One of the areas that are going through the most dramatic change in the break of the Fourth Industrial Revolution is agriculture. Today, more than ever, manufacturers are under huge pressure to increase yields, reduce costs and grow quality products. This is not an easy task and requires dedication, hard work and investments, and the only way to succeed is to make the decisions exceptionally efficient and timely. In the era of satellites, powerful computers and the internet, we finally have the opportunity to make decisions based on facts, and the key to success lies in data.


Sensors, yield monitors, drones and soil sampling probes have become a common sight on our fields, and they all are nothing more than different platforms for obtaining the data. Using them, we are getting information about the soil type, terrain elevation, crops growth and yields, and counting in the specifics of a particular field, we are optimising the choice of varieties and hybrids, planting date, amount and type of fertiliser, as well as the right moment for irrigation. In this way, we are providing all the necessary conditions for good plant growth and preventing excessive use of agrochemicals. However, a huge data problem with data is that until now the most of them have been hardly available for the farmers. They were located in different corners, in different formats, and some were not even available in the digital form. Today these problems are past. Satellite, meteorological and other data are widely available to Serbian farmers through AgroSense, the web platform of digital agriculture of Serbia, which gives them the opportunity to run their farms digitally, professionally and efficiently, completely free of charge.


AgroSense is a platform developed to ensure that the farmers have a quick and easy access to data, but also to let the community benefit from BioSense’s top class research in the domain of artificial intelligence, satellite image processing and the development of sensor networks. Since the platform launch in October 2017, we have gained more than 10,000 satisfied users, which confirms that Serbian farmers have recognised the importance of digital transformation in their area. The platform consists of a number of modules and their number is increasing constantly from one month to the other, to answer to the requirements in the production.



Using the service


Everyone interested in AgroSense can register at completely free of charge, for what you they only need to have the e-mail address. The first step after the registration is creating the production parcel. This can be done through the cadastre, if the user possesses the data about the municipality and the cadastral ID. The user can also select more than one parcel, if the crop production spans over a few of them. On the other hand, it often happens that multiple crops are produced on one cadastral parcel, which is why there is a possibility for the user to draw the boundaries of the production parcel using the map and a simple drawing tool. The user can have an unlimited number of production parcels on its profile and immediately after creating the parcel in the system, the whole range of modules become available. These modules are described in detail in the following chapters.



Meteorological data


Meteorological data are one of the key indicators of the state of the field and many activities are planned based on them. Soil temperature is a decisive parameter for sowing, rain forecast determines the optimal moment for the application of fertilisers and pesticides and solar radiation and air temperature are influencing the plant development. There are two modules in AgroSense that are dealing with meteorology. The first is devoted to the historical weather records. Data from the Nation Hidro-meteorological Agency are uploaded to the system and the user can choose up to 3 weather stations of interest, for which he can observe the soil and air meteorological parameters in the previous 7 days and get an insight about the state of the field. The second module is devoted to the weather forecast. Researchers at BioSense Institute have a long history of using the forecast issued by the Norwegian Meteorological Institute, as it is by our experience the most accurate one. What is more, it is given at a resolution of 2 km, which allows the user to have an accurate forecast for the location of its field.





Thanks to the involvement of BioSense in the European scientific projects and the success it had in its research, Serbia has become the only country outside of the EU that has the wideband access to imagery from Sentinel satellites of the European Space Agency, and AgroSense is fully utilising this opportunity. Sentinel satellites are taking footage of crops every 5 days at a resolution of 10 m and allowing the user to monitor its fields from a completely new perspective. Just as the human eye recognises 3 components of light (red, green and blue), Sentinels recognise 13 spectral components. This means that they see what we do not see and sometimes it is the infrared and thermal bands that are telling the most about the “health status” of the plants. Using 13 spectral channels, we are calculating vegetation indices that are telling us about the growth of crops, drought intensity, dispersion of plant diseases and the plants’ needs for irrigation and fertilisation. In this way, farmers can see which parts of the field are growing better and which are growing worse, and they can distribute the fertiliser according to the actual needs of crops in different parts of the field.



Digital field records (ili Digital diary of the production)


Every proper production should be well-documented. In agriculture, it is a practice to keep the field records and, as Serbia is transitioning towards the EU legislation, it will soon be obligatory. From the data analytics point of view, paper field records are a huge missed opportunity. Datasets regarding sowing, application of fertilisers and pesticides and harvesting are a “gold mine” for BioSense, but, as a rule, they are written in notebooks and planners and remain on the shelves, hidden from everyone’s eyes. With digital data, through application of advanced methods of artificial intelligence, machine learning and Big Data analytics, we are able to process huge amounts of data, find hidden dependencies within them, give automatic recommendations to the farmers and make their decisions more efficient. An old Chinese proverb says: “the best time to plant a tree was 20 years ago. The second best time is now”. Currently we do not have Big Data about agricultural production, but keeping digital field records is an ideal chance to start gathering information, based on which BioSense will develop new modules for the farmers. AgroSense’s Activities module allows users to enter the information about the date of the activities conducted on the field and specify details about the type of tillage, applied pesticides, yields and operation costs, while all these activities can be also documented with photographs using AgroSense mobile app. Besides the module with activities, there is a separate module for soil analysis that displays the soil data acquired through the years, and a module that consists of photos that the user has acquired using smartphone and their map based on their GPS locations.



