In this study, we evaluate artificial neural network (ANN) models that estimate the positions of gamma-ray sources from plastic scintillating fiber (PSF)-based radiation detection systems using different filtering ratios. The PSF-based radiation detection system consists of a single-stranded PSF, two photomultiplier tubes (PMTs) that transform the scintillation signals into electric signals, amplifiers, and a data acquisition system (DAQ). The source used to evaluate the system is Cs-137, with a photopeak of 662 keV and a dose rate of about 5 μSv/h. We construct ANN models with the same structure but different training data. For the training data, we selected a measurement time of 1 minute to secure a sufficient number of data points. Conversely, we chose a measurement time of 10 seconds for extracting time-difference data from the primary data, followed by filtering. During the filtering process, we identified the peak heights of the gaussian-fitted curves obtained from the histogram of the time-difference data, and extracted the data located above the height which is equal to the peak height multiplied by a predetermined percentage. We used percentage values of 0, 20, 40, and 60 for the filtering. The results indicate that the filtering has an effect on the position estimation error, which we define as the absolute value of the difference between the estimated source position and the actual source position. The estimation of the ANN model trained with raw data for the training data shows a total average error of 1.391 m, while the ANN model trained with 20%-filtered data for the training data shows a total average error of 0.263 m. Similarly, the 40%-filtered data result shows a total average error of 0.119 m, and the 60%-filtered data result shows a total average error of 0.0452 m. From the perspective of the total average error, it is clear that the more data are filtered, the more accurate the result is. Further study will be conducted to optimize the filtering ratio for the system and measuring time by evaluating stabilization time for position estimation of the source.

In terms of habitat conservation, it is essential to develop a habitat assessment system that can evaluate not only the suitability of the current habitat, but also the health and stability of the habitat. This study aimed to develop a methodology of habitat quality assessment for endangered species by analyzing various existing habitat assessment methods. The habitat quality assessment consisted of selecting targeted species, planning of assessment, selecting targeted sites, assessing performance, calculating grade, and expert verification. Target sites were selected separately from core and potential habitats using a species distribution model or habitat suitability index. Habitat assessment factors were classified into ecological characteristic, landscape characteristic, and species-habitat characteristic. Ecological characteristic consisted of thirteen factors related to health of tree, vegetation, and soil. Landscape characteristic consisted of five factors related to fragment and connectivity of habitat. Species-habitat characteristic consisted of factors for evaluating habitat suitability depending on target species. Since meanings are different depending on characteristics, habitat quality assessment of this study could be used by classifying results for each characteristic according to various assessment purposes, such as designation of alternative habitats, assessment of restoration project, and protected area valuation for endangered species. Forest habitat quality assessment is expected to play an important role in conservation acts of endangered species in the future through continuous supplementation of this system in regard to quantitative assessment criteria and weighting for each factor with an influence.

In this study, the positions of Cs-137 gamma ray source are estimated from the plastic scintillating fiber bundle sensor with length of 5 m, using machine learning data analysis. Seven strands of plastic scintillating fibers are bundled by black shrink tube and two photomultiplier tubes are used as a gamma ray sensing and light measuring devices, respectively. The dose rate of Cs-137 used in this study is 6 μSv·h−1. For the machine learning modeling, Keras framework in a Python environment is used. The algorithm chosen to construct machine learning model is regression with 15,000 number of nodes in each hidden layer. The pulse-shaped signals measured by photomultiplier tubes are saved as discrete digits and each pulse data consists of 1,024 number of them. Measurements are conducted separately to create machine learning data used in training and test processes. Measurement times were different for obtaining training and test data which were 1 minute and 5 seconds, respectively. It is because sufficient number of data are needed in case of training data, while the measurement time of test data implies the actual measuring time. The machine learning model is designated to estimate the source positions using the information about time difference of the pulses which are created simultaneously by the interaction of gamma ray and plastic scintillating fiber sensor. To evaluate whether the double-trained machine learning model shows enhancement in accuracy of source position estimation, the reference model is constructed using training data with one-time learning process. The double-trained machine learning model is designed to construct first model and create a second training data using the training error and predetermined coefficient. The second training data are used to construct a final model. Both reference model and double-trained models constructed with different coefficients are evaluated with test data. The evaluation result shows that the average values calculated for all measured position in each model are different from 7.21 to 1.44 cm. As a result, by constructing the double-trained machine learning model, the final accuracy shows 80% of improvement ratio. Further study will be conducted to evaluate whether the double-trained machine learning model is applicable to other data obtained from measurement of gamma ray sources with different energy and set a methodology to find optimal coefficient.

Due to changes in the agricultural market environment and both overseas and domestic farming conditions, uncertainties in agricultural production and management are becoming greater. Hence, there is a stronger need for farmers to choose crops in the optimal condition. This research aims to introduce the result and process of developing a decision support system for selecting crops, aimed to assist farmers in selecting the optimal crops most suitable in the given situation. There are basically three main factors to consider in the decision-making process for farmers when selecting a crop to introduce to their lands. First of all, one must consider how much profit crop A will produce when it is cultivated. Secondly, one must consider which crop to cultivate in order to earn a certain amount of profit. Thirdly, one must consider what is the best way to maximize Farm A’s business profit. For instance, a farm may have land as its resource, and one must research which location, type of crop, level of technology, and so forth, to maximize profit.This research creates a database of the profitability of a total of 180 crop types by analyzing Rural Development Administration’s survey of agricultural products income of 115 crop types, small land profitability index survey of 53 crop types, and Statistics Korea’s survey of production costs of 12 crop types. Furthermore, this research presents the result and developmental process of a web-based crop introduction decision support system that provides overseas cases of new crop introduction support programs, as well as databases of outstanding business success cases of each crop type researched by agricultural institutions.