NEC is currently developing “CropScope”, an agricultural digital transformation (DX) solution for open-field farming. With CropScope, growers can remotely monitor field conditions in real time using data collected from satellites and various sensors. Additionally, CropScope can automatically determine the optimal amounts and schedules for water and fertilizer application based on the collected data, facilitating seamless execution of these actions. In field trials conducted in processing tomato fields across Europe, CropScope successfully reduced water usage by 15% and fertilizer usage by 20%, while simultaneously increasing yields by 20%. This paper introduces the features of CropScope that make these achievements possible and discusses its potential role in minimizing environmental impact.

1. Introduction

With a growing global awareness of sustainability challenges, there is a widespread effort in agriculture to reduce environmental impact while maintaining productivity. Examples of environmental burdens from agriculture include greenhouse gas emissions from tractor operations and soil biochemical processes, as well as water pollution from fertilizer runoff into bodies of water such as oceans, rivers, and lakes. To address these challenges and maximize yields, there is a need for solutions that can optimize agricultural activities such as tillage, planting, weeding, pesticide application, irrigation, fertilization, and harvesting.

NEC has developed “CropScope”, an agricultural DX solution aimed at open-field farming to realize “low-input, high-output” agriculture. CropScope aggregates, visualizes, and analyzes a wide range of farming data, including crop growth, weather conditions, soil moisture, disease risk, and farm operation records (Fig. 1). In addition, it remotely controls irrigation and fertilization equipment through integration with irrigation and fertilization systems (Fig. 2). CropScope is deployed in 14 countries and adapted for 14 different types of open-field crops. A key feature is its capability to automatically optimize irrigation and fertilization levels using sensing data (Fig. 3), which plays a vital role in achieving low-input, high-output agriculture.

This paper is organized as follows: In section 2, we explain the techniques for learning irrigation and fertilization practices from expert growers using an integrated crop model and machine learning. Section 3 introduces the remote automatic control feature of irrigation systems, which can potentially enhance crop yields. Following this, section 4 discusses how CropScope contributes to reducing the environmental impact of agriculture. Finally, we conclude the paper with a summary.

2. Reproducing Expert Farming Practices Using an Integrated Crop Model and Machine Learning

Fig. 4 illustrates the process for deriving optimal irrigation and fertilization amounts within CropScope. By utilizing this feature, it is possible to accurately determine the precise amounts of water and fertilizer that crops truly need while maintaining crop yields.

As a first step in establishing optimal farming practices, CropScope collects crop growth data through satellite imagery and gathers weather and soil moisture information using IoT sensors installed in the field.

This data is then fed into an integrated crop model to simulate crop growth. The integrated crop model is composed of several modules: a soil module that describes the movement of water and nutrients in the soil, a weather module that outlines the impact of weather conditions on soils and crops, and a crop module that details how crops grow under specific soil and weather conditions. By performing simulations using this integrated model, CropScope can estimate field parameters that cannot be directly obtained from satellites and sensors, such as evapotranspiration from leaves and nutrient requirements of crops.

Finally, the collected data, estimated parameters, and irrigation and fertilization history data from expert growers are used to model their farming practices through machine learning.

By utilizing the integrated crop growth model as a comprehensive repository of agricultural knowledge, we have successfully transformed raw sensing data into agronomically meaningful features. This approach has enabled us to accurately learn the farming practices of expert growers, even with a small amount of training data (approximately 10 fields) as shown in Fig. 5.

The expert growers involved in the study typically use less fertilizer while achieving yields similar to those of their peers, indicating that adopting their modeled practices could lead to low-impact farming. In a 2019 field trial at a processing tomato field in Portugal, the model’s recommended daily irrigation and fertilization schedules were implemented to test their efficacy. The results demonstrated that the irrigation, fertilization, and yield achieved were comparable to those of the expert growers.¹⁾ Notably, compared to regional averages, fertilizer usage was reduced by 20%, and yield increased by 30%, exemplifying low-input, high-output agriculture.

