Intelligent Agriculture Research Articles

Few-shot learning in intelligent agriculture: A review of methods and applications

Due to the high cost of data acquisition in many specific fields, such as intelligent agriculture, the available data is insufficient for the typical deep learning paradigm to show its superior performance. As an important complement to deep learning, few-shot learning focuses on pattern recognition tasks under the constraint of limited data, which can be used to solve practical problems in many application fields with data scarcity. This survey summarizes the research status, main models and representative achievements of few-shot learning from four aspects: model fine-tuning, meta-learning, metric learning and data enhancement, and especially introduces the few-shot learning-driven typical applications in intelligent agriculture. Finally, the current challenges of few-shot learning and its development trends in intelligent agriculture are prospected.

Intelligent Farming based on Uncertainty Expert System with Butterfly Optimization Algorithm for Crop Recommendation

Meeting the current population's food demands has become challenging, given the rising population, frequent climate fluctuations, and limited resources. Smart farming, also known as precision agriculture, has emerged as an advanced approach to tackle modern challenges in crop production. At the heart of this cutting-edge technology is machine learning, serving as the driving force behind its implementation. Though, there are many algorithms are available in crop prediction process, the problem of predicting vague information is still a challenging issue. Unfortunately, existing algorithms mostly avoids the complicated instances in crop recommendation dataset by not handling them effectively, due to imbalance class distribution. Hence in this research work to conduct an intelligent farming, two different uncertain theories are adopted to handle the issue of vagueness in appropriate recommendation of crop by considering soil fertility and climatic condition. The proposed is developed based on uncertainty expert system with both neutrosophicalparaconsistent inference model. The neutrosophic inference model is integrated with the paraconsistent logic to overcome the problem of uncertainty in prediction of appropriate crop by representing the factors in terms of certainty degree and contradiction degree. The rule generated by paraconsistent model is validated to improve the accuracy of crop prediction by fusing the knowledge of butterfly optimization algorithm. The nectar searching behavior of the butterflies are used for searching potential rules as a validation process. With the pruned rules generated by uncertainty expert model, the suitable crop is predicted more accurately compared to the other existing prediction models.

Glimpse of Artificial Intelligence in the Field of Agriculture

Artificial intelligence in agriculture is a developing topic with enormous potential to transform the way we produce and manage food. With the world's population growing and the need for more sustainable and effective farming practises, AI technologies have the potential to improve crop yields, reduce waste, and optimise resource allocation. The chapter will argue that incorporating AI into agriculture can lead to major improvements in output, sustainability, and security of food. The review paper intends to illustrate the relevance of embracing this transformational approach by evaluating the current issues encountered by the agriculture business and the anticipated benefits of AI technologies. As we progress through the chapter, we will look at specific uses of AI in agriculture and examine its significance for the future of farming.

Digital Technology Empowers the High-quality Development of Agriculture: Taking Intelligent Agriculture in Dasheng Town, Yubei District, Chongqing as an Example

Taking the smart agriculture in Dasheng Town, Yubei District, Chongqing as an example, this study explores the enabling role of digital technology in the high-quality development of agriculture. By reviewing existing research and combining field investigations, this paper summarizes the positive impacts of digital technology on improving agricultural production efficiency, enhancing the quality of agricultural products, and promoting agricultural sustainability, as well as the successful experience of the Dasheng Town smart agriculture project, providing experience and inspiration for other regions to use digital technology to promote agricultural modernization better. Through these above, the goal of high-quality agricultural development is achieved. Concurrently, in response to the shortcomings identified during the project's implementation, this paper proposes feasible suggestions from multiple aspects, including strengthening talent cultivation and introduction in the field of smart agriculture, enhancing policy support and incentive mechanisms, and promoting collaboration and scaled production. These suggestions serve as references for how digital technology can empower high-quality agricultural development in agriculture and the development of other smart agriculture projects under the context of rural revitalization.

