Artificial Intelligence was shown its importance in other fields. It reduces the need of man power and increases efficiency. AI has recently entered in the agriculture field. Due to the rapid increase in global population we need to increase our production but the problem is we have only limited land for crop cultivation, there are also other problems such as irregular monsoon, uneven rainfall, weed, pest and disease problems. Artificial Intelligence played a vital role in overcoming these problems. With the help of AI we can manage soil health, weed, pest and crop disease. Robotics and automation reduce the agricultural budget and the need for resources. Drones can help in irrigation, fertility, weed management. We can also predict rainfall. This study outlines the use of AI in agriculture (such as drones, robots, digital twin, chat box, disease, weed, pest management) and the future scope of AI in agriculture.
Introduction
Earlier, agricultural activities were limited to the production of food and crop and food production. (M. Fan et al. 2012). However, in the last 2 decades, it was involved in crop, production, processing, distribution and marketing. It is also a vital role in the GDP of many countries (O. Oyakhilomen et al. 2014). According to the Food and Agriculture Organization in 2020 there are 720 to 811 million people who are facing undernourished problem in the world. So in percentage about 8.4 to 9.9% people are undernourished. And the population of people is increasing day by day. So we need to increase agricultural production. So to overcome this problem we use artificial intelligence.
Artificial Intelligence (AI) is a major field of study in computer science. With its rapid technological advancement and wide range of applications, AI is rapidly becoming widespread due to its strong applicability in problems that cannot be solved well by humans or traditional computing structures. (E. Rich et al. 1991). Artificial Intelligence (AI) has recently re-entered the public consciousness as a result of its impressive demonstrations of what it can now accomplish, incorporated into commonly used digital tools such as mobile phones, and clear indications by companies that it will be a key component of their future products and services (Wolfert et al. 2017).
Common and widespread examples of artificial intelligence include the development of powerful chatbot conversational interfaces with capabilities such as text understanding, speech interpretation, image recognition, and language translation to help humans figure out what they are looking for faster (Burgess 2017). The agriculture industry also needs AI because it has many features that make it a great target for AI; For example, their tendency to cover large and remote areas makes them good targets for AI-enabled remote-monitoring capabilities (Patrioccio and Rider 2018). In this review, I will describe various artificial intelligence methods used in agriculture such as soil management, crop management, weed management, pest management, disease management, and robotics and drones in agriculture.
AI used in agriculture
1. General crop management
Crop management systems serve as an interface for total crop management, covering all aspects of farming. The use of artificial intelligennce in crop management for the first time was proposed by Mackinnion and Leaman in their paper and the title of the paper was “Expert systems for agriculture” (J. M. Mackinnion et al. 1983). Boulanger proposed another corn crop protection expert system in his doctoral thesis. (A. G. Boulanger. 1983) Roch et al. et al. 1989). Italy forms severe frost (C. Robinson et al. 1997).
To train and test the network, both inputs and outputs were programmed in binary form. The authors used different input configurations to get the most accurate model. The best model discovered has a 94 retention accuracy with two output classes and six inputs. Lee, S. K. et al. proposed an image-based AI technique for wheat harvesting. (S. K. Lee et al. 2002). It uses a pixel labeling algorithm followed by a Laplace transformation to consolidate image information. One of the best networks acquired had five hidden units that were trained for up to 300,000 iterations and had an average accuracy of 85.9%. (C. Prakash et al. 2013).
2. Soil Management
Soil management is an essential component of agricultural operations. A thorough understanding of different soil types and conditions will increase crop yield while conserving soil resources. It refers to the application of operations, practices, and treatments to improve soil performance. Pollutants in urban soils can be examined using a conventional soil survey strategy (C.R.D. Kimpe et al. 2000) and manure application improves soil porosity and aggregation. A high level of aggregation indicates the presence of organic matter, which plays an important role in preventing soil crust formation. To prevent physical erosion of soil, alternative tillage systems can be used. The use of organic matter is important for improving soil quality C.R.D. . . . and silt content. (2005. Researchers developed an artificial neural network-based system to estimate soil moisture in paddy (C. Arif et al.
3. Insect Pests
Insect pest infestations are one of the most concerning issues in agriculture, resulting in significant economic losses. For decades, researchers have attempted to mitigate this threat by developing computerized systems capable of identifying active pests and recommending control measures. Several rule-based expert systems have been proposed, including Pascal and Mansfield, 1998, W.D. Batchelor et al. 1989, M. Mozni et al. 1993, J.D. Knight et al. 1994, B.D. Mehman et al. 2003, M. Lee et al. 2002, A.K. Chakraborti et al. 2013, Ghosh, 2015.
