• Research Article

    Cucumber fruit weight prediction using deep learning models on Korea's smart agriculture big data platform
    Im-joung Choi, Shahriar Ahmed, Sun-Ok Chung, Myongkyoon Yang
    The integration of artificial intelligence (AI) in smart agriculture has enabled precise crop monitoring and yield prediction, contributing to improved resource management … + READ MORE
    The integration of artificial intelligence (AI) in smart agriculture has enabled precise crop monitoring and yield prediction, contributing to improved resource management and productivity. However, integrating multiple data types to build an AI model for crop monitoring and yield could be challenging. This study develops and evaluates deep learning models for predicting cucumber fruit weight by utilizing structured time-series data from a smart farm database. The dataset, collected from a protected horticulture environment, incorporates key growth parameters (leaf length, leaf count, and leaf width) and environmental factors (internal temperature, relative humidity, and solar radiation). Three deep learning models—Gated Recurrent Unit (GRU), One-Dimensional Convolutional Neural Network (1D-CNN), and Transformer—were implemented to analyze the relationships between these factors and fruit weight. Several preprocessing steps, including anomaly handling and data purification, were applied to improve data quality and model performance. Complex variables reflecting the interaction between growth and environmental factors were generated, and ideal values were adjusted to minimize distortion without data loss. The models were trained to predict cucumber harvest overload using 10-fold cross-validation to avoid overfitting. The results indicate that all models demonstrated strong predictive capabilities, with the Transformer model achieving the highest accuracy (MAE = 0.1584, R2 = 0.9999), followed by GRU (MAE = 3.5735, R2 = 0.9728) and 1D-CNN (MAE = 3.8413, R2 = 0.9733). The superior performance of the Transformer model is attributed to its ability to capture complex dependencies within time-series data. GRU effectively modeled sequential dependencies, while 1D-CNN efficiently extracted local features from temporal sequences. This study highlights the potential of deep learning models in smart agriculture for growth monitoring and precision farming. However, the research is limited by the dataset’s geographic scope and the exclusion of additional agronomic variables such as soil conditions and nutrient levels. Future research should explore model optimization, diverse environmental datasets, and real-time deployment in smart farming systems. - COLLAPSE
    31 March 2025
  • Research Article

    Stability evaluation of walking-type rotary cultivators through sensor data analysis
    Hye-In Kim, Kyu-ho Lee, Sun-Ok Chung
    To ensure the operational stability of walking-type rotary cultivators is one of the essential things to improve safety and efficiency in small-scale … + READ MORE
    To ensure the operational stability of walking-type rotary cultivators is one of the essential things to improve safety and efficiency in small-scale agricultural environments, particularly in regions with uneven and rocky terrains such as South Korea. This study aims to quantitively assess the stability and safety risks associated with rotary cultivators by sensor data collected during field operations. A smartphone-based sensing system integrating accelerometers and gyroscopes was used to monitor dynamic response under varying operational conditions. The experiments were conducted in both standard and gravelly soil environments, evaluating the effects of different rotary configuration, such as front-mounted vs. rear-mounted, and working depths on machine stability. The results indicated that rear-mounted rotary configurations showed superior stability, as they reduced impact forces and minimized abrupt vibrations compared to front-mounted setups, which were prone to excessive tilting and instability. Additionally, deeper working depth in gravelly fields significantly amplified hazardous situations, including abrupt stops and overturns, because of greater resistance and interaction with embedded rocks. Sensor data analysis confirmed that higher acceleration spikes and angular velocity deviations correlated strongly with critical instability scenarios, emphasizing the necessity for adaptive control mechanisms and optimized rotary positioning. The findings emphasize the necessity of structural improvements in rotary cultivator designs and the development of real-time stability monitoring systems to enhance safety in mechanized farming. - COLLAPSE
    31 March 2025
  • Review Article

