Artificial Intelligence and Robotics-Driven Digital Agriculture

Artificial Intelligence and Robotics-Driven Digital Agriculture

Published: 2019.11.22
Accepted: 2019.11.28
110
Agricultural Robotics & Automation Research Center, Chonnam National University, Gwangju, Korea

ABSTRACT

The global agriculture and food industry are facing four main challenges: increased global population, constrained natural resources, climate change reducing productivity, and food waste causing market inefficiency and environment threat. Digital agriculture system (DAS), which uses the technologies of sensors, information communication, big data, artificial intelligence, unmanned aerial vehicle (UAV), and robotics, will be effective strategies that work for the challenges. In supporting DAS, UAVs can scan soil health, monitor crop health, assist in planning irrigation schedules, apply fertilizers, estimate yield data, and provide valuable data for weather analysis. Digital agriculture management solution, which connects all producers and objects into a cloud server, can help farmers remotely monitor and control their farm at anytime and anywhere. As farms are getting bigger, multi-unmanned ground vehicle (UGV) systems will be widely used in cooperation with UAVs. The cooperation between the UAV and UGV can detect obstacles in a larger field-of-view and monitor the operation of multi-UGVs. The core of DAS technology development is to integrate diverse digital technologies and stabilize the technologies within the agriculture environment. For this, a pilot digital agriculture complex should be built urgently.

Keywords: Digital agriculture, unmanned aerial vehicle (UAV), unmanned ground vehicle (UGV), digital agriculture management solution, pilot digital agriculture complex

INTRODUCTION

The global agriculture and food industry are facing several challenges. Clercq et al. (2018) and Trendov et al. (2019) introduced those challenges. With the global population projected to grow to almost 10 billion in 2050, there will be a significant increase in the demand of food. Farmers will have to produce 70 percent more food by 2050. And this food will need to be customized to the needs of a growing urban population. Although food needs and demand are increasing, the rural population is decreasing. At the same time, rural population are rapidly aging, which has major implications for the workforce, production pattern, land tenure, social organization within rural communities. Also, the availability of natural resources such as fresh water and productive arable land is becoming increasingly constrained.

Climate change is rapidly altering the environment. A side effect of climate change is an increase in the variability of precipitation and a rise in the frequency of droughts and floods, which tend to reduce crop yields. Climate change also impacts on global food security relating not merely to food supply, but also food quality, food access, and utilization (Clercq et al., 2018).

About 33-50 percent of all foods produced globally is never eaten, and the value of this wasted food is more than $1trillion. Food waste is a massive market inefficiency, the kind of which does not persist in other industries. Food waste is bad for the environment. It takes a land mass larger than Chian to grow food that ultimately goes uneaten. In addition, food that is never eaten accounts for 25 percent of all fresh water consumtion globally (Clercq et al., 2018). 

A digital agricultural revolution will be the newest shift which could help ensure agriculture meets the needs of the global population in the future. Digitalization will function in real time in a hyper-connected way, driven by data. Digital agriculture will create systems that are highly productive, anticipatory and adaptable to changes such as those caused by climate change. This, in turn, could lead to greater food security, profitability and sustainability (Trendov et al., 2019). 

DIGITAL AGRICULTURE

Digital agricultrue integrates new and advanced technologies into one system (United Nations Global Compact, 2019). The technologies used in digital agricultre system (DAS) include sensors, Information Communication Technology (ICT), big data, Artificial Intelligence (AI), Unmanned Aerial Vehicle (UAV), and robotics. The inputs of DAS are labor, chemicals, water, energy and the outputs are yield and the quality of product. DAS consists of three components (Fig. 1): sensing system for monitoring crop and field, decision making system for making optimum decisions in supplying inputs using big data and AI, and actuating system for delivering the decided inputs to crop and field precisely using UAVs and robots. The ouputs are feedbacked to the decision making system and the decision will be updated based on both the feedbacked outputs and the planned inputs. DAS keeps run through the feedback loop. It will be stablized when the inputs and outputs reach minimum and maximum level, respectively. Therefore, DAS enables farmers and other stakeholders within the agriculture value chain to improve production efficiency. It also enables to build trust between producer and consumer by sharing information gathered during agricultural production.     

UAV-based sensing system

DAS combines sensor data and imaging with real-time data analytics to improve farm producitivty through mapping spatical variability in the field (Food and Agriculture Organization of the United Nations, 2018). Data collected through UAVs provide the much-needed wealth of raw data to activate analytical models for agriculture. In supporting DAS, UAVs can scan soil health, monitor crop health, assist in planning irrigation schedules, apply fertilizers, estimate yield data, and provide valuable data for weather analysis. The image sensors commonly mounted on UAVs are RGB camera, multispectral and hyperspcectral camera, lidar, and multi-module AR camera as shown in fig. 2.       