A new method of crop yield prediction based on artificial intelligence

With the rapid growth of population, the continuous reduction of arable land and the acceleration of urbanization, the challenges facing agriculture are becoming increasingly severe. The rapid development of the new generation of digital economy, such as the industrial Internet, cloud computing and ChatGPT, has brought all-round impact on people’s lives and the global economy. At the same time, the changes of social economy and technology have promoted the development of agricultural production, and the application of intelligent stage number to create smart agriculture can improve the land output rate, labor productivity and resource utilization rate, and achieve sustainable agricultural development. Agricultural intelligence will become an important development direction of agriculture in the future. In this paper, an artificial intelligence-based method of crop yield prediction was proposed to improve the prediction accuracy of crop yield by using smart agriculture.

Frequency Offset Methods for Large-Scale Deployment of LoRa: A New Strategy to Reduce Data Retransmission

LoRa technology, renowned for its extensive range and robust anti-interference capabilities, has emerged as a linchpin for myriad IoT applications, ranging from agricultural intelligence to smart metering. Yet, its potential for large-scale deployment remains stymied by significant packet conflict issues at the gateway, particularly pronounced in extensive networks involving thousands of endpoints. Such conflicts, exacerbated during network reboots, instigate a domino effect of retransmissions, power consumption, and channel resource wastage, culminating in severely impeded communication efficiency. A pivotal challenge arises from the concurrent data uploads by numerous terminals, coupled with the restrictive uplink channel choices due to limited carrier frequencies and spreading factors. While existing solutions like phase, power, and inherent frequency offsets offer avenues for distinguishing between packet conflicts, they are riddled with practical constraints. This paper introduces a novel approach, leveraging a controlled frequency offset for LoRa symbols. Grounded in the fundamentals of LoRa demodulation, this innovative strategy promises to tackle large-scale congestion without adding system intricacy.

Application of Smart Farming Technologies in Sustainable Agriculture Development: A Comprehensive Review on Present Status and Future Advancements

The pressure on agricultural food production to produce high-quality and sustainable food has increased dramatically as a result of global urbanization, population growth and climate change. Food shortage and population expansion are two of the world's major challenges to long-term sustainable development. So, in recent years technological development such as smart farming has resulted substantial changes in the agricultural production allowing for 3rd green revolution and progression in farming methods. Smart farming (SF) is a method of agricultural production that integrates information and communication technologies to maximise resource efficiency. Smart farming technologies such as automated tractors, artificial agricultural intelligence (AAI), automated irrigation and weather forecasting, Internet of Things (IoT), kisan drones and remote sensing (GIS and GPS) can provide realistic solutions to the world's challenges. The accurate application of this technology in conjunction with the Internet of Things (IoT) is expected to be a beneficial for the farmers to improve their living standards through increased production and profit which is a good indicator for sustainable food security. However, due to lack of internet access, high costs and lack of application knowledge there are still barriers to the adaptation and conversion of smart farming. As a result, more research is needed in this area.

Three-dimension object detection and forward-looking control strategy for non-destructive grasp of thin-skinned fruits

The dynamic non-destructive grasp of thin-skinned fruits using flexible robotic hands, which requires obtaining three-dimension(3D) spatial structure information along with adaptive planning and motion control toward the target object, is a challenging topic in agricultural intelligence. To tackle the issue of 3D detection, we utilize RGB images and LiDAR point clouds for feature extraction and construct a multi-modal depth fusion convolution neural network (MDF-CNN) to obtain the classification information and perform image segmentation. Incorporating the advantages of a variable palm structure, we establish an evaluation mechanism of the optimal grasping stability (EM-OGS) using a hybrid method of the best configuration and force closure to build a new comprehensive performance optimal configuration planning (CPO-CP) method that is based on the multiple grasping performance indexes. We also create three cross-related nonlinear prediction models P-MGF, P-ODAP, and P-OBA along with a forward-looking non-destructive grasp control algorithm (FL-NGCA) using minimum grasping force to grasp thin-skinned fruits. The control algorithm carries out online, self-directed learning in the actual grasping process of the flexible hand and constantly optimizes the accuracy of the prediction model. Experimental results show that our proposed approach greatly improves the flexible hand comprehensive performance of grasping, outperforming state-of-the-art methods for non-destructive grasping of delicate fruits in most areas.

