• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.55 no.4
• /
• pp.312-318
• /
• 2010

This study was carried out to develop and test a prototype program that recommends the nitrogen topdressing rate using the color digital camera image taken from rice field at panicle initiation stage (PIS). This program comprises four models to estimate shoot N content (PNup) by color digital image analysis, shoot N accumulation from PIS to maturity (PHNup), yield, and protein content of rice. The models were formulated using data set from N rate experiments in 2008. PNup was found to be estimated by non-linear regression model using canopy cover and normalized green values calculated from color digital image analysis as predictor variables. PHNup could be predicted by quadratic regression model from PNup and N fertilization rate at panicle initiation stage with $R^2$ of 0.923. Yield and protein content of rice could also be predicted by quadratic regression models using PNup and PHNup as predictor variables with $R^2$ of 0.859 and 0.804, respectively. The performance of the program integrating the above models to recommend N topdressing rate at PIS was field-tested in 2009. N topdressing rate prescribed for the target protein content of 6.0% by the program were lower by about 30% compared to the fixed rate of 30% that is recommended conventionally as the split application rate of N fertilizer at PIS, while rice yield in the plots top-dressed with the prescribed N rate were not different from those of the plots top-dressed with the fixed N rates of 30% and showed a little lower or similar protein content of rice as well. And coefficients of variation in rice yield and quality parameters were reduced substantially by the prescribed N topdressing. These results indicate that the N rate recommendation using the analysis of color digital camera image is promising to be applied for precise management of N fertilization. However, for the universal and practical application the component models of the program are needed to be improved so as to be applicable to the diverse edaphic and climatic condition.

• Economic and Environmental Geology
• /
• v.41 no.3
• /
• pp.299-314
• /
• 2008

The NE-trending Honam shear zone is a broad, dextral strike-slip fault zone between the southern margin of the Okcheon Belt and the Precambrian Yeongnam Massif in South Korea and is parallel to the trend of Sinian deformation that is conspicuous in Far East Asia. In this paper, we report geochemical and isotopic(Sr and Nd) data of mylonitic quartz-muscovite Precambrian gneisses and surrounding foliated hornblende-biotite granitoids near the Myeongho area in the Yecheon Shear Zone, a representative segment of the Honam Shear Zone. Foliated hornblende-biotite granitoids commonly plot in the granodiorite field($SiO_2=61.9-67.1\;wt%$ and $Na_2O+K_2O=5.21-6.99\;wt%$) on $SiO_2$ vs. $Na_2O+K_2O$ discrimination diagram, whereas quartz-muscovite Precambrian orthogneisses plot in the granite field. The foliated hornblende-biotite granitoids are mostly calcic and calc-alkalic and are dominantly magnesian in a modified alkali-lime index(MALI) and Fe# [$=FeO_{total}(FeO_{total}+MgO)$] versus $SiO_2$ diagrams, which correspond with geochemical characteristics of Cordilleran Mesozoic batholiths. The foliated hornblende-biotite granitoids have molar ratios of $Al_2O_3/(CaO+Na_2O+K_2O)$ ranging from 0.89 to 1.10 and are metaluminous to weakly peraluminous, indicating I type. In contrast, Paleoproterozoic orthogneisses have peraluminous compositions, with molar ratios of $Al_2O_3/(CaO+Na_2O+K_2O)$ ranging from 1.11 to 1.22. On trace element spider diagrams normalized to the primitive mantle, the large ion lithophile element(LILE) enrichments(Rb, Ba, Th and U) and negative Ta-Nb-P-Ti anomalies of foliated hornblende-biotite granitoids and mylonitized quartz-muscovite gneisses in the Yecheon Shear Zone are features common to subduction-related granitoids and are also found in granitoids from a crustal source derived from the arc crust of active continental margin. ${\varepsilon}_{Nd}(T)$ and initial Sr-ratio ratios of foliated hornblende-biotite granitoids with suggest the involvement of upper crust-derived melts in granitoid petrogenesis. Foliated hornblende-biotite granitoids in the study area, together with the Yeongju Batholith, show not changing contents of specific elements(Ti, P, Zr, V and Y) from shear zone to the area near the shear zone. These results suggest that no volume changes and geochemical alterations in fluid-rich foliated hornblende-biotite granitoids may occur during deformation, which mass transfer by fluid flow into the shear zone is equal to the mass transfer out of the shear zone.

