• Journal of Intelligence and Information Systems
• /
• v.26 no.2
• /
• pp.1-25
• /
• 2020

In this paper, we suggest an application system architecture which provides accurate, fast and efficient automatic gasometer reading function. The system captures gasometer image using mobile device camera, transmits the image to a cloud server on top of private LTE network, and analyzes the image to extract character information of device ID and gas usage amount by selective optical character recognition based on deep learning technology. In general, there are many types of character in an image and optical character recognition technology extracts all character information in an image. But some applications need to ignore non-of-interest types of character and only have to focus on some specific types of characters. For an example of the application, automatic gasometer reading system only need to extract device ID and gas usage amount character information from gasometer images to send bill to users. Non-of-interest character strings, such as device type, manufacturer, manufacturing date, specification and etc., are not valuable information to the application. Thus, the application have to analyze point of interest region and specific types of characters to extract valuable information only. We adopted CNN (Convolutional Neural Network) based object detection and CRNN (Convolutional Recurrent Neural Network) technology for selective optical character recognition which only analyze point of interest region for selective character information extraction. We build up 3 neural networks for the application system. The first is a convolutional neural network which detects point of interest region of gas usage amount and device ID information character strings, the second is another convolutional neural network which transforms spatial information of point of interest region to spatial sequential feature vectors, and the third is bi-directional long short term memory network which converts spatial sequential information to character strings using time-series analysis mapping from feature vectors to character strings. In this research, point of interest character strings are device ID and gas usage amount. Device ID consists of 12 arabic character strings and gas usage amount consists of 4 ~ 5 arabic character strings. All system components are implemented in Amazon Web Service Cloud with Intel Zeon E5-2686 v4 CPU and NVidia TESLA V100 GPU. The system architecture adopts master-lave processing structure for efficient and fast parallel processing coping with about 700,000 requests per day. Mobile device captures gasometer image and transmits to master process in AWS cloud. Master process runs on Intel Zeon CPU and pushes reading request from mobile device to an input queue with FIFO (First In First Out) structure. Slave process consists of 3 types of deep neural networks which conduct character recognition process and runs on NVidia GPU module. Slave process is always polling the input queue to get recognition request. If there are some requests from master process in the input queue, slave process converts the image in the input queue to device ID character string, gas usage amount character string and position information of the strings, returns the information to output queue, and switch to idle mode to poll the input queue. Master process gets final information form the output queue and delivers the information to the mobile device. We used total 27,120 gasometer images for training, validation and testing of 3 types of deep neural network. 22,985 images were used for training and validation, 4,135 images were used for testing. We randomly splitted 22,985 images with 8:2 ratio for training and validation respectively for each training epoch. 4,135 test image were categorized into 5 types (Normal, noise, reflex, scale and slant). Normal data is clean image data, noise means image with noise signal, relfex means image with light reflection in gasometer region, scale means images with small object size due to long-distance capturing and slant means images which is not horizontally flat. Final character string recognition accuracies for device ID and gas usage amount of normal data are 0.960 and 0.864 respectively.

• Journal of Biomedical Engineering Research
• /
• v.25 no.4
• /
• pp.261-268
• /
• 2004

Osteoporosis is a clinical condition in which the amount of bone tissue is reduced and the likelihood of fracture is increased. It is known that the electrical property of the bone is related to its density, and, in particular, the electrical resistance of the bone decreases as the bone loss increases. This implies that the electrical property of bone may be an useful parameter to diagnose osteoporosis, provided that it can be readily measured. The study attempted to evaluate the electrical conductivity of bone using a technique of electrical impedance tomography (EIT). It nay not be easy in general to get an EIT for the bone due to the big difference (an order of 2) of electrical properties between the bone and the surrounding soft tissue. In the present study, we took an adaptive mesh regeneration technique originally developed for the detection of two phase boundaries and modified it to be able to reconstruct the electrical conductivity inside the boundary provided that the geometry of the boundary was given. Numerical simulation was carried out for a tibia phantom, circular cylindrical phantom (radius of 40 mm) inside of which there is an ellipsoidal homeogenous tibia bone (short and long radius are 17 mm and 15 mm, respectively) surrounded by the soft tissue. The bone was located in the 15 mm above from the center of the circular cross section of the phantom. The electrical conductivity of the soft tissue was set to be 4 mS/cm and varies from 0.01 to 1 ms/cm for the bone. The simulation considered measurement errors in order to look into its effects. The simulated results showed that, if the measurement error was maintained less than 5 %, the reconstructed electrical conductivity of the bone was within 10 % errors. The accuracy increased with the electrical conductivity of the bone, as expected. This indicates that the present technique provides more accurate information for osteoporotic bones. It should be noted that tile simulation is based on a simple two phase image for the bone and the surrounding soft tissue when its anatomical information is provided. Nevertheless, the study indicates the possibility that the EIT technique may be used as a new means to detect the bone loss leading to osteoporotic fractures.

