• Proceedings of the Korean Society of Crop Science Conference
• /
• 2022.10a
• /
• pp.80-80
• /
• 2022

Estimation of heading date of paddy rice is laborious and time consuming. Therefore, automatic estimation of heading date of paddy rice is highly essential. In this experiment, deep learning classification models were used to classify two difference categories of rice (vegetative and reproductive stage) based on the panicle initiation of paddy field. Specifically, the dataset includes 444 slanted view images belonging to two categories and was then expanded to include 1,497 images via IMGAUG data augmentation technique. We adopt two transfer learning strategies: (First, used transferring model weights already trained on ImageNet to six classification network models: VGGNet, ResNet, DenseNet, InceptionV3, Xception and MobileNet, Second, fine-tuned some layers of the network according to our dataset). After training the CNN model, we used several evaluation metrics commonly used for classification tasks, including Accuracy, Precision, Recall, and F1-score. In addition, GradCAM was used to generate visual explanations for each image patch. Experimental results showed that the InceptionV3 is the best performing model in terms of the accuracy, average recall, precision, and F1-score. The fine-tuned InceptionV3 model achieved an overall classification accuracy of 0.95 with a high F1-score of 0.95. Our CNN model also represented the change of rice heading date under different date of transplanting. This study demonstrated that image based deep learning model can reliably be used as an automatic monitoring system to detect the heading date of rice crops using CCTV camera.

• Journal of IKEEE
• /
• v.27 no.4
• /
• pp.512-517
• /
• 2023

In this study, we propose a liquid crystal-based polarization control technology that can control polarization by adjusting the voltage applied to the liquid crystal, and apply it to a Closed-circuit Television (CCTV) to transmit only the desired angle of polarized light. CCTV with conventional polarizing films cannot control polarization because they focus on backlight compensation, so light reflected from the water surface or highlights reflected from vehicles interfere with subject identification. However, the Vertical Alignment mode allows the polarization to be adjusted electrically, so that only the polarized light at the user's desired angle is transmitted, eliminating reflected highlights. The images obtained using this technique are optimized by computer software. Liquid crystal polarization panels, which can electrically control the polarization angle, transmittance, and polarization rate, have been applied to polarized image monitoring device to improve subject identification in conventional CCTV.

• Journal of Microbiology and Biotechnology
• /
• v.33 no.9
• /
• pp.1228-1237
• /
• 2023

The CRISPR-Cas system has emerged as the most efficient genome editing technique for a wide range of cells. Delivery of the Cas9-sgRNA ribonucleoprotein complex (Cas9 RNP) has gained popularity. The objective of this study was to develop a quantitative polymerase chain reaction (qPCR)-based assay to quantify the double-strand break reaction mediated by Cas9 RNP. To accomplish this, the dextransucrase gene (dsr) from Leuconostoc citreum was selected as the target DNA. The Cas9 protein was produced using recombinant Escherichia coli BL21, and two sgRNAs were synthesized through in vitro transcription to facilitate binding with the dsr gene. Under optimized in vitro conditions, the 2.6 kb dsr DNA was specifically cleaved into 1.1 and 1.5 kb fragments by both Cas9-sgRNA365 and Cas9-sgRNA433. By monitoring changes in dsr concentration using qPCR, the endonuclease activities of the two Cas9 RNPs were measured, and their efficiencies were compared. Specifically, the specific activities of dsr365RNP and dsr433RNP were 28.74 and 34.48 (unit/㎍ RNP), respectively. The versatility of this method was also verified using different target genes, uracil phosphoribosyl transferase (upp) gene, of Bifidobacterium bifidum and specific sgRNAs. The assay method was also utilized to determine the impact of high electrical field on Cas9 RNP activity during an efficient electroporation process. Overall, the results demonstrated that the qPCR-based method is an effective tool for measuring the endonuclease activity of Cas9 RNP.

• Korean Journal of Remote Sensing
• /
• v.39 no.6_1
• /
• pp.1401-1411
• /
• 2023

In the face of escalating utilization of the marine spatial domain, conflicts have emerged among stakeholders, necessitating effective management strategies beyond conventional government permits and regulations. Particularly within the domain of aquaculture, operational oversight relies on a localized licensing system, posing challenges in accurately assessing the prevailing circumstances. This research employs synthetic aperture radar (SAR) imagery as a tool to monitor coastal aquaculture fish farms, aimed at enhancing insights into management protocols. Leveraging Sentinel-1A imagery and time series SAR data integration, a superimposition technique is utilized, facilitating noise reduction while retaining crucial information regarding smaller-scale facilities, such as fish farms. Through analysis of VH polarization data, a detection overall accuracy of approximately 88% for coastal fish farms was achieved. The findings of this study offer potential applications in the continuous monitoring of aquaculture farms in correspondence with seasonal variations in aquaculture yields, thereby proposing frameworks for the establishment of effective management cycles for marine space utilization.