Because the knowledge involved in agricultural management is often incomplete, ambiguous, and imprecise, rule-based expert systems can lead to uncertainty. Several fuzzy logic-based expert systems including SAINI have been proposed to capture this uncertainty Ghosh et al. (2003) developed an expert system for pest management in tea, using an object-oriented approach to frame a rule base. A step-by-step identification and consultation process was also used in this case. After that Samanta and Ghosh modified this system using a multilevel back propagation neural network (R. K. Samanta, & Indrajit Ghosh, 2012).
4. Disease Management
There are many factors that affect the farm yield in which plant and animal diseases are the main factors. So we need to control the diease of PALNT for better yield. To completely control the diseases and reduce the losses, a farmer should use an integrated disease control and management model that integrates physical, chemical and biological controls. (B, 2018). Chemical and biological testing to accomplish these goals would be time consuming and not very cost effective (Weed Science Society of America) Computer-aided systems are being used worldwide to diagnose diseases and recommend control measures. Rule based systems, such as those developed by D.W. Boyd et al. 1994, S.K. Sarma et al., 2010, K, Balleda et al. 2014,. were developed at an early stage. Thus we need the use of AI approaches for disease control and management. Explanation Block (EB) provides a clear picture of a kernel of the expert system, followed by reasoning. (K. Balleda et al. 2014). Some hybrid systems have also been proposed. To classify Phalaenopsis seedling diseases, an image processing model combined with an artificial neural network model (K.Y. Huang et al. 2007), designed by (SS. Sannakki, et al. 2011). A scientist used a fuzzy logic approach in combination with image processing to detect the percentage of infection in a leaf. H. Al-Hari, et al. 2011, D. Al Bashish et al. 2011 created a system based on K-means segmentation algorithm.
5. Farmer Chatbots
Chatbots are communicative virtual assistants that automate interactions with end customers. Chatbots powered by artificial intelligence and machine learning techniques have enabled us to understand our native language and interact with users in a more personalized way. (Tanha Talviya et al. 2020).
6. Weed Management
Herbicide use has a direct impact on both human health and the environment. Modern AI methods are being used to reduce herbicide use through proper and precise weed management. Pascal. (G. M. Pascal, 1994). The researcher created a rule-based expert system for detecting and eradicating weeds in crops such as oats, barley, triticale and wheat. . (T. F. Burke, et al. 2005). Compared three different neural network models, mainly back propagation, counter propagation and radial basis function-based models, with the same inputs as the previous paper and discovered that back propagation network outperformed the others.
7. Agricultural Drones
Agricultural drones in a mechanical setting, driverless aeronautical vehicles (UAVs) or (UAS), also known as automatons, are unmanned aircraft that are controlled by remotes which they communicate with GPS and other sensors that are being installed in the drones for monitoring crop health, irrigation equipment, weed identification, wildlife monitoring and disaster management. In agriculture we can use remote sensing with the help of UAVs for the purpose of image capture, process and analysis. (Abdullahi et. al, 2015). Wireless sensor networks are used in the development of UAS (WSN). The data recovered by WSN allows UAS to advance in their use. For example, it can limit the use of synthetic compounds to specific areas. Because ecological conditions are constantly changing, the control cycle must almost certainly respond as quickly as can be reasonably expected. Harmonization with WSN can contribute to this goal. (B.S., Costa et al. 2014).
8. Digital Twin
On a farm, the amount of data about any single entity grows over time. The ‘digital twin’ is a new paradigm for organizing such information. (A digital twin is a real-world entity iteration like a specific cow, specific farm and filed. Digital twins give an organizing system to determine a good response to a particular query. Such digital twins will not only provide access to entities’ historical and near-real-time status, but will also make predictions about those entities. (Driessen C, Heutinck LF 2015).
9. Use of Robotics and Automation in Agriculture
The agriculture sector had to adapt to new breakthroughs and inventions. A scientist in the field of automation proposed a new field of research of embedded intelligence (EI). For a country to grow embedded intelligence smart crop management, smart farming and other agricultural technologies, intelligent irrigation and greenhouses, it has to incorporate these emerging technologies into its agriculture sector. Agriculture is important to many sectors of the economy. Furthermore, the authors of this paper demonstrated a technology roadmap (TRM) that clarifies the concerns about the above-mentioned agricultural sectors (smart farming, smart irrigation, etc.) (Yong, W. et al. 2018).