• The Journal of Korean Society for Radiation Therapy
• /
• v.20 no.2
• /
• pp.109-113
• /
• 2008

Purpose: To evaluate the detector dependency in the various collimator size for Stereotactic Radiosugery (SRS). Materials and Methods: This study was performed with 6 MV photon beam (Varian 21EX, Varian, US) and the measurement detectors are used by ion chamber CC01, CC13 (Wellhofer, Germany) and stereotactic diode detector (SFD, Wellhofer, Germany). SRS collimator size was used by ${\varphi}$5, 10, 20, 30 mm (Brain Lab, Germany). Percentage depth dose (PDD) was measured at SSD 100 cm and field size 10×10 cm from individual detectors. Ouput factor was measured by using same setup of PDD and with maximum dose depth. Data was normalized at field size $10{\times}10\;cm$. Beam profile was measured at SSD 100 cm in SRS collimator ${\varphi}$10, 30 mm and field $10{\times}10\;cm$ and a comparison of FWHM (full width half maximum), penumbra width (20~80%). Results: The CC13 detector was overestimated 16% than other detectors from the PDD in the 5 mm collimator. Output factors were underestimated CC01 28%, CC13 72% in the 5 mm collimator and CC01 9.6%, CC13 25% in the 10 mm collimator than the SFD. Maximum difference was 3% at the FWHM of the dose profile in the 10 mm collimator and difference of the 30 mm collimator was 0% at the FWHM. Penumbra width was increased CC01 122%, CC13 194% in the 10 mm collimator and CC01 68%, CC13 185% in the 30 mm collimator than the SFD. Conclusion: It is very important for accurate dosimetry to select a detector in small field. The SFD was considered with the most accurate dosimeter for small collimator dosimetry in this study.

• The Korean Journal of Nuclear Medicine
• /
• v.32 no.3
• /
• pp.225-237
• /
• 1998

Purpose: We investigated the statistical methods to compose the functional brain map of human working memory and the principal factors that have an effect on the methods for localization. Materials and Methods: Repeated PET scans with successive four tasks, which consist of one control and three different activation tasks, were performed on six right-handed normal volunteers for 2 minutes after bolus injections of 925 MBq $H_2^{15}O$ at the intervals of 30 minutes. Image data were analyzed using SPM96 (Statistical Parametric Mapping) implemented with Matlab (Mathworks Inc., U.S.A.). Images from the same subject were spatially registered and were normalized using linear and nonlinear transformation methods. Significant difference between control and each activation state was estimated at every voxel based on the general linear model. Differences of global counts were removed using analysis of covariance (ANCOVA) with global activity as covariate. Using the mean and variance for each condition which was adjusted using ANCOVA, t-statistics was performed on every voxel To interpret the results more easily, t-values were transformed to the standard Gaussian distribution (Z-score). Results: All the subjects carried out the activation and control tests successfully. Average rate of correct answers was 95%. The numbers of activated blobs were 4 for verbal memory I, 9 for verbal memory II, 9 for visual memory, and 6 for conjunctive activation of these three tasks. The verbal working memory activates predominantly left-sided structures, and the visual memory activates the right hemisphere. Conclusion: We conclude that rCBF PET imaging and statistical parametric mapping method were useful in the localization of the brain regions for verbal and visual working memory.

• The Journal of the Petrological Society of Korea
• /
• v.12 no.1
• /
• pp.16-31
• /
• 2003

The granites in the southern Gimcheon area can be divided into two parts, marginal hornblende biotite granodiorite (Mgd) and central biotite granodiorite to granite (Cgd). Mgd and Cgd are gray in color and display gradational contact relations and are mainly composed of coarse-grained and medium-grained rocks, respectively. Mgd has more frequent and larger mafic enclaves than Cgd, and the two granites partly show parallel foliation at thire contact with gneisses. From representative samples of the granites, K-Ar biotite ages of 197∼207 Ma were obtained. Considering the blocking temperature of biotite, it is suggested that the emplacement age of the granitic magma was probably late Triassic. The anorthite contents of plagioclases in Mgd display less variation than those of Cgd, indicating that Mgd crystallized within a narrow range of temperatures. In the Al$\_$total/-Mg diagram, the biotites from the granites plot within the subalkaline field, and the smooth slope indicates differentiation from a single magma. All amphiboles from the granites belong to magnesio-hornblende. The linear trends of major oxides, AFM and Ba-Sr-Rb indicate that Mgd and Cgd were fractionally differentiated from a single granitic magma body crystallizing from the margin inwards. The relations of modal (Qz+Af) vs. Op, K$_2$O vs. Na$_2$O, Fe$_2$ $O_3$ vs. FeO, Fe$\^$+3/(Fe$\^$+3/+Fe$\^$+2/) and K/Rb vs. Rb/Sr show that they belong to I-type and magnetite-series granitic rocks developed by the progressive melting products of fixed sources. REE data, normalized to chondrite value, have trends of enriched LREE and depleted HREE together with weakly negative Eu anomalies.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.21 no.4
• /
• pp.327-336
• /
• 2019