Typically, excessive fertilizer application does not significantly harm crop growth, whereas inadequate fertilization can severely inhibit growth. Consequently, many growers habitually apply more fertilizer than necessary. However, as demonstrated in the field trials, by referencing the recommended values provided by CropScope's advanced irrigation and fertilization guidance feature, anyone is　expected to be able to execute timely and appropriate irrigation and fertilization practices.

3. Automated Remote Control for Enhanced Water Conservation and Crop Yield

By utilizing the techniques for determining optimal irrigation and fertilization amounts discussed in section 2, we demonstrated that it is possible to achieve yields equal to or greater than regional averages while reducing fertilization by 20%. Currently, CropScope includes a feature for automating “Small-amount and high-frequency” irrigation, allowing anyone to easily adopt farming practices that not only reduce water usage but also have the potential to further increase yields.

Small-amount and high-frequency irrigation is a method that divides the necessary irrigation amount for a given day into multiple applications, maintaining optimal soil moisture levels for crops at all times (Fig. 6). However, for growers managing large numbers of extensive fields, the complexity of irrigation management can increase the workload significantly, hindering widespread adoption.

By utilizing CropScope, schedules for Small-amount and high-frequency irrigation and fertilization can be easily created automatically for each day. This capability is especially advantageous for large-scale fields consisting of multiple plots, as it enables the development of an optimal schedule that considers the varying irrigation and fertilization needs of each plot (Fig. 7). The automation of Small-amount and high-frequency irrigation allows for remarkably precise irrigation management, moving us closer to realizing low-input, high-output agriculture.

To verify the effects of Small-amount and high-frequency irrigation on reducing irrigation amounts and increasing yield, we have been conducting ongoing field trials in processing tomato fields across various countries, including Portugal, Spain, Italy, Turkey, the United States, and Chile, since 2022. Although some cases do not yield the expected results due to external factors, the trials generally indicate a potential reduction of irrigation volume by about 15% and an increase in yield by 20%.^{2) 3)} Factors that can inhibit these effects include inadequate soil preparation before planting and the occurrence of pests and diseases. To minimize these issues and more reliably achieve the low-input, high-output effect through the use of CropScope, we have partnered with Kagome Co., Ltd., a company rich in agricultural knowledge and experience, to establish the joint venture DXAS Agricultural Technology LDA. Moving forward, we aim to expand CropScope's business by integrating the agronomy and technology expertise of both companies.

4. Contribution of CropScope to Reducing Environmental Impact

By utilizing CropScope's advanced irrigation and fertilization guidance feature, alongside its Small-amount and high-frequency irrigation automation, growers can achieve low-input, high-output agriculture. This approach not only boosts profitability for growers but also plays a significant role in climate change mitigation and minimizing environmental impact. Expected benefits include reduced water usage, addressing global drought concerns, and preventing fertilizer over-application, which can lead to water body eutrophication and increased greenhouse gas emissions.

Nitrous oxide (N₂O), a greenhouse gas emitted from nitrogen fertilizers used on farmland, is 265 times more potent than carbon dioxide (CO₂).⁴⁾ N₂O and CO₂ emissions from these fertilizers contribute approximately 1.23% of the world’s greenhouse gas emissions. When accounting for the entire supply chain, including the production and transportation of nitrogen fertilizers, this figure rises to 2.1%, totaling 1.13 Gt in CO₂ equivalents.⁵⁾ As highlighted in section 2, the application of CropScope in processing tomato fields has led to instances where nitrogen fertilizer usage was reduced by 20% compared to traditional methods. By expanding its reach globally and targeting a variety of crops in the future, CropScope aims to significantly reduce greenhouse gas emissions from agriculture.

5. Conclusion

In this paper, we introduced CropScope, an application designed to optimize irrigation and fertilization by leveraging field sensing and data analysis. In recent years, cutting-edge technologies such as IoT and AI have started to make inroads into agriculture. However, for growers who have traditionally relied on their senses—like touching soil and crops, and even smelling or tasting them—to assess field conditions, translating digital "mere numbers" into practical insights can be challenging. NEC envisions CropScope as a "grower-friendly" tool that bridges this gap, enabling more precise decision-making through digital information. By expanding CropScope's reach globally, we aim to significantly contribute to achieving a sustainable society.

References

Authors’ Profiles

Related URL: