Artificial Intelligence (AI) in Agriculture: Revolutionizing Farming for a Sustainable Future

Introduction 

In the 19th century during the times of the industrial revolution machines were deployed as a substitution or reduction for human labor. This in the course of time, with the advancements in information technology in the 20th century, post the arrival of the computers, initiated the vision for artificial intelligence (AI) powered machines. In the present day, it’s a reality that AI is tardily taking over human labor. 

What AI in Agriculture?

involves the development of computer systems for performing tasks that normally require human intelligence which aims at increasing productivity in the agricultural sector.

Scope 

In agriculture, there is a quick adaptation to AI in its various farming techniques, the concept of cognitive computing is one that imitates human thought processes as a model in computers. This results in turbulent technology in AI-powered agriculture, rendering its service in interpreting, acquiring, and reacting to different situations (based on the learning acquired) to enhance efficiency. To harvest benefits in the field by catching up with the recent advancements in the farming sector, the farmers can be offered solutions via platforms like Chatterbot. At present in India, Microsoft Corporation is working in the state of Andhra Pradesh with 175 farmers rendering services and solutions for land preparation, sowing, addition of fertilizers, and other nutrient supplements for

crops. On average, a 30% increase in crop yield per ha has already been witnessed in comparison to the previous harvests. The various areas where the solutions for benefitting agriculture involving cognition possess knowledge are furnished below. 

The Internet of Things (IoT) driven development 

There are massive volumes of data getting generated each day in structured and unstructured formats, these data are regarding weather patterns, soil reports, new research, rainfall, vulnerability to pest attacks, and imaging through drones and cameras. IoT solutions relating to cognition would sense, recognize, and yield innovative solutions to enhance crop yields. Two primary technologies have been deployed for intelligent data fusion, namely proximity and remote sensing which is meant for testing the soil. Unlike remote sensing, proximity sensing doesn’t need sensors to be built into aerial or satellite systems; it only requires sensors that are in contact with the soil at a close range. This facilitates the characterization of the soil based on the soil beneath the surface at a particular region. Hardware solutions like Rowbot (concerning crops like corn) have already begun pairing software that collects data with robotics to develop the best fertilizer for the cultivation of corn to maximize the most possible crop yield. 

Image-based insight generation 

In the current world scenario one of the most dissertated areas in farming today is Precision farming, imaging through drones can assist in rigorous field analysis, monitoring crops, and scanning of fields. A combination of Computer vision technology, drone data, and IoT will ascertain that the farmers take rapid actions. Data fed from drone images could bring forth alerts in real-time which would accelerate precision farming. Commercial drone makers like Aerialtronics have enforced the IBM Watson IoT Platform and the Visual Recognition APIs for real-time image analysis. Some areas in computer vision technology that can be put to use are as follows, 

• Identification of disease 

Images of plant leaves are segmented into surface sections like background, infected area, and non-diseased portion of the leaf using image sensing and analysis to make sure this happens. The unhealthy or infected area is then cut off and sent to the lab for additional diagnosis to provide further support in identifying pests and detecting nutrient deficiencies.

• Assessing Crop Ripeness Using White Light and UVA Illumination

Farmers are employing innovative techniques to gauge the readiness of their crops by capturing images of various agricultural products under both white light and UVA (ultraviolet A) light which allows them to determine the optimal stage of ripeness for green fruits and other crops, helping farmers categorize the produce into distinct readiness levels and then arrange them in assorted stacks for efficient distribution to the market.

• Field supervision 

By creating a field map and identifying the places where the crops need water, fertilizer, and pesticides, real-time calculations may be made during the period of cultivation using high-definition photos from drone and copter systems to aid in the resource optimization process.

Identification of optimal mix for agronomic products 

Cognitive solutions recommend the farmers on the best choice of crops and hybrid seeds which is grounded on multiple parameters like soil condition, weather forecast, type of seeds, and pest infestation in a specific area. A personalized recommendation based on the farm’s requirements, native conditions, and data pertaining to successful farming in the past. Other external factors like trends in the marketplace, crop prices, consumer needs, requirements, and aesthetics may also be factored in to enable farmers to make a clued-up decision. 

Crop health surveillance 

Building agricultural metrics across thousands of acres of arable land requires the use of remote sensing (RS) techniques, hyperspectral imagery, and 3D laser scanning from a time and effort standpoint, it has the potential to bring about a revolutionary change in how farmers monitor their farmlands and also this technology will be used to keep an eye on crops during their entire life cycle, including the creation of reports in the event of anomalies. 