Advanced Technologies and Artificial Intelligence in Agriculture

According to the Food and Agriculture Organization, the world’s food production needs to increase by 70 percent by 2050 to feed the growing population. However, the EU agricultural workforce has declined by 35% over the last decade, and 54% of agriculture companies have cited a shortage of staff as their main challenge. These factors, among others, have led to an increased interest in advanced technologies in agriculture, such as IoT, sensors, robots, unmanned aerial vehicles (UAVs), digitalization, and artificial intelligence (AI). Artificial intelligence and machine learning have proven valuable for many agriculture tasks, including problem detection, crop health monitoring, yield prediction, price forecasting, yield mapping, pesticide, and fertilizer usage optimization. In this scoping mini review, scientific achievements regarding the main directions of agricultural technologies will be explored. Successful commercial companies, both in the Russian and international markets, that have effectively applied these technologies will be highlighted. Additionally, a concise overview of various AI approaches will be presented, and our firsthand experience in this field will be shared.

Wearable Electrochemical Biosensors for In Situ Pesticide Analysis from Crops

On-site monitoring of pesticide presence on the surface of crops is a growing trend in precision and intelligent agriculture. In this study, the researchers prepared a wearable biosensor for the in situ detection of pesticides on the surfaces of crops. Here, a flexible fiber membrane substrate was prepared via the electrospinning technology. After the three-electrode system was transferred onto the membrane substrate used for screen printing, this resulted in a screen printing electrode(SPE) that can effectively adapt to the irregular surface of crops or fruits. By modifying the SPE with acetylcholinesterase(AchE), the biosensor showed excellent selectivity and recognition for the methyl parathion. To further enhance the electrochemical performance, the surface of the work electrode was modified with gold nanoparticles (AuNPs) and reduced graphene oxide (rGo). The developed wearable sensor successfully detected methyl parathion, showing a good linear relationship within the range of 1 ppb to 2 ppm. The detection limit was measured to be 0.48 ppb, enabling on-site monitoring of pesticide levels in plants. This work presents straightforward, sensitive, and efficient biosensors to analyze pesticides, which hold potential for application in detecting other agrochemicals. Moreover, this advancement could significantly contribute to the progress of precision agriculture in the future.

Machine-Learning-Based UAV-Assisted Agricultural Information Security Architecture and Intrusion Detection

In recent years, Unmanned Aerial Vehicle (UAV) remote sensing has developed rapidly in the field of farmland information monitoring. Real-time and accurate access to farmland information and crop growth dynamics is a prerequisite for the implementation of precision agriculture. Machine Learning identifies existing knowledge to acquire new knowledge, promotes the development of Artificial Intelligence, and brings a large number of data training sets for Machine Learning. The present work aims to ensure the safe operation of agricultural information systems and guarantee the data security of intelligent agriculture. The Machine Learning method explores the wireless network deployment of the UAV system. The geographical location deployment of agricultural information security can effectively carry out rapid security detection of agricultural information security. Firstly, the UAV-assisted information acquisition system was studied. Besides, a Double Deep Q-network (DDQN) algorithm for location deployment based on Geography Position Information (GPI) was proposed to quickly optimize the deployment location of UAVs. GPI can avoid the complicated calculation process of channel state information. The DDQN algorithm was introduced to obtain the functional relationship between GPI and the optimal UAV deployment position, forming a new GPI-Learning strategy. In addition, the Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) are integrated as the CNN-LSTM algorithm to build the intrusion detection system for agricultural IoT for agricultural. In the integrated network structure of the system, LSTM is responsible for data transmission, and CNN is capable of network model building. Combined with the influence of various parameters on the performance of the UAV deployment location algorithm, the simulation experiment set the population size as 36, the discovery probability as 0.25, the step scaling factor as 0.8, and the Levy flight index as 1.25. The network throughput performance of the GPI-Learning algorithm combined with Cuckoo Search was better than other algorithms under different numbers of UAVs. On the KDD-CUP99 data set, the accuracy and detection rate of the agricultural IoT intrusion detection system based on the CNN+LSTM algorithm reached 93.5% and 94.4%, respectively. In general, the agricultural IoT intrusion detection system reported here has crucial practical reference value for the safe operation of agricultural information systems.

