6 Predictive models in precision agriculture

Global overview

Fast-growing patent activity mainly focusing on North America

Analysis of data from 1,529 international patent families in the field of Predictive models in precision agriculture shows a substantial CAGR of +27.1% from 2017 to 2021, indicating a surge in interest in the subject matter (Figure 2.22).

The USPTO, WIPO and EPO are widely recognized as the leading global authorities for patent filings, with 1,138, 1,055, and 823 international patent families respectively (Figure 6.1). Asia is also rapidly emerging as a significant contributor in the field of Predictive models in precision agriculture. China leads the pack with 635 international patents, followed closely by Japan with 391 international patents and the Republic of Korea with 284 international patents. This trend highlights the region's increasing presence and influence in the global market.

On the other hand, North America sees Canada contributing 459 international patent families, while Latin America and the Caribbean are represented by Brazil with 472 international patent families. Oceania is also making its mark, with Australia filing 378 international patent families. It is important to note that this data only considers international patent families, which may not fully capture the impact of regional jurisdictions in Asia.

Inventive regions

North America as the innovation hub for Predictive models in precision agriculture

According to Figure 6.2, the United States is the leader in R&D for predictive models dedicated to the Agrifood sector, with a total of 809 international patent families. Following close behind is China with 136 international patent families, and Japan with 106 international patent families. All the regions have shown significant growth in this area, with a CAGR between 2017 and 2021, exceeding 20% for several countries. Specifically, EPO has seen a growth rate of 58.5%, Japan has experienced a growth rate of 53.4%, the Republic of Korea has seen a growth rate of 49.6%, India has grown at a rate of 35.1%, Germany has seen a growth rate of 29.7%, and China has grown at a rate of 27.9%.

Top players

Mixed panorama between AgriTech pure players and IT corporate heavyweights

The predictive model segment is largely dominated by industrial actors, with 90% of the market share. The top players in this segment originate from the United States, Germany and Japan, with global machine manufacturers and agrochemical companies leading the way (Figure 6.3).

As a global main actor for AgriTech in general, Deere is also the main innovator of predictive computational models as well as their integration within automated machine controls, holding 116 international patent families. Deere holds patents covering predictive crop state and characteristic mapping, predictive nutrient mapping, predictive harvesting models for machine control, predictive yield mapping and more. Deere exhibits a strong and diversified portfolio. However, this portfolio only shows extensions mainly in Brazil and Europe to a lesser extent.

Bayer and BASF are two German chemical and agrochemical companies, holding 72 and 53 international patent families respectively in the field of Predictive models in precision agriculture, ranking second and third. Bayer developed machine learning-based systems for soil analysis, multivariate algorithms for yield prediction in horticulture, image processing for disease detection in plants, fertility prescription and more. BASF shows a large expertise in crop protection and pest control based on infestation predictions, weather image capture complemented by analysis algorithms and prediction models for formulation properties. Both portfolios are strong and diversified, with worldwide extensions.

Mineral Earth Sciences and Alphabet are both US companies specializing in cutting-edge technology. They hold 21 and 12 international patent families respectively in this field, ranking fourth and sixth. Mineral Earth Science offers a wide range of services including soil composition prediction, crop yield prediction, climate forecasting and plant imaging analysis. In contrast, Alphabet has developed a comprehensive suite of solutions tailored specifically for aquaculture and fish feeding industries.

TBG AG (Thyssen-Bornemisza Group) is a Swiss company which acquired DTN (former Telvent and Schneider Electric) in 2017. TBG AG currently holds 10 international patent families in this field, ranking eighth. Together with Iteris, DTN developed models for soil compaction from diagnosis and prediction of soil and weather conditions to ultimately improve ground performance.

IBM (United States) shows a restricted portfolio related to the AgriTech domain from the development of IoT agricultural ecosystems, machine learning models for livestock value chain, crop disorder detection and systems for crop identification using satellite. It currently holds nine international patent families related to Predictive models in precision agriculture, ranking ninth.

TATA Sons PVT is an Indian actor involved in system development for crop features and quality management. Tata’s portfolio is mostly extended to the United States, Brazil and Europe. It currently holds seven international patent families related to Predictive models in precision agriculture, ranking fourteenth.