• Journal of the Korean Society of Radiology
• /
• v.81 no.2
• /
• pp.302-309
• /
• 2020

Fabry disease is a rare X-linked metabolic disorder that is characterized by the accumulation of glycosphingolipids in various organs, resulting from the deficiency of alpha-galactosidase A. Cardiac involvement is relatively common; myocardial inflammation, left ventricular hypertrophy, and myocardial fibrosis secondary to abnormal lipid deposition in myocytes are often observed. Hence, the diagnosis of cardiac involvement is crucial for evaluating patient prognosis. Cardiac MRI is the standard technique for measuring the function, volume, and mass of the ventricles. It is also useful for myocardial tissue characterizations. The evaluation of native myocardial T1 values can facilitate early diagnosis of cardiac involvement, while measurements of left ventricular myocardial mass can be used to monitor treatment outcomes, in patients with Fabry disease. Consequently, cardiac MRI can provide useful information for diagnosing, monitoring, and treating patients with Fabry disease.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.2
• /
• pp.348-369
• /
• 2024

With the advancement of Industry 4.0 and Industrial Internet of Things (IIoT), manufacturing increasingly seeks automation and intelligence. Temperature and vibration monitoring are essential for machinery health. Traditional abnormal state detection methodologies often overlook the intricate frequency characteristics inherent in vibration time series and are susceptible to erroneously reconstructing temperature abnormalities due to the highly similar waveforms. To address these limitations, we introduce synergistic, end-to-end, unsupervised Frequency-Time Domain Memory-Enhanced Autoencoders (FTD-MAE) capable of identifying abnormalities in both temperature and vibration datasets. This model is adept at accommodating time series with variable frequency complexities and mitigates the risk of overgeneralization. Initially, the frequency domain encoder processes the spectrogram generated through Short-Time Fourier Transform (STFT), while the time domain encoder interprets the raw time series. This results in two disparate sets of latent representations. Subsequently, these are subjected to a memory mechanism and a limiting function, which numerically constrain each memory term. These processed terms are then amalgamated to create two unified, novel representations that the decoder leverages to produce reconstructed samples. Furthermore, the model employs Spectral Entropy to dynamically assess the frequency complexity of the time series, which, in turn, calibrates the weightage attributed to the loss functions of the individual branches, thereby generating definitive abnormal scores. Through extensive experiments, FTD-MAE achieved an average ACC and F1 of 0.9826 and 0.9808 on the CMHS and CWRU datasets, respectively. Compared to the best representative model, the ACC increased by 0.2114 and the F1 by 0.1876.

• The Korean Journal of Nuclear Medicine Technology
• /
• v.21 no.2
• /
• pp.74-79
• /
• 2017

Purpose Proton therapy can deliver an optimal dose to tumor while reducing unnecessary dose to normal tissue as compared the conventional photon therapy. As proton beams are irradiated into tissue, various positron emitters are produced via nuclear fragmentation reactions. These positron emitters could be used for the dose verification by using PET. However, the short half-life of the radioisotopes makes it hard to obtain the enough amounts of events. The aim of this study is to investigate the effect of off-line PET imaging scan time on the PET image quality. Materials and Methods The various diameters of spheres (D=37, 28, 22 mm) filled with distilled water were inserted in a 2001 IEC body phantom. Then proton beams (100 MU) were irradiated into the center of the each sphere using the wobbling technique with the gantry angle of $0^{\circ}$. The modulation widths of the spread out bragg peak were 16.4, 14.7 and 9.3 cm for the spheres of 37, 28 and 22 mm in diameters respectively. After 5 min of the proton irradiation, the PET images of the IEC body phantom were obtained for 50 min. The PET images with different time courses (0-10 min, 11-20 min, 21-30 min, 31-40 min and 41-50 min) were obtained by dividing the frame with a duration of 10 min. In order to evaluate the off-line PET image quality with the different time courses, the contrast-to-noise ratio (CNR) of the PET image calculated for each sphere. Results The CNRs of the sphere (D=37 mm) were 0.43, 0.42, 0.40, 0.31 and 0.21 for the time courses of 0-10 min, 11-20 min, 21-30 min, 31-40 min and 41-50 min respectively. The CNRs of the sphere (D=28 mm) were 0.36, 0.32, 0.27, 0.19 and 0.09 for the time courses of 0-10 min, 11-20 min, 21-30 min, 31-40 min and 41-50 min respectively. The CNR of 37 mm sphere was decreased rapidly after 30 min of the proton irradiation. In case of the spheres of 28 mm and 22 mm, the CNR was decreased drastically after 20 min of the irradiation. Conclusion The off-line PET imaging time is an important factor for the monitoring of the proton therapy. In case of the lesion diameter of 22 mm, the off-line PET image should be obtained within 25 min after the proton irradiation. When it comes to small size of tumor, the long PET imaging time will be beneficial for the proton therapy treatment monitoring.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.17 no.1
• /
• pp.15-24
• /
• 2015