Soybean is one of the most important crops of which the grains contain high protein content and has been consumed in various forms of food. Soybean plants are generally cultivated on the field and their yield and quality are strongly affected by climate change. Recently, the abnormal climate conditions, including heat wave and heavy rainfall, frequently occurs which would increase the risk of the farm management. The real-time assessment techniques for quality and growth of soybean would reduce the losses of the crop in terms of quantity and quality. The objective of this work was to develop a simple model to estimate the growth of soybean plant using a multispectral sensor mounted on a rotor-wing unmanned aerial vehicle(UAV). The soybean growth model was developed by using simple linear regression analysis with three phenotypic data (fresh weight, dry weight, leaf area index) and two types of vegetation indices (VIs). It was found that the accuracy and precision of LAI model using GNDVI (R²= 0.789, RMSE=0.73 ㎡/㎡, RE=34.91%) was greater than those of the model using NDVI (R²= 0.587, RMSE=1.01 ㎡/㎡, RE=48.98%). The accuracy and precision based on the simple ratio indices were better than those based on the normalized vegetation indices, such as RRVI (R²= 0.760, RMSE=0.78 ㎡/㎡, RE=37.26%) and GRVI (R²= 0.828, RMSE=0.66 ㎡/㎡, RE=31.59%). The outcome of this study could aid the production of soybeans with high and uniform quality when a variable rate fertilization system is introduced to cope with the adverse climate conditions.

• The Korean Journal of Archival Studies
• /
• no.12
• /
• pp.81-135
• /
• 2005

The release of administrative information has been the challenge of our age following the maturation of democratic ideology in our society. However, differences of opinion and conflict still exist between the government and private sectors regarding the issue, and it seems that the technical and policy-related insufficiencies of information and record management that actually operate the release of information are the main causes. From the perspective of records management, records or information are variable in their nature, value, and influence during their life span. The most controversial issue is the records and information in the current stage of carrying out business activities. This is because the records and information pertaining to finished business are but evidence to ascertain the past, and have only a limited relationship to the ideal of the 'democratic participation' by citizens in activities of the public sector. The current information release policies are helpless against the 'absence of information,' or incomplete records, but such weakness can be supplemented by enforcing record management policies that make obligatory the recording of all details of business activities. In addition, it is understood that the installation of 'document offices("Jaryogwan")' that can manage each organization's information and records will be an important starting point to integrate the release, management, and preservation of information and records. Nevertheless, it seems that the concept of 'release' in information release policies refers not to free use by all citizens but is limited to the 'provision' of records according to public requests, and the concept of 'confidential' refers not to treating documents with total secrecy but varies according to the particulars of each situation, making the actual practice of information release difficult. To solve such problems, it is absolutely necessary to collect the opinions of various constituents associated with the recorded information in question, and to effectively mediate the collective opinions and the information release requests coming from applicants, to carry out the business more practically. Especially crucial is the management of the process by which the nature and influence of recorded information changes, so that information which has to be confidential at first may become available for inquiry and use over time through appropriate procedures. Such processes are also part of the duties that record management, which is in charge of the entire life span of documents, must perform. All created records will be captured within a record management system, and the record creation data thus collected will be used as a guide for inquiry and usage. With 'document offices(Jaryogwan)' and 'archives' controlling the entire life span of records, the release of information will become simpler and more widespread. It is undesirable to try to control only through information release policies those records the nature of which has changed because, unlike the ones still in the early stages of their life span and can directly influence business activities, their work has finished, and they have become historical records or evidences pointing to the truth of past events. Even in the past, when there existed no formal policy regarding the release of administrative information, the access and use of archival records were permitted. A more active and expanded approach must be taken regarding the 'usage' of archival records. If the key factor regarding 'release' lies in the provision of information, the key factor regarding 'usage' lies in the quality and level of the service provided. The full-scale usage of archival records must be preceded by the release of such records, and accordingly, a thorough analysis of the nature, content, and value of the records and their changes must be implemented to guarantee the release of information before their use is requested. That must become a central task of document offices and "Today's information" will soon become "yesterday's records," and the "reality" of today will become "history" of the past. The policies of information release and record management share information records as their common objective. As they have a mutual relationship that is supplementary and leads toward perfection, the two policies must both be differentiated and integrated with each another. It is hoped that the policies and business activities of record management will soon become normalized and reformed for effective and fair release of information.