Zheijiang University and Climate Foundation are the two academic players of this list. Zheijiang University currently holds 12 international patent families related to Predictive models in precision agriculture, ranking fifth. This reflects the university's extensive expertise in yield prediction for crops and seeds, as well as in nitrogen measurement in soil content. Climate Foundation shows innovation in machine learning strategies for geospatial and weather data processing. It holds seven international patent families related to this technology, ranking thirteenth.

Emerging technologies: data collection, processing, and controlling

Technologies for data collection, processing, and controlling systems are driving the adoption of Predictive models in precision agriculture

By analyzing the IPC subclasses within the field of Predictive models in precision agriculture, statistics were compiled on the number of international patent families associated with each IPC subclass, along with the CAGR of these subclasses from 2017 to 2021. Figure 6.4 illustrates the most-utilized technologies in the area of predictive models and their applications in agriculture.

Through the analysis of international patent families related to Predictive models in precision agriculture, the technologies involved in predictive models can be categorized into three main areas: data collection, data processing and controlling systems.

• Data collection forms the foundation of Predictive models in precision agriculture. By utilizing sensor data, accelerometers, captured images, cameras and communication interfaces, a variety of real-time information about the farmland can be obtained. This data is used to monitor crop growth conditions, soil moisture, temperature and other environmental parameters, thereby providing the basis for subsequent data processing and decision-making. The relevant IPC classifications for this technology include: G06T, G06V, G01N, G01S, G01C and H04N.

• In the data processing stage, predictive models analyze the collected data through machine learning, data classification and correlation techniques. These models can identify patterns and trends within the data, assess crop health, predict pest infestations and optimize the growth environment. The key to this stage is transforming complex raw data into valuable information that helps farmers and agricultural enterprises make informed decisions. The IPC classifications relevant to data processing include G06Q, G06N, G06F and G06K.

• Controlling systems are employed in precision agriculture to manage and regulate various agricultural processes. These systems can control non-electric variables and other critical parameters to optimize the operation of agricultural equipment and processes. For example, controlling systems can automatically adjust irrigation systems, fertilization equipment and pesticide sprayers, thereby enhancing the efficiency and effectiveness of agricultural production. The IPC classifications pertaining to controlling systems include G05D and G05B.

By integrating advanced predictive model technologies at every stage – data collection, data processing and controlling systems — agricultural enterprises can streamline operations, optimize decision-making processes and ultimately achieve better production outcomes. The application of these technologies not only enhances the precision and efficiency of agriculture but also promotes the shift towards intelligent and data-driven agricultural practices.

In the field of Predictive models in precision agriculture, the majority of patent applications are concentrated on data processing technologies, followed by data collection.

Applications of predictive models within the precision agriculture sub-domain was achieved by counting international patent families in-related IPC subclasses (Figure 6.5).

In recent years, there has been a significant increase in the use of predictive models to enhance soil management, plant culture and animal husbandry practices worldwide. This growth is evidenced by a CAGR of over 20% from 2017 to 2021. Innovations in the application of predictive models have revolutionized various sectors including soil working, horticulture, harvesting, planting, animal husbandry, aviculture, apiculture, pisciculture, fishing and animal trapping.

The integration of predictive models in soil management, includes soil working (A01B), harvesting and mowing (A01D) as well as planting, sowing, fertilizing (A01C). This has led to more efficient and sustainable agricultural practices. By analyzing data on soil composition, moisture levels and nutrient content, farmers can make informed decisions on irrigation, fertilization and crop rotation. This not only improves crop yields but also helps preserve soil health for future generations.

In horticulture (A01G), predictive models are being used to optimize plant growth and development. By monitoring factors such as temperature, humidity and sunlight exposure, growers can create ideal conditions for plant growth and minimize the risk of pests and diseases. This has led to higher-quality produce and increased profitability for farmers.

The use of predictive models in animal husbandry has also transformed the way livestock is managed. Animal husbandry, aviculture, apiculture, pisciculture and fishing have A01K as an IPC subclass, and catching, trapping or scaring of animals is A01M. By analyzing data on animal behavior, health and nutrition, farmers can identify potential issues before they escalate and take proactive measures to ensure the well-being of their animals. This has not only improved animal welfare but also increased productivity and profitability for farmers.

However, if predictive models are found in major areas of AgriTech segments, it fails to deeply penetrate and support FoodTech applications, for now (including food supply chain, food services, food delivery).