Implementasi Profil Pelajar Pancasila melalui Kegiatan Ekstrakurikuler Robotik dan Smart Farmer di SMAS Santo Tarcisius Dumai

Santo Tarcisius High School is the first driving school in Dumai City, so Santo Tarcisius High School has the responsibility to implement the Pancasila Student Profile for its students. One of the characteristics of the Pancasila Student Profile is creative, namely being able to modify and produce something original, meaningful, useful and impactful. To realize the Pancasila Student Profile, an effort is needed to provide knowledge and skills to students in an interesting way, one of which is through robotics and smart farmer extracurricular learning activities. Based on the results of discussions with the school principal, the problems faced in implementing robotics and smart farming extracurricular activities include 1) Students have never had an understanding of computer and robot-based intelligent farming system technology, 2) They do not yet have teachers who are competent in the field of robotics, 3) Limited costs purchasing robotic electronic equipment and installing smart farmers at schools. Based on the problems above, the aim of this activity is to realize the Pancasila Profile, especially in the IT sector, through education on making Line Follower robots and smart farming systems for hydroponic plants. The implementation method is by dividing students into several groups to make Line Follower robots. The number of students in this activity is 150 students/class XI SMA Santo Tarcisius. Success in Beginner Community Service (PMP) is to improve students' skills in making robots and intelligent systems which is measured through the final product results, pre test and post test questionnaires, the product results from this PMP are line follower robots and smart farmer systems, the conclusion of the PMP This is to increase efficiency, productivity and sustainability in the agricultural sector through the application of the latest technology.

Smart greenhouse construction and irrigation control system for optimal Brassica Juncea development.

This paper contributes to smart greenhouses and IoT (Internet of Things) research. Our pioneering achievement centers on successfully designing, constructing, and testing a 30m2 smart greenhouse, explicitly focusing on the cultivation and development of Brassica Juncea, a mustard variety commonly grown in Vietnam. The construction phase entailed the meticulous integration of diverse IoT technologies and systems, culminating in the creation of a finely tuned environment to meet the unique needs of Brassica Juncea cultivation. Notably, our research team has realized the physical infrastructure and developed and implemented a robust web interface. This interface empowers users to monitor and remotely control the smart greenhouse conveniently. It provides real-time visualization of critical parameters, including temperature, humidity, soil moisture, and light intensity, enabling precise monitoring and supporting informed decision-making in crop management. In addition to the web interface, we have meticulously designed and completed an Android mobile application, further enhancing accessibility and convenience. This mobile app allows users to monitor and control the smart greenhouse while on the move. It is imperative to underscore that this work marks a significant milestone as the first complete smart greenhouse IoT solution dedicated to developing Brassica Juncea. Our pioneering accomplishments not only advance the frontiers of innovative greenhouse and IoT research but also contribute substantially to the progress of intelligent agriculture.

An iterative noisy annotation correction model for robust plant disease detection.

Previous work on plant disease detection demonstrated that object detectors generally suffer from degraded training data, and annotations with noise may cause the training task to fail. Well-annotated datasets are therefore crucial to build a robust detector. However, a good label set generally requires much expert knowledge and meticulous work, which is expensive and time-consuming. This paper aims to learn robust feature representations with inaccurate bounding boxes, thereby reducing the model requirements for annotation quality. Specifically, we analyze the distribution of noisy annotations in the real world. A teacher-student learning paradigm is proposed to correct inaccurate bounding boxes. The teacher model is used to rectify the degraded bounding boxes, and the student model extracts more robust feature representations from the corrected bounding boxes. Furthermore, the method can be easily generalized to semi-supervised learning paradigms and auto-labeling techniques. Experimental results show that applying our method to the Faster-RCNN detector achieves a 26% performance improvement on the noisy dataset. Besides, our method achieves approximately 75% of the performance of a fully supervised object detector when 1% of the labels are available. Overall, this work provides a robust solution to real-world location noise. It alleviates the challenges posed by noisy data to precision agriculture, optimizes data labeling technology, and encourages practitioners to further investigate plant disease detection and intelligent agriculture at a lower cost. The code will be released at https://github.com/JiuqingDong/TS_OAMIL-for-Plant-disease-detection.