    IoT-enabled LoRaWAN gateway for monitoring and predicting spatial environmental parameters in smart greenhouses: A review
    Emmanuel Bicamumakuba, Eliezel Habineza, Samsuzzaman, Md Nasim Reza, Sun-Ok Chung
    The integration of the Internet of Things (IoT) with Long Range Wide Area Network (LoRaWAN) technology has revolutionized precision agriculture, particularly in … + READ MORE
    The integration of the Internet of Things (IoT) with Long Range Wide Area Network (LoRaWAN) technology has revolutionized precision agriculture, particularly in smart greenhouse environments. This review explores the role of IoT-enabled LoRaWAN gateways in monitoring and predicting spatial environmental parameters, which are crucial for optimizing crop growth, reducing resource consumption, and enhancing productivity. LoRaWAN ensures seamless communication between data management platforms and sensor networks by minimizing packet loss and enhancing network reliability through the use of adaptive data rate (ADR) and Just-In-Time (JIT) scheduling. An IoT-based automated irrigation system utilizing LoRaWAN communication demonstrated up to a 34% improvement in water use efficiency by enabling precise soil moisture monitoring and data-driven irrigation scheduling. Blockchain-based systems and AES-128 encryption strengthen security by guaranteeing distributed access and thereby guarding against cyberattacks. Key challenges such as network scalability, data security, interoperability, and energy efficiency are also analyzed. By synthesizing recent advancements and emerging trends, this review highlights the potential of IoT-enabled LoRaWAN gateways in transforming greenhouse agriculture and provides insights into future research directions. - COLLAPSE
    31 March 2025
  • Research Article

    Development of a soil environment monitoring device using wireless sensor nodes for fruit tree smart farming
    YeongGil Kim, DongHyun Kang, Cheolwoo Han
    Fruit cultivation in domestic orchards requires significant labor for managing growth and harvesting. However, due to an aging farming population and rising … + READ MORE
    Fruit cultivation in domestic orchards requires significant labor for managing growth and harvesting. However, due to an aging farming population and rising labor costs, the agricultural workforce is steadily decreasing, posing annual difficulties for farmers. It is essential to continuously monitor the crop growth environment in orchards; however, conventional methods often result in excessive fertilization, nutritional deficiencies, and increased management costs due to indiscriminate pesticide spraying. To address these challenges, this study aimed to develop a soil environment monitoring system to enhance efficient orchard management and productivity. Sensor modules capable of measuring soil temperature, humidity, electrical conductivity (EC), and pH were selected and connected to sensor nodes using RS485 Modbus communication. Data collected from sensor nodes were transmitted to local control units using LoRa wireless communication technology. Solar panels were installed for sustainable power supply, and data measurements and transmissions were scheduled at 1 hour intervals to minimize power consumption. A soil map was generated using the collected data. The results demonstrated the feasibility of automating soil environment monitoring and labor reduction in open field conditions, providing foundational data for implementing fruit tree smart farms in orchard management. - COLLAPSE
    31 March 2025
  • Research Article

    Estimation of napa cabbage fresh weight using uav-based multispectral images and accumulated temperature
    Chang-Hyeok Park, Chan-Seok Ryu, Ye-Seong Kang, Gang-In Je, Ho-Jun Kwon
    This study aimed to develop a regression model to accurately estimate napa cabbage fresh weight using UAV-based multispectral imagery, incorporating accumulated temperature … + READ MORE
    This study aimed to develop a regression model to accurately estimate napa cabbage fresh weight using UAV-based multispectral imagery, incorporating accumulated temperature (AT) to improve prediction accuracy under varying environmental conditions. Growth data and multispectral images were collected for two cultivars, Cheongmyeonggael and Bulam No.3, during the 2022 and 2023 growing seasons, and ten vegetation indices (VIs) were calculated. Both linear regression models (Multiple Linear Regression, Ridge, Lasso) and nonlinear models (Support Vector Regression, K-Nearest Neighbors) were applied, and their performance was evaluated using K-Fold Cross Validation. As a result, Ridge Regression showed the highest prediction accuracy in cultivar-specific models, while Multiple Linear Regression performed best in the integrated model. NDRE and TCARI were the most influential variables selected in the Ridge Regression models of Cheongmyeonggael and Bulam No.3, respectively. Furthermore, the inclusion of accumulated temperature significantly improved model performance, confirming its potential to reflect environmental growth conditions. This study presents the potential of integrating remote sensing imagery with climate data to enhance crop biomass estimation and suggests the feasibility of applying this precision agriculture-based yield prediction model under diverse environmental conditions. - COLLAPSE
    31 March 2025
  • Research Article

