Aditi Sharma, Inderjit Singh, Rahul Kumar, Yogesh Kumar, and Ashutosh Sharma’s “Plant Disease Detection Using Image Sensors: A Step Towards Precision Agriculture ” (Ch. 5) explores the potential for plant disease detection through image sensors, given the ambiguity of visual cues and a mix of potential underlying causes for various symptoms. Examples of relevant visuals include “changes in leaf colour, shape, morphology, transpiration rate, and plant density” (Sharma, Sharma, Singh, Kumar, Kumar, & Sharma, 2021, p. 91). The cameras and sensors may be hyperspectral and thermal, to enable multiple streams of information. The distribution of the symptoms may indicate something of disease incidence and spread. They write:
The inability of the human eye to work beyond the visible range of the electromagnetic spectrum limits the visual assessment to a short region of the spectrum. Various optical sensors viz., RGB (red, green, and blue) wavebands, 3D-imaging, chlorophyll-fluorescence imaging, thermography, and multispectral (MSI) and hyperspectral imaging (HIS) have recently been used for the detection of plant diseases” for early detection of plant disease and forecasting progressions and suggesting possible interventions. (Sharma, Sharma, Singh, Kumar, Kumar, & Sharma, 2021, p. 91)
Once anomalies of interest are detected, perhaps samples are collected and tested in labs, records kept by the farmers and ranchers are checked, and plant pathologists and others analyze the data and suggest possible interventions. A variety of causes may underlie a lack of thriving in plants: mineral and other deficiencies, watering regimes, fertilization practices, environment events, biotic actions, and others. Sifting through the data to find actual root causes requires information, deep knowledge, and analytical skills. Then, this also assumes that there are the various potential interventions: chemical, mechanical, and others, that are cost-effective to address the risks in the least harmful ways to the environment and the farmers’ livelihoods.
Analyzing Trends in Modern Agriculture
S. Kannadhasan, R. Nagarajan, and M. Shanmuganantham’s “Recent Trends in Agriculture using IoT, Challenges and Opportunities” (Ch. 6) brings up the harnessing of blockchain technologies (more often linked to cryptocurrencies and non-fungible tokens or “NFTs”) to track “the entire agricultural product lifetime process,” with IoT harnessed alongside that to “track the entire commodity lifecycle to prevent the possibility of bad food safety and expiry” (p. 132).
Cobbling Technologies to Detect Disease in Agriculture
Various devices used to detect disease in agriculture has to be accurate, with as low rates of false positives as possible (otherwise, it will drive human labor to be wasted at tracking down the false indicators). N. Ambika’s “Early Detection of Infection / Disease in Agriculture” (Ch. 7) describes the uses of an Arduino microcontroller board (with hardware and simple software) and a Raspberry Pi to link farm workers with sensors in the field for an analysis and response. This work describes various test frameworks deployed to fields for different types of monitoring and surveillance and interventions, to lay the groundwork. Then, the researcher proposes a different approach, which reads as theoretical more than applied.
Learning from Smart Agricultural Irrigation for Smart Cities 5.0?
Rohit Rastogi, Sunil Kumar Prajapati, Shiv Kumar, Satyam Verma, and Pardeep Kumar’s “Application of Agriculture Using IoT: Future Perspective for Smart Cities Management 5.0” (Ch. 8) focuses on what is needed for an automated smart irrigation system. Their system…
…aims to maintain an adequate amount of water needed by the crop by monitoring the amount of soil moisture, temperature, and humidity in the soil. Data of temperature and humidity is maintained in the database for backup. The data is used for crop rotation and also helps the farmer with the selection of appropriate crops. We can also verify the different types of soil appropriate for different crops using this model. (Rastogi, Prajapati, Kumar, Verma, & Kumar, 2021, p. 172)
The world is in a global water crisis, with human uses outstripping supply in many cases, and with issues made worse with anthropocene-based global warming. Making raw water potable at population scale requires fuel energy and expensive infrastructure. With the human population growing at a fast rate, with billions more expected by 2050 and the need for double or triple the amounts of food available today, agriculture is under pressure to provide or risk catastrophic hunger. This work proposes the use of an Arduino microprocessor board linked to various sensors and the Android mobile operating system to create a simple agricultural irrigation system. This work uses effective data flow and other diagrams to explain the functioning. The title’s reference to Smart Cities 5.0 does not seem to fit with the content.
IoT for Sustainable Agriculture
M.S. Sadiq, I.P. Singh, M. M. Ahmad, and N. Karunakaran’s “The Internet of Things (IoT) for Sustainable Agriculture” (Ch. 9) opens with a potent observation: “Most of the food consumed in developed nations is from half a billion small family farmers” (p. 191) and the others from big agriculture. [Various visuals are included in this and other chapters, but many of them are unintelligible without extreme zooming. Many seem to be merely decorative vs. informational.] This work also reads as a survey or overview piece focused on smart agriculture, with a focus on geospatial technologies, like GPS and GIS.
Precision Farming for Green Revolution Concerns and Crop Protection
Figure 3. Leaves
Conclusion
Agriculture is central to people’s survival, but by nature, it is under threat from various factors. There are apparent limits to production. Every advantage should be harnessed with clear thinking about effects of every intervention. Vishal Jain and Jyotir Moy Chatterjee’s Internet of Things and Machine Learning in Agriculture (2021) may help readers warm up to harnessing technologies for agriculture in various practical and efficient ways, in the developed and developing world.
About the Author
Shalin Hai-Jew works as an instructional designer / researcher at Kansas State University. Her email is shalin@ksu.edu.
| Previous page on path | Cover, page 16 of 22 | Next page on path |
Discussion of "Book review: Tilling with IoT and machine learning"
Add your voice to this discussion.
Checking your signed in status ...