A lightweight rice disease identification network based on attention mechanism and dynamic convolution

Ensuring the quality and yield of rice depends heavily on the accurate identification of early-stage rice diseases. Existing models face significant challenges in balancing lightweight requirements and precise classification of rice disease types due to the noisy background and scattered distribution of disease symptoms in real-world environments. To address the above issues, this study proposes DGLNet, a novel lightweight and highly accurate network for rice disease identification. DGLNet includes two low-complexity modules, the global attention module (GAM) and the dynamic representation module (DRM). The GAM is designed to capture key information in complex noisy scenes, thus improving the generalization ability of the model. Meanwhile, the self-developed four-dimensional flexible convolution (4D-FConv) in the DRM can dynamically generate adaptive convolutional kernel parameters from four dimensions. This allows DRM to maintain diversity among different sample inputs to enhance the model's ability to fit complex functions. As a result, DRM enhances feature representation without the need for additional network layers and channels. The proposed method achieves 99.82% and 99.71% recognition accuracy on two real plant disease datasets, outperforming current popular methods. The experimental results demonstrate that the proposed method is not only computationally lightweight, but also capable of accurately identifying rice diseases in real-life scenarios. Furthermore, this study provides robust technical support for disease identification and control and offers guidance for the implementation of agricultural intelligence and precision farming.

Artificial intelligence and new business models in agriculture: the “ZERO” case study

PurposeThis study aims to investigate the use of Artificial Intelligence (AI) applied to vertical farms to evaluate whether disrupting technology supports sustainability and increases strategic business model choices in the agricultural sector. The study responds through empirical analysis to the gap on the subject of AI-driven business models present in the growing sector literature.Design/methodology/approachThe paper analyzes the case of “ZERO”, a company linked to the strategy innovation ecosystem of the Ca’ Foscari University of Venice, Italy. The empirical data were collected through a semi-structured questionnaire, interviews and the analysis of public news on the business model available in the analyzed case study. The research is empirical and uses exploratory, descriptive analysis to interpret the findings. The article focuses on the evaluation of AI impact on the agricultural sector and its potential to create new business models.FindingsThe study identified how AI can support the decision-making process leading to an increase in productivity, efficiency, product quality and cost reduction. AI helps increase these parameters through a continuous learning process and local production, and the possible decrease in prices directed toward the goal of zero km food with fresh products. AI is a winning technology to support the key elements of the vertical farm business model. However, it must be coupled with other devices, such as robots, sensors and drones, to collect enough data to enable continuous learning and improvement.Research limitations/implicationsThe research supports new research trends in AI applied to agriculture. The major implication is the construction of ecosystems between farms, technology providers, policymakers, universities, research centers and local consumer communities.Practical implicationsThe ZERO case study underlines the potential of AI as a destructive technology that, especially in vertical farms, eliminates external conditions by increasing productivity, reducing costs and responding to production needs with adequate consumption of raw materials, boosting both environmental and social sustainability.Originality/valueThe study is original, as the current literature presents few empirical case studies on AI-supporting business models in agriculture. The study also favors valuable strategic implications for the policies to be adopted in favor of new business models in agriculture.

A New Semantic Segmentation Framework Based on UNet.

This paper discusses a semantic segmentation framework and shows its application in agricultural intelligence, such as providing environmental awareness for agricultural robots to work autonomously and efficiently. We propose an ensemble framework based on the bagging strategy and the UNet network, using RGB and HSV color spaces. We evaluated the framework on our self-built dataset (Maize) and a public dataset (Sugar Beets). Then, we compared it with UNet-based methods (single RGB and single HSV), DeepLab V3+, and SegNet. Experimental results show that our ensemble framework can synthesize the advantages of each color space and obtain the best IoUs (0.8276 and 0.6972) on the datasets (Maize and Sugar Beets), respectively. In addition, including our framework, the UNet-based methods have faster speed and a smaller parameter space than DeepLab V3+ and SegNet, which are more suitable for deployment in resource-constrained environments such as mobile robots.