• Journal of the Korean Institute of Traditional Landscape Architecture
• /
• v.35 no.2
• /
• pp.87-97
• /
• 2017

Paulownia species has long been recognized in Korean traditional culture and the values of the species have been researched in various focuses. However, studies on distribution and ecological characteristics of the species are still needed. This study aimed to identify distribution trends and ecological characteristics of two Paulownia species in Busan metropolitan city using the MaxEnt model. The MaxEnt model was established based on the environmental factors such as positioning information of the Paulownia species, topography, climate and degree of anthropogenic disturbance potentiality (ADP), which was collected in the on-site research. The study verified that the accuracy of the model was appropriate as the AUC value of Paulownia coreana and P. tomentosa was 0.809, respectively. In terms of the distribution trends of the two Paulownia species in the research area depending on the distribution model, they were both mainly distributed in downtown where built-up area and bare ground were densely concentrated. The potential distribution area of the two species was identified as $137.4km^2$ for P. coreana and $135.0km^2$ for P. tomentosa. The distribution probability was high in Jung-gu, Dongrae-gu, Busanjin-gu and Yeonje-gu. As a result of the analysis on contribution of the environmental factors, it was turned out that the degree of anthropogenic disturbance potentiality (ADP) contributed to distribution of P. coreana and P. tomentosa by about 50%, and the contribution of the environmental factors had a positive correlation with the degree of ADP. The elevation had a negative correlation with both the two species, which was considered because the species must compete more with native species in natural habitats as the altitude above sea level rises. The research findings demonstrated numerically that the distribution of P.coreana and P. tomentosa depended on artificial activities, and indicated the relevance with the Korean traditional landscape. These findings are expected to provide meaningful information in using, preserving and restoring Paulownia species.

• Korean journal of applied entomology
• /
• v.54 no.4
• /
• pp.401-407
• /
• 2015

This study describes the development of a method for monitoring Aphis gossypii in greenhouse cucumber fields that was used during 2013 and 2014. The dispersion pattern of A. gossypii was determined by commonly used methods: Taylor's power law (TPL) and Iwao's patchiness regression (IPR). The sample unit was determined by linear regression analysis between mean density of sample unit versus whole plant. The optimum sample unit for different plant growth stages was two leaves (median and the lowest + 1 leaf) when the total number of leaves was less than nine, and three leaves (4th, 7th from canopy, and the lowest +1 leaf) when the total number of leaves was greater than nine. A. gossypii showed an aggregated distribution pattern, as the slopes of both TPL and IPR lines were greater than 1. TPL provided a better description of the mean-variance relationship than did IPR. The slopes and intercepts of TPL and IPR from leaf samples did not differ between the surveyed years. Fixed precision levels (D) for a sequential sampling plan were developed using Green's and Kuno's equations based on the number of aphid in a leaf sample. Green's method was more efficient than Kuno's to stop sampling. The number of samples needed to estimate the density of A. gossypii increased at higher D levels and lower mean densities. The cumulative number of aphids needed to stop sampling increased at higher D levels and with fewer plants sampled. Thus to estimate 10 aphids per leaf, 13 plants needed to be sampled, and the cumulative number of aphids to stop sampling was 131.

• Korean Journal of Agricultural and Forest Meteorology
• /
• v.20 no.4
• /
• pp.277-283
• /
• 2018

The crop growth model has been widely used for climate change impact assessment. Crop growth model require genetic coefficients for simulating growth and yield. In order to determine the genetic coefficients, regional growth monitoring data or yield trial data of crops has been used to calibrate crop growth model. The aim of this study is to verify that yield trial data of corn is appropriate to calibrate genetic coefficients of CERES-MAIZE. Field experiment sites were Suwon, Jinju, Daegu and Changwon. The distance from the weather station to the experimental field were from 1.3km to 27km. Genetic coefficients calibrated by yield trial data showed good performance in silking day. The genetic coefficients associated with silking are determined only by temperature. In CERES-MAIZE model, precipitation or irrigation does not have a significant effect on phenology related genetic coefficients. Although the effective distance of the temperature could vary depending on the terrain, reliable genetic coefficients were obtained in this study even when a weather observation site was within a maximum of 27 km. Therefore, it is possible to estimate the genetic coefficients by yield trial data in study area. However, the yield-related genetic coefficients did not show good results. These results were caused by simulating the water stress without accurate information on irrigation or rainfall. The yield trial reports have not had accurate information on irrigation timing and volume. In order to obtain significant precipitation data, the distance between experimental field and weather station should be closer to that of the temperature measurement. However, the experimental fields in this study was not close enough to the weather station. Therefore, When determining the genetic coefficients of regional corn yield trial data, it may be appropriate to calibrate only genetic coefficients related to phenology.

• Korean Journal of Remote Sensing
• /
• v.38 no.6_2
• /
• pp.1691-1707
• /
• 2022

Cheonji Lake, the caldera of Baekdu Mountain, located on the border of the Korean Peninsula and China, alternates between melting and freezing seasonally. There is a magma chamber beneath Cheonji, and variations in the magma chamber cause volcanic antecedents such as changes in the temperature and water pressure of hot spring water. Consequently, there is an abnormal region in Cheonji where ice melts quicker than in other areas, freezes late even during the freezing period, and has a high-temperature water surface. The abnormal area is a discharge region for hot spring water, and its ice gradient may be used to monitor volcanic activity. However, due to geographical, political and spatial issues, periodic observation of abnormal regions of Cheonji is limited. In this study, the degree of ice change in the optimal region was quantified using a Landsat -5/-7/-8 optical satellite image and a Modified U-Net regression model. From January 22, 1985 to December 8, 2020, the Visible and Near Infrared (VNIR) band of 83 Landsat images including anomalous regions was utilized. Using the relative spectral reflectance of water and ice in the VNIR band, unique data were generated for quantitative ice variability monitoring. To preserve as much information as possible from the visible and near-infrared bands, ice gradient was noticed by applying it to U-Net with two encoders, achieving good prediction accuracy with a Root Mean Square Error (RMSE) of 140 and a correlation value of 0.9968. Since the ice change value can be seen with high precision from Landsat images using Modified U-Net in the future may be utilized as one of the methods to monitor Baekdu Mountain's volcanic activity, and a more specific volcano monitoring system can be built.