A researcher presented the use of the Losant platform for monitoring agricultural farmland and notifying farmers via SMS or e-mail if some anomalies are detected by the system. Losant is by far the most powerful cloud platform based on IoT. It allows real-time viewing of stored data. Mobile robots and automated equipment have already been efficiently used for some indoor agricultural activities, such as identifying animals and feeding them based on their nutritional needs and expected output, grouping and separating them, selectively feeding cows, shucking sheep, cleaning shelters, and slaughtering animals (Matias Collar Narock et al.
2009). Significant progress has also been made in outdoor activities, including establishment crops, plant care, and selective harvesting. Commercially available mobile robots are able to selectively harvest almost all types of fruits (strawberries, pears, grapes, watermelons), legumes, and flowers (Scott A., 2010). Automated guidance systems installed on conventional tractors are already in use, significantly reducing driver effort and being credited with improved technical and financial efficiency (Dave Franzen, 2009). A robotic driverless tractor able to follow a predefined route and react to unknown obstacles has been successfully tested, but it cannot be used in unattended areas at this time, especially for safety purposes (Scott A., 2010).
Milking robots, which first appeared in the industry in 1992, have been the only representative category of robots widely used in farming, and they rank second (after military robots) in the service robot class, accounting for 25% of all service robots. Between 2006 and 2009, the number of connected milking robots increased by 272%, from 6,180 to 22,980 units (European Robotics Research Network, 2008, 2010), and it is estimated that milking robots now account for around 20% of new milking installations in the UK (Dunn, 2009).
Future scope of AI in agriculture
The world population is projected to reach nine billion by 2050, requiring a 70% increase in agricultural production to meet demand. Only 10% of this higher production can come from uncultivated land, with the rest met by current production intensification. In this framework, the use of cutting-edge technological solutions to improve farming efficiency is a very critical and important need. Current agricultural intensification strategies require high energy inputs, while the market demands high-quality food. (D. G. Panapette, 2018). Artificial intelligence techniques are rapidly developing, and they can be used to detect plant disease or any unwanted weeds in the field using CNN, RNN, or any other computational network. Greenhouse farming can provide plants with a specific environment, but it is not possible without human intervention. Here, wireless technology and IoT come into play, and using the most recent communication protocols and sensors, we can implement weather control and monitoring without the need for human intervention on the farm. Fruit and crop harvesting can be integrated by robots that are specialized in working tirelessly for quick harvesting. (Kirtan Jha et al. 2019).
AI technology will be useful in predicting weather and other agricultural conditions such as land quality, crop cycle pest attack, groundwater land quality, groundwater, crop cycle, pest attack, and so on. The accurate projection or prediction made possible by AI technology will reduce the majority of farmers’ worries. AI-powered sensors are extremely useful for extracting important agricultural data. The information will be useful in improving production. These sensors have a huge potential in agriculture. Agricultural scientists can obtain data such as soil quality, weather, and groundwater levels, among other things, which will be useful in improving the farming process. AI enabled sensors can also be installed in robotic harvesting equipment to collect data. (Tanha Talvia et al. 2020).
Conclusion
Artificial intelligence can be the milestone of agriculture. It reduces the requirement of man power and other resource. AI can play a vital role to increase the yield of crops. It reduces the cost of crop cultivation. The problem of farmers was that the accurate weeding method outweighed the large amount of crops lost during the weeding process. These autonomous robots not only improve efficiency, but they also reduce the need for unnecessary pesticides and herbicides. In addition, farmers can use drones to effectively spray pesticides and herbicides on their fields, and plant monitoring is no longer necessary. A barrier to start-ups, resource and job scarcity, can be underestimated with the aid of man-made agribusiness brainpower.
Farming robots and autonomous equipment are still in their infancy. Only in greenhouses and dairy farms can robotic activity be considered relevant; for other activities, robots have only recently become commercially available or are in the prototype stage with little visibility and impact on global agriculture. New technologies in the robotics and automation fields still require time and investment to improve current technology. The rate and magnitude of adoption trends are difficult to predict. The world’s socioeconomic system is not considered to be changing at the same rate as technological scientific advances, preserving undue access to resources.
Read Also:
- A Glimpse About Artificial Intelligence (AI) In Agriculture
- Benefits And Challenges Of Artificial Intelligence (AI) In Agriculture
- Artificial Intelligence (AI) Advanced Interactive Systems
- Challenges And Future Of Adoption Of Artificial Intelligence (AI) In Educational Sectors
- Potential Of Artificial Intelligence (AI) In Healthcare
- Advantages And Disadvantages Of The Use Of Artificial Intelligence (AI) In Management
- Ethical, Legal And Other Challenges In The Use Of Artificial Intelligence (AI) In Health Care
- 50 Negative Effects Of Artificial Intelligence (AI) On Software Developers
- Disadvantages of Artificial Intelligence (AI) in Education
Leave a Reply