Parameter maps


Besides satellites that are covered by a separate module, there are many other sources of data about the parcel, such as yield, elevation and soil electrical conductivity maps and drone footage. These maps are additional layers of information that can be extremely important in planning and monitoring the growth of crops. There is a possibility to import these maps into AgroSense and to compare them with one another. The image below is an example how the soil electrical conductivity maps are corresponding to the yield and it also shows the clear influence of centre-pivot irrigation system. On the other hand, drones are allowing us to observe crops on a resolution of just a couple of centimetres and spot the finest changes in crop growth. Additional parameter maps are giving us a new insight into the agricultural production, and AgroSense is giving us an opportunity to show them and analyse adequately.



Sensors and meteo stations


Scientific results are telling us that the weather is responsible for 70 % of the variability in yield. Due to the climate change, extreme weather events such as droughts and floods have become very frequent and have a distinctively local character. National Hidro-meteorological Agency’s network covers around 30 geographic locations, but this number is far from sufficient. It often happens that weather events occur between two weather stations and remain unregistered, which is a serious problem for farmers, who base their decisions in production on soil moisture, soil and air temperature and other parameters. For this reason, BioSense will deploy a staggering number of 500 weather stations throughout the Republic of Serbia, that will allow for measurements of meteorological parameters on a high spatial resolutions. When it comes to wireless integration of sensors and meteo stations, today’s state-of-the-art are LORA systems and such networks are often called IoT (the Internet of Things). Just as the mobile network is optimised for sending large files, such as high-quality images and video clips, LORA is optimised for sending small batches of data (sensory readings) over vast distances with small energy consumption. In this way, a sensor battery can last throughout the whole growing season and data logging is possible from fields up to 30 km away from the base station. Based on the readings from sensors and weather stations, farmers can have a precise insight into the state of the field through AgroSense and make optimal decisions in their production.



Cost analysis


One of the direct benefits of keeping comprehensive digital field records is that the cost analysis can be conducted by a single click of the mous. This module uses the information about the activities and their prices and lists the costs according to the crop type or the total costs for the whole farm. The cost analysis is supported by graphs that visually represent the numbers in a clear and understandable way, and which are helping the farmer get a better insight in the financial aspect of the production.





Consultancy is one of the fields that will undergo the biggest change due to the digital transformation, as the precise field data will allow advisory services and experts to give more accurate and more efficient advice to the farmer. Based on this module, advisors and farmers will be able to connect more efficiently and, upon the signing of the contract, the advisor will get a full range of information about the crops, from satellite images and parameter maps, over photographs from the field, all the way to the digital field records. Such a detailed dataset leaves a possibility to the expert to perceive the problem from multiple angles and determine the optimal steps in the production, often even without the field visit. Besides the module for advisors, drone pilots will also have their section within the Services module. Conducting business over AgroSense even eliminates the need for any personal contact between the providers and receivers of UAV imaging services. After establishing the contact, the pilot gets GPS coordinates and boundaries of the parcel, while the images are digitally uploaded to the platform. This makes the business run faster and more efficient and promotes the use of high-tech in agriculture. Service providers can limit the area of their operation to Serbian counties of interest, so that they could avoid deals from remote locations that are not financially justifiable, due to high travel costs. Another factor for choosing a service provider is the price. Namely, we set up a system for giving ratings to the providers that allows farmers to rate advisors and drone pilots based on the quality of work conducted.

This system puts quality service providers on top and allows them to set up their prices according to the quality of their service and the demand. Besides advisors and UAV pilots, we are planning to include sections for laboratories for soil analysis, companies for electrical conductivity scanning of the fields and others. This network will gather different actors in agricultural production and integrate them under one roof – the AgroSense platform.



Pest alarms


Plant diseases and pests near the parcel are a serious threat to the production and farmers need to act urgently upon information about their occurrence. This module logs the warnings issued by the Agricultural Prognostic and Reporting Agency of Serbia and sends them to the farmer. Warnings are filtered and the user sees only those related to diseases that affect the crops he is cultivating. In this way, the farmer is alarmed and warned that he needs to act according to the recommended procedures.





During the preparation for every operation on the field, farmers must calculate the necessary amounts of raw materials and calibrate the sprayers and fertiliser spreaders. This module allows them to do this precisely and quickly.





BioSense sees itself as a leader in the digital transformation of Serbian agriculture. AgroSense contains a large number of practical functionalities, but education is another important aspect. The Technologies module has a role to direct the farmer to the latest trends in agriculture. It contains articles about the modern fertiliser spreaders, tractors, irrigation systems and other products that have recently found their way to the market and represent today’s state-of-the-art.