• Proceedings of the Korean Society of Crop Science Conference
• /
• 2022.10a
• /
• pp.81-81
• /
• 2022

Recently, many studies on big data based smart farming have been conducted. Research to quantify morphological characteristics using image data from various crops in smart farming is underway. Rice is one of the most important food crops in the world. Much research has been done to predict and model rice crop yield production. The number of productive tillers per plant is one of the important agronomic traits associated with the grain yield of rice crop. However, modeling the basic growth characteristics of rice requires accurate data measurements. The existing method of measurement by humans is not only labor intensive but also prone to human error. Therefore, conversion to digital data is necessary to obtain accurate and phenotyping quickly. In this study, we present an image-based method to predict leaf number and evaluate tiller number of individual rice crop using YOLOv5 deep learning network. We performed using various network of the YOLOv5 model and compared them to determine higher prediction accuracy. We ako performed data augmentation, a method we use to complement small datasets. Based on the number of leaves and tiller actually measured in rice crop, the number of leaves predicted by the model from the image data and the existing regression equation were used to evaluate the number of tillers using the image data.

• Proceedings of the Korean Society of Crop Science Conference
• /
• 2022.10a
• /
• pp.315-315
• /
• 2022

• IEMEK Journal of Embedded Systems and Applications
• /
• v.18 no.6
• /
• pp.285-292
• /
• 2023

With the wide spread of smart farms and the advancements in IoT technology, it is easy to obtain additional data in addition to crop images. Consequently, deep learning-based crop disease diagnosis research utilizing multimodal data has become important. This study proposes a crop disease diagnosis method using multimodal supervised contrastive learning by expanding upon the multimodal self-supervised learning. RandAugment method was used to augment crop image and time series of environment data. These augmented data passed through encoder and projection head for each modality, yielding low-dimensional features. Subsequently, the proposed multimodal supervised contrastive loss helped features from the same class get closer while pushing apart those from different classes. Following this, the pretrained model was fine-tuned for crop disease diagnosis. The visualization of t-SNE result and comparative assessments of crop disease diagnosis performance substantiate that the proposed method has superior performance than multimodal self-supervised learning.

• Korean Journal of Remote Sensing
• /
• v.33 no.5_2
• /
• pp.751-762
• /
• 2017

A crop information system can provide information regarding crop distribution, crop growth conditions, crop yield in various forms such as monitoring, forecasting, estimation or analysis. This paper presents the design and construction of a crop information system based on data collected in Korea, USA, and China. Therein, climate data including temperature, precipitation,solar radiation are used to evaluate the impact on crop growth, NDVI (Normalized Difference Vegetation Index) data is used in crop monitoring, and crop map data is utilized for the management of crop distribution. The system has achieved three prominent results: 1) Providing information with high frequency, 2) Automatically creating the report through the analysis of the data, 3) The users to easily approach the system and retrieve the information.

• Magazine of the Korean Society of Agricultural Engineers
• /
• v.39 no.6
• /
• pp.31-40
• /
• 1997

Potential impacts for unfavourable weather conditions and the assessment of the magnitudes of their adverse effects on crop yields were studied. EPIC model was investigated for its capability on crop yield predictions for rice and soybean. Weather generationmodel was used to generate long-term climatic data. The model was verified with ohserved climate data of Suwon city. Fifty years weather data including abnormal conditions were generated and used for crop yield simulation by EPIC model. Crop yield probability function was derived from simulated crop yield data, which followed normal distribution. Probable crop yield reductions due to abnormal weather conditions were also analyzed.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.52 no.4
• /
• pp.393-402
• /
• 2007

Remote sensing data can be integrated into crop models, making simulation improved. A crop model that uses remote sensing data was evaluated for its capability, which was performed through comparing three different methods of canopy measurement for cotton(Gossypium hirsutum L.). The measurement methods used were leaf area index(LAI), hand-held remotely sensed perpendicular vegetation index(PVI), and satellite remotely sensed PVI. Simulated values of cotton growth and lint yield showed reasonable agreement with the corresponding measurements when canopy measurements of LAI and hand-held remotely sensed PVI were used for model calibration. Meanwhile, simulated lint yields involving the satellite remotely sensed PVI were in rough agreement with the measured lint yields. We believe this matter could be improved by using remote sensing data obtained from finer resolution sensors. The model not only has simple input requirements but also is easy to use. It promises to expand its applicability to other regions for crop production, and to be applicable to regional crop growth monitoring and yield mapping projects.