Automation techniques in irrigation and enabling farmers 

AI-trained machines aware of historical weather patterns, soil quality and the kind of crops to be grown, can automate irrigation and increase overall yield. Nearly 70% of the world’s freshwater resource is utilized for irrigation; such automation can conserve water and benefit farmers in managing their water probs. 

Significant of drone 

According to a recent PWC (Price Water House Coopers) study, the total available market for drone-based solutions throughout the world is $127.3 billion. And for agriculture is at $32.4 billion, drone-based solutions in the agriculture sector have a lot of implications like dealing with adverse climatic conditions, productivity gains, precision farming, and crop yield management. 

Fig.1 Disease Detection  

Fig.2 Plant Stress recognition using machine learning and intelligence 

                                      

Fig 3 Robotics in Digital Farming

A detailed 3D map of the field, its terrain, irrigation drainage, and soil viability must be developed using the drone. This has to be carried out before the crop cycle begins. The soil N2 levels management can also be done by solutions powered by drones. Drone-powered aerial spraying of pods with seeds and plant nutrients into the soil supplies necessary supplements for plants, Also drones can be programmed to atomize liquids by regulating the distance from the ground surface depending on the terrain. Crop monitoring and crop health assessment prevail as one of the most important domains in agriculture to offer drone-based solutions in coactions with computer vision technology and AI. Drones with high-resolution cameras gather precision field images that can flow through convolution neural networks to detect areas with weeds, individual crops requiring more water, and plant stress levels in various growth stages.

In the case of infected plants, by scanning crops in both RGB (Red Green Blue) and infra-red light, potential multispectral images can be generated using drone devices. Through this individual and specific cluster of plants infected in any region of the field can be spotted and supplied with remedies at once. The multi-spectral images taken from the drone cameras blend hyperspectral images with 3D scanning techniques to define the spatial information system employed for acres of farmland. This renders guidance throughout the lifecycle of the plant as a temporal component.

Precision farming 

Refers to a controlled technique of farming that substitutes the repetitive and labor-intensive part of farming, providing guidance regarding crop rotation. These distinguished key technologies that enable precision farming are high precision positioning systems, geological mapping, remote sensing, integrated electronic communication, variable rate technology, optimum planting and harvesting time estimator, water resource management, plant and soil nutrient management, and attacks by pests and rodents. 

Goals for Precision Farming Profitability 

Recognize crops and market strategically as well as prefiguring ROI (Return on Investment) based on cost and gross profit. 

Efficiency 

By putting in a precision algorithm, improved, rapid, and low-cost farming opportunities can be utilized. 

Sustainability 

Better socioeconomic and environmental operation assures additive improvements in each season for all the performance indicators. 

Cases of Precision Farming Management 

The detection of different stress levels in a plant via high-resolution images and multiple sensor data by AI. This entire set of data generated from multiple sources needs to be utilized as input data for AI machine learning. This enables the fusion of these data and feature identification parameters for plant stress recognition (Figure 2). AI machine learning models developed are trained on a wide range of plant images and can recognize the different levels of stress in plants. This total approach can be categorized into four sequential stages recognition, categorization, quantification, and forecasting to make better and improved decisions (Figure 2). 

AI adoption in agriculture Challenges 

There is currently a lack of understanding of cutting-edge high-tech machine learning solutions at farms all over the world. Agriculture is highly exposed to external elements such as weather, soil, and insect attack vulnerability. A crop-raising plan that was set up at the beginning of the season could not look ideal when harvesting begins because it is affected by outside factors. To train their computers and provide accurate forecasts or predictions, AI systems also need a large amount of data. The collection of spatial data is simple for very large agricultural land areas, but the collection of temporal data is more difficult. Only during the growing season of the crops could the various crop-specific data be collected. 

Building a reliable AI machine-learning model requires a significant amount of time since the database needs time to mature. The use of AI in agronomic items like seeds, fertilizer, and insecticides rather than on-field precise solutions is largely due to this. 

Conclusion

While using AI decision-making systems and predictive solutions to address the real-world demands and issues faced by farmers, farming with AI is still in its infancy.  

It will then be able to manage frequent shifts and changes in the environment on its own. This would make it easier to make decisions at the moment and use the right models and programs sequentially to collect contextual data effectively. The exorbitant cost of the many cognitive farming solutions that are readily available on the market is the other important factor, to guarantee that this technology reaches the farming community, AI solutions need to become more commercially viable. 

Dharmaraj, V. and Vijayanand, C. 2018. Artificial Intelligence (AI) in Agriculture. Int.J.Curr.Microbiol.App.Sci. 7(12): 2122-2128. doi: https://doi.org/10.20546/ijcmas.2018.712.241