• KOREAN JOURNAL OF CROP SCIENCE
• /
• v.66 no.2
• /
• pp.105-111
• /
• 2021

Rice lodging is an annual occurrence caused by typhoons accompanied by strong winds and strong rainfall, resulting in damage relating to pre-harvest sprouting during the ripening period. Thus, rapid estimations of the area of lodged rice are necessary to enable timely responses to damage. To this end, we obtained images related to rice lodging using a drone in Gimje, Buan, and Gunsan, which were converted to 128 × 128 pixels images. A convolutional neural network (CNN) model, a deep learning model based on these images, was used to predict rice lodging, which was classified into two types (lodging and non-lodging), and the images were divided in a 8:2 ratio into a training set and a validation set. The CNN model was layered and trained using three optimizers (Adam, Rmsprop, and SGD). The area of rice lodging was evaluated for the three fields using the obtained data, with the exception of the training set and validation set. The images were combined to give composites images of the entire fields using Metashape, and these images were divided into 128 × 128 pixels. Lodging in the divided images was predicted using the trained CNN model, and the extent of lodging was calculated by multiplying the ratio of the total number of field images by the number of lodging images by the area of the entire field. The results for the training and validation sets showed that accuracy increased with a progression in learning and eventually reached a level greater than 0.919. The results obtained for each of the three fields showed high accuracy with respect to all optimizers, among which, Adam showed the highest accuracy (normalized root mean square error: 2.73%). On the basis of the findings of this study, it is anticipated that the area of lodged rice can be rapidly predicted using deep learning.

• Journal of the Korean earth science society
• /
• v.43 no.1
• /
• pp.128-139
• /
• 2022

This study aimed to investigate the tectonic setting of the volcanic edifice at Mt. Baekdu by analyzing petrochemical characteristics of Holocene felsic volcanic rocks distributed in the Baekdusan stratovolcano edifice and summit of the Cheonji caldera rim, as well as Pleistocene mafic rocks of the Gaema lava plateau and Changbaishan shield volcano edifice. During the early eruption phases, mafic eruption materials, with composition ranging from alkali basalt to trachybasalt, or from subalkaline (tholeiitic) basalt to basaltic andesite formed the Gaema lava plateau and Changbaishan shield volcanic edifice, whereas the Baekdusan stratovolcano edifice and Holocene tephra deposits near the summit of the Cheonji caldera comprises trachytic and rhyolitic compositions. Analysis results revealed bimodal compositions with a lack of 54-62 SiO₂, between the felsic and mafic volcanic rocks. This suggested that magmatic processes occurred at the locations of extensional tectonic settings in the crust. Mafic volcanic rocks were plotted in the field of within-plate volcanic zones or between within-plate alkaline and tholeiite zones on the tectonic discrimination diagram, and it was in good agreement with the results of the TAS diagram. Felsic volcanic rocks were plotted in the field of within-plate granite tectonic settings on discrimination diagrams of granitic rocks. None of the results were plotted in the field of arc islands or continental margin arcs. The primitive mantle-normalized spider diagram did not show negative (-) anomalies of Nb and Ti, which are distinctive characteristics of subduction-related volcanic rocks, but exhibited similar patterns of ocean island basalt. Trace element compositions showed no evidence of, magmatic processes related to subduction zones, indicating that the magmatic processes forming the Baekdusan volcanic field occurred in an intraplate environment. The distribution of shallow earthquakes in this region supports the results. The volcanic rocks of the Baekdusan volcanic field are interpreted as the result of intraplate volcanism originating from the upwelling of mantle material during the Cenozoic era.

• Journal of Intelligence and Information Systems
• /
• v.28 no.4
• /
• pp.251-273
• /
• 2022

Recently, research on applying text analysis to deep learning has steadily continued. In particular, researches have been actively conducted to understand the meaning of words and perform tasks such as summarization and sentiment classification through a pre-trained language model that learns large datasets. However, existing pre-trained language models show limitations in that they do not understand specific domains well. Therefore, in recent years, the flow of research has shifted toward creating a language model specialized for a particular domain. Domain-specific pre-trained language models allow the model to understand the knowledge of a particular domain better and reveal performance improvements on various tasks in the field. However, domain-specific further pre-training is expensive to acquire corpus data of the target domain. Furthermore, many cases have reported that performance improvement after further pre-training is insignificant in some domains. As such, it is difficult to decide to develop a domain-specific pre-trained language model, while it is not clear whether the performance will be improved dramatically. In this paper, we present a way to proactively check the expected performance improvement by further pre-training in a domain before actually performing further pre-training. Specifically, after selecting three domains, we measured the increase in classification accuracy through further pre-training in each domain. We also developed and presented new indicators to estimate the specificity of the domain based on the normalized frequency of the keywords used in each domain. Finally, we conducted classification using a pre-trained language model and a domain-specific pre-trained language model of three domains. As a result, we confirmed that the higher the domain specificity index, the higher the performance improvement through further pre-training.