$CH_4$ is a trace gas and one of the key greenhouse gases, which requires continuous and systematic monitoring. The application of eddy covariance technique for $CH_4$ flux measurement requires a fast-response, laser-based spectroscopy. The eddy covariance measurements have been used to monitor $CO_2$ fluxes and their data processing procedures have been standardized and well documented. However, such processes for $CH_4$ fluxes are still lacking. In this note, we report the first measurement of $CH_4$ flux in a rice paddy by employing the eddy covariance technique with a recently commercialized wavelength modulation spectroscopy. $CH_4$ fluxes were measured for five consecutive days before and after the rice transplanting at the Gimje flux monitoring site in 2012. The commercially available $EddyPro^{TM}$ program was used to process these data, following the KoFlux protocol for data-processing. In this process, we quantified and documented the effects of three key corrections: (1) frequency response correction, (2) air density correction, and (3) spectroscopic correction. The effects of these corrections were different between daytime and nighttime, and their magnitudes were greater with larger $CH_4$ fluxes. Overall, the magnitude of $CH_4$ flux increased on average by 20-25% after the corrections. The National Center for AgroMeteorology (www.ncam.kr) will soon release an updated KoFlux program to public users, which includes the spectroscopic correction and the gap-filling of $CH_4$ flux.

• Journal of Environmental Impact Assessment
• /
• v.27 no.6
• /
• pp.674-684
• /
• 2018

The purpose of this study is to construct ecological pond using LID technique in order to create naturally comfortable community space in urban idle space. The specification of the ecological pond is $110m^2$ of surface area, $0.45{\pm}0.02m$ of average depth, and bed material is composed of gravel (diameter ${\leq}60mm$), sand (diameter ${\leq}2mm$) and bentonite. Rainfall and water depth monitoring were conducted to determine the annual characteristics of inflow of the water for the ecological pond, result of total rainfall was 1,287 mm and showed a seasonal imbalance that accounted for 71.3% (918 mm) during July to August, but the annual mean water depth was kept constant at $0.45{\pm}0.02m$ due to the secondary water source. Annual trends of basic water quality showed a significant changes according to the season, such as water temperature ($5.2{\sim}28.8^{\circ}C$), DO (5.0 ~ 13.8 mg/L), EC ($113{\sim}265{\mu}S/cm$). BOD, COD, TN, and TP in physicochemical water quality tended to increase after October, but the ion parameters such as $NH_3$ and $PO_4{^{3-}}$ were generally low. Phytoplankton indicators Chl-a and BGA (blue green algae) showed a sharp increase from July to August, and green algae (Selenastrum bibraianum, Pediastrum boryanum etc.) and filamentous blue green algae (Phormidium sp.) emerged as a dominant species. The ion parameters ($F^-$, $Na^+$, $K^+$, $Mg^{2+}$, $Ca^{2+}$) were strongly correlated with the $Cl^-$ as a conservative substance (R=0.70~0.97, p<0.05). Water quality was influenced by the ambient environment such as seasonal changes or rainfall, and it was closely related to fluctuation of the inflow of the water. In the future, it is necessary to consider ecological connections by referring to the characteristics surveyed in this study in order to effectively manage the water quality and biodiversity of the ecological pond in idle space.

• Korean Journal of Soil Science and Fertilizer
• /
• v.38 no.3
• /
• pp.127-133
• /
• 2005

The measurement technique using laser distometer to estimate soil erosion on sloped upland was assessed for its possibility for application. This technique was practiced in lysimeter installed in Chuncheon and Pyeongchang in 2001. The lysimeter installed at Chuncheon has 12% slope, 10 m slope length, 2 m width, and soil texture was sandy loam, while the lysimeter installed at Pyeongchang showed 23% slope, 15 m slope length, 5 m width, and sandy loam soil. Change of surface soil height was monitored using laser distance meter before and after rainy season on same spots. The Investigated periods in Chuncheon and Pyeongchang were from 22 June to 22 September and from July to October, respectively. Precipitation in Chuncheon and Pyeongchang in this period was 892.2 and 931.9 mm, respectively. Rainfall over 60% of annual precipitation was concentrated on July and August in Chuncheon and September and October in Pyeongchang, respectively. By monitoring the change of surface soil height using laser distance meter before and after rainy season, eroded soil surface in up-down fallow field in Chuncheon was 0.874 cm, while eroded soil surface in slant furrow field in Pyeongchang was 1.127 cm, which correspond to 79.5 and $98.0MT\;ha^{-1}$. Soil erosion increased as furrow length increased. Amounts of estimated soil loss using laser distance meter was 0.98-1.18 times higher compared to the estimated values through Iysimeter experiment, which implies possibility for application to monitor soil loss particularly in up-down and slant furrowed field. However, in the lysimeter with contoured tillage, amount of estimated soil loss using laser distance meter was approximately a half compared to that from lysimeter experiment, which implies inadequacy of distance meter application in contour- tillage field. The great soil loss difference between distance meter and lysimeter might be caused by disruption of some of the contoured furrows in lysimeter. The measurement technique using distometer in this study could be useful to estimate soil loss especially in up-down and slant-tillage fields.