Water Fertilizer Integration Equipment Research Status and Development Trend

Integrated water and fertilizer technology can accurately adjust the mix of water and fertilizer on demand, and then through the irrigation system, water and fertilizer are delivered to the roots of the crop in quantity, on time, and with precision. As an important part of the integrated system of water and fertilizer, the fertilizer has a critical influence on the quality of the fertilizing system. Nowadays, the development of intelligent agriculture has become a general trend of agricultural development. By exploring the development process of water and fertilizer mixing, this paper sorted out the research status of various kinds of fertilizers in China, compared different mixing methods of water and fertilizer, and analyzed and summarized the excellent working performance of different fertilizers. Secondly, the current research emphases were discussed, including precision fertilization technology, intelligent control, comprehensive utilization of resources and circular agriculture. Then, the development trend of integrated water and fertilizer equipment is put forward, including the improvement of precision fertilization, the development of multifunctional integrated equipment, and the application of data management and decision support system. Through continuous technological innovation and application, the integrated equipment of water and fertilizer is expected to play a greater role in the agricultural field.

Artificial intelligence in postharvest agriculture: mapping a research agenda

IntroductionThe implementation of artificial intelligence (AI) in postharvest agriculture has significantly improved in recent decades, thanks to extensive scientific research. The study aimed to identify research gaps and hotspots for future research based on keyword co-occurrence and clustering analyses, as well as to discuss the results and highlight the research trends.MethodsThis study analyses research trends in AI application in postharvest agriculture using novel scientometric tools such as the Bibliometrix R package, biblioshiny, and VosViewer. The research analysed 586 published papers on AI application in postharvest agriculture research between 1994 and September 2022, retrieved from the Scopus database.Results and discussionThe results showed that publications on AI applications in postharvest agriculture research have been increasing for almost 30 years, with significant growth in the subject area in the last decade. China, the USA, and India were found to be the top three most productive countries, accounting for 52.4%, 22%, and 18.6% of the total selected publications, respectively. The analysis also revealed that topics such as the Internet of Things, cold chain logistics, big data, decision-making, and real-time monitoring have low development degrees in the knowledge domain. This study demonstrated increased research on AI applications in postharvest agriculture, aiming to reduce postharvest losses, enhance food nutrition and quality, and mitigate food insecurity. It also provides valuable scientific references on AI applications in postharvest agriculture research for researchers and scholars. By identifying research gaps and hotspots, this study can guide future research in AI applications in postharvest agriculture to further improve the industry.

Improved MobileNetV2 crop disease identification model for intelligent agriculture.

Using intelligent agriculture is an important way for the industry to achieve high-quality development. To improve the accuracy of the identification of crop diseases under conditions of limited computing resources, such as in mobile and edge computing, we propose an improved lightweight MobileNetV2 crop disease identification model. In this study, MobileNetV2 is used as the backbone network for the application of an improved Bottleneck structure. First, the number of operation channels is reduced using point-by-point convolution, the number of parameters of the model is reduced, and the re-parameterized multilayer perceptron (RepMLP) module is introduced; the latter can capture long-distance dependencies between features and obtain local a priori information to enhance the global perception of the model. Second, the efficient channel-attention mechanism is added to adjust the image-feature channel weights so as to improve the recognition accuracy of the model, and the Hardswish activation function is introduced instead of the ReLU6 activation function to further improve performance. The final experimental results show that the improved MobilNetV2 model achieves 99.53% accuracy in the PlantVillage crop disease dataset, which is 0.3% higher than the original model, and the number of covariates is only 0.9M, which is 59% less than the original model. Also, the inference speed is improved by 8.5% over the original model. The crop disease identification method proposed in this article provides a reference for deployment and application on edge and mobile devices.