• Proceedings of the Korean Society of Crop Science Conference
• /
• 2017.06a
• /
• pp.58-58
• /
• 2017

The advent of next generation sequencing technology has elicited plenty of sequencing data available in agriculturally relevant plant species. For most crop species, it is too expensive to obtain the whole genome sequence data with sufficient coverage. Thus, many approaches have been developed to bring down the cost of NGS. Genotyping-by-sequencing (GBS) is a cost-effective genotyping method for complex genetic populations. GBS can be used for the analysis of genomic selection (GS), genome-wide association study (GWAS) and constructing haplotype and genetic linkage maps in a variety of plant species. For efficiently dealing with plant GBS data, the TASSEL-GBS pipeline is one of the most popular choices for many researchers. TASSEL-GBS is JAVA based a software package to obtain genotyping data from raw GBS sequences. Here, we describe application of GBS and bioinformatics pipeline of TASSEL-GBS for analyzing plant genetics data.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.22 no.2
• /
• pp.68-78
• /
• 2020

Crop models have been used to predict vegetable crop yield, which would have a considerable economic impact on consumers as well as producers. A small number of models have been developed to estimate growth and yield of vegetables due to limited availability of growth observation data in high-quality. In this study, we aimed to analyze the protocols designed for collection of the observation data for major vegetable crops including cabbage, radish, garlic, onion and pepper. We also designed the protocols suitable for development and verification of a vegetable crop growth model. In particular, different measures were proposed to improve the existing protocol used by Statistics Korea (KOSTAT) and Rural Development Administration (RDA), which would enhance reliability of parameter estimation for the crop model. It would be advantageous to select sampling sites in areas where reliable weather observation data can be obtained because crop models quantify the response of crop growth to given weather conditions. It is recommended to choose multiple sampling sites where climate conditions would differ. It is crucial to collect time series data for comparison between observed and simulated crop growth and yield. A crop model can be developed to predict actual yield rather than attainable yield using data for crop damage caused by diseases and pests as well as weather anomalies. A bigdata platform where the observation data are to be shared would facilitate the development of crop models for vegetable crops.

• Proceedings of the Korean Society of Crop Science Conference
• /
• 2004.04a
• /
• pp.100-101
• /
• 2004

[ $\bigcirc$ ] In the growth simulation using genetic coefficients calculated with fooled data under various condition, WAGT was not higher and LAI, WLVG, WSO were higher, but WST was similar before grain-filling stage after the became lower because of higher translocation of carbohydrates than in the growth simulation using genetic coefficients calculated with data under high nitrogen applicated condition. $\bigcirc$ Genetic coefficients should be calculated with data showing potential in ORYZA2000, but under 180 kg and 240 kg N condition in 2003, plants were infected by panicle blast and also yield was not higher than under 120 kg N condition showing not potential condition and therefore not appropriate for genetic coefficients estimation compared with pooled data from various condition.

• Proceedings of the KSRS Conference
• /
• 1999.11a
• /
• pp.479-484
• /
• 1999

Rice crop has relatively short growing season during the summer in Korea and, therefore, it is often difficult to acquire cloud-free imagery on time. This study was attempt to define the temporal characteristics of radar backscattering observed from satellite L-band SAR data on different growing stages of rice crop. Six scenes of multi-temporal JERS SAR data were obtained from the transplanting season to the harvesting month of October. Six layers of multi-temporal SAR data were registered on a common geographic coordinate system. Using topographic maps, field collected data, and Landsat TM data, several sample rice fields were delineated from the imagery and their relative radar backscatters were calculated by using a set of reference targets. The temporal pattern of radar backscattering was very distinctive by the growing stage of rice crop. It was also separable between two types of rice fields having different cultivation practices. Considering the temporal characteristics of radar backscattering observed from the study, it is obvious that a certain date of the growing season can be more effective to delineate the exact area of the cultivated rice crop field.