UAVs were used to monitor the status of crops and fields in the full-cycle of rice production (Fig. 3): land prepartation, transplanting, disease and pest meanagement, and harvesting steps. In land preparation step, UAVs could measure the surface elevation of paddy field precisely using aerial photographic survey technique. These also could analyze the uniformity of plowing. Right after transplanting, UAVs took the image of paddy field at a low altitude and then could recognize each plant using deep learning technology. With the recognition of each plant, the distance between plants could be obtained and then the planting density (i.e. missing planting rate) based on the plant distance could be analyzed. Also, the position of each plant was given in the GPS coordinates. In growth step, diseases and pests could be diagnosed using the vegetaton indices.

Rice yield could be estimated by segmenting grain areas using low altitude UAV RGB images (Reza, et al., 2019). An image processing method that combined K-means clustering with a graph-cut algorithm was used to segment the rice grain areas. The graph-cut algorithm was applied to extract the foreground and background of the images. The foreground RGB images were converted to the Lab colour space and then K-means clustering was used to label pixels based on colour information. The area of the rice grains in the images was calculated from the clustered images (Fig. 3). The rice yield based on the grain area could be estimated.

Digital agriculture management solution

The leading companies in agriculture prodcution such as John Deere and Trimble provides digital agricultrue management solution, which connets all producers and objects into a cloud server, and therefore a farmer can remotely monitor and control the farm at anytime and anywhere. This new technology is changing the characteristics of the agricultural industy from hardware-orinted to software-orinted and from producer-oriented into consumer-orinted. 

The digital agricultural management solution was also developed by Agricultural Robotics and Automation Research Center (ARARC) at Chonnam National University in collaboration with Geospatial Informaton Cooperation in Korea. Through the management solution, data and statistics regarding crop and field management, analysis of farmland, and monitoring and control of agricultural machineries are provided. Users can access the management solution via mobile phone.  

Cooperation between UAV and UGV

In DAS, unmanned ground vehicle (UGV) will play a key role in farm operation. As farms are getting bigger, multi-UGV systems will be widely used in cooperation with UAVs. Currently, the fight time of UAVs is limited by the capacity of battery. Therefore, a wired UAV, which a UGV can supply power to through a wire, is connected to the UGV. The cooperaton between the UAV and UGV can detect obstacles in a larger field-of-view and monitor the operaton of multi-UGVs.

Pilot digital agriculture complex

The core of DAS technolgy development is to integrate diverse digital technologies, stabilize the integrated technologies within the agricultural environment, let the system teach itself, and to achieve high-efficiency of agricultural production. Within a pilot digital agriculture complex, researchers can test diverse and complicated external factors, accumulate long-term data, and develop solutions to counter possible problems. For this, it is necessary to build a pilot digital agriculture complex.  

CONCLUSION

With decreased arable land and water resources, the global agriculture and food system are challenged to produce enough food for an estimated 9.7 billion people by 2050. Digital agriculture system (DAS), which deploys intelligent information and automation technologies in farming, will play a key role in developing innovative solutions of the challenges. This review focused the work flow of DAS and the applications of DAS components. DAS consists of three components: sensing system for monitoring crop and field, decision making system for making optimum decisions in farming, and actuating system for delivering the decided inputs to crop and field. The ouputs of DAS are feedbacked to the decision making system and the decision will be updated based on both the feedbacked outputs and the planned inputs. DAS keeps run through the feedback loop. In sensing system, UAVs were used to monitor the status of crops and fields in the full-cycle of rice production: land prepartation, transplanting, disease and pest meanagement, and harvesting steps. Rice yield could be estimated by segmenting grain areas using low altitude UAV RGB images. An image processing method that combined K-means clustering with a graph-cut algorithm was used to segment the rice grain areas. The rice yield based on the grain area could be estimated. Digital agriculture management solution, which connects all producers and objects into a cloud server, was developed in Korea. To speed up the application of DAS to fields, a pilot digital agriculture complex, which can connect and integrate diverse digital technologies within agricultural environment, needs to be built.

REFERENCES

Clercq, M.D., A. Vats, and A. Biel. 2018. Agriculture 4.0: the future of farming technology (https://www.worldgovernmentsummit.org/api/publications/document?id=95df8ac4-e97c-6578-b2f8-ff0000a7ddb6;Accessed 1 Septemebr 2019).  

Food and Agriculture Organization of the United Nations. 2018. E-agriculture in action: drones for agriculture (http://www.fao.org/3/I8494EN/i8494en.pdf; Accessed 1 Septemebr 2019).  

Reza, M.N., I.S. Na, S.W. Baek, and K.H. Lee. 2019. Rice yield estimation based on K-means clustering with graph-cut segmentation using low-altitude UAV images. Biosystems Engineering 177: 109-121.

Trendov, M.M., S. Varas, and M. Zeng. 2019. Digital technologies in agriculture and reural areas (http://www.fao.org/3/ca4887en/ca4887en.pdf; Accessed 1 Septemebr 2019).  

United Nations Global Compact. 2019. Digital agriculture (http://breakthrough.unglobalcompact.org/disruptive-technologies/digital-agriculture/; Accessed 1 Septemebr 2019).  

Date submitted: October 29, 2019
Reviewed, edited and uploaded: November 28, 2019

Comment