    Analysis of fluidization characteristics of grains in a vertical tube for the improvement of separation performance of combine harvester
    Choung-Keun Lee, Nam-Kyu Yun, Duck-Kyu Choi
    This study was performed to investigate the separation characteristics of grains by fluidization. The rapeseeds, rough rice and soybeans were fluidized by … + READ MORE
    This study was performed to investigate the separation characteristics of grains by fluidization. The rapeseeds, rough rice and soybeans were fluidized by air using air distributors with wire screens (18 mesh). The minimum fluidization velocities according to the changes of the grain moisture contents were 0.65~0.74 m/s in rapeseeds, and 1.30~1.63 m/s in rough rice, and 2.05~2.21 m/s in soybeans, respectively. As bed volume was changed, the minimum fluidization velocities were varied as 0.61~0.71 m/s in rapeseeds, and 1.25~1.65 m/s in rough rice, and 1.54~2.13 m/s in soybeans. The terminal velocities by change of the grain moisture contents were 5.97~6.60 m/s in rapeseeds, and 7.10~7.78 m/s in rough rice, and 14.4~14.7 m/s in soybeans, respectively. With changes of bed volume, the terminal velocity were 5.96~7.33 m/s in rapeseeds, and 7.26~8.68 m/s in rough rice, and 13.6~15.6 m/s in soybeans. The minimum fluidization and terminal velocity for the grain were highly correlated with the grain moisture content and bed volume except soybeans. - COLLAPSE
    31 March 2025
  • Research Article

    Development of a self-propelled riding garlic clove planter for film mulching -Static falling down angle, center of gravity, seed metering performance-
    Choung-Keun Lee, Duck-Kyu Choi
    Garlic planting is one of the most laborious works during garlic cultivation executed within a short period of time, especially when southern-type … + READ MORE
    Garlic planting is one of the most laborious works during garlic cultivation executed within a short period of time, especially when southern-type garlic are planted almost manually. This study was conducted to develop a self-propelled riding type garlic clove planter for the southern-type garlic. The garlic clove planter which is capable of drilling holes in the soil and planting garlic cloves at the same time was designed and tested for film mulched beds. The designed prototype planter consisted of main body, engine, travelling part, metering and planting device, metering and planting cam, metering bucket, seed hopper, and hydraulic equipment. Static turn-over angles toward left, right, front and rear side of the prototype were adjusted wider than 35°. Load distribution on the left-and right-side showed 45.9% and 54.1%, and those of the front-and rear-side were 37.3% and 62.7%, respectively. Results imply that difference of the load distribution on each wheel would not affect significantly on the straightness and driving ability of the prototype. When the planter was tested at variable working speed, the best performance was achieved with speed at 0.22 m/s with single clove planting rate, double clove metering, miss-planted rate and fall rate 62.5%, 24.0%, 13.5%, and 24%, respectively. The length and width of film opening by the seeding device were 6.8 cm and 3.2 cm that can be considered as sufficient for garlic sprouting. Yet there is need to improve the developed prototype garlic planter in the metering and seeding devices in order to increase the single clove metering rate and to reduce the miss-planted rate, and fall rate. - COLLAPSE
    31 March 2025