• Journal of the Korea Society of Computer and Information
• /
• v.27 no.9
• /
• pp.13-20
• /
• 2022

It's proposed and analyzed ML(Machine Learning) models to predict vehicle FC(Fuel Consumption) in real-time. The test driving was done for a car to measure vehicle speed, acceleration, road gradient and FC for training dataset. The various ML models were trained with feature data of speed, acceleration and road-gradient for target FC. There are two kind of ML models and one is regression type of linear regression and k-nearest neighbors regression and the other is classification type of k-nearest neighbors classifier, logistic regression, decision tree, random forest and gradient boosting in the study. The prediction accuracy is low in range of 0.5 ~ 0.6 for real-time FC and the classification type is more accurate than the regression ones. The prediction error for total FC has very low value of about 0.2 ~ 2.0% and regression models are more accurate than classification ones. It's for the coefficient of determination (R²) of accuracy score distributing predicted values along mean of targets as the coefficient decreases. Therefore regression models are good for total FC and classification ones are proper for real-time FC prediction.

• Journal of Korean Society of Forest Science
• /
• v.106 no.3
• /
• pp.275-287
• /
• 2017

The purpose of this study was to analyze forest vegetation type classification and relationships between the type and environment factor in Mt. Dutasan. Data were collected by total of forty six plots using Z-M phytosociological method from June to October, 2016, and analyzed by vegetation classification, canopy layer structure and relationships between vegetation unit and environment factor using coincidence methods. As a result of vegetation type classification, Quercus mongolica community group was classified at a top level of vegetation hierarchy that was classified into Rhododendron schlippenbachii community and Betula costata community. R. schlippenbachii community was divided into Lychnis cognata group and R. schlippenbachii typical group. L. cognata group was subdivided into Veratrum oxysepalum subgroup and L. cognata typical subgroup. B. costata community was divided into Fraxinus mandshurica group and Betula schmidtii group. F. mandshurica group was subdivided into Weigela subsessilis subgroup and Cimicifuga heracleifolia subgroup. Therefore the forest vegetation was composed of six vegetation units with two kinds of bisected species groups and fourteen species groups. As the result of an analysis of canopy layer structure, there were two kinds of structures with monotonous structures V. oxysepalum subgroup (vegetation units 1), L. cognata typical subgroup (vegetation units 2), W. subsessilis subgroup (vegetation units 4) and complicated structures R. schlippenbachii typical group (vegetation units 3), C. heracleifolia subgroup (vegetation units 5), Betula schmidtii group (vegetation units 6). The vertical layer structure of vegetation unit 5 was the most developed and vegetation unit 6 had the lowest coverage of herb layer. According to the correlation between vegetation unit and environmental factor, R. schlippenbachii community (vegetation units 1~3) and B. costata community (vegetation units 4~6) were classified based on 1,100 m of altitude, middle slope, twenty of slope degree, twenty percents of bare rock and thirty centimeters of DBH in tree layer. R. schlippenbachii community (vegetation units 1~3) showed positive correlation with altitude, topography and B. costata community (vegetation units 4~6) showed negative correlation tendency with them.

• Journal of Korean Society of Forest Science
• /
• v.109 no.3
• /
• pp.249-258
• /
• 2020

The purpose of this study examined the succession by vegetation type after forest ecosystem restoration in the Baramjae area. Vegetation of the Baramjae area was classified using a survey of 81 sites from May to October 2019. The vegetation type was classified as Pinus densiflora community group with both Quercus mongolica community and P. densiflora typical community. The group unit was further classified as the Quercus dentata typical subgroup, Salix koreensis subgroup, and Q. mongolica typical subgroup. Such as Q. mongolica, Quercus variavilis in vegetation unit 1, Q. mongolica, Q. dentata in vegetation unit 2, P. densiflora in vegetation unit 3 and S. koreensis in vegetation unit 4 were shown a high importance value. The difference in species by vertical layer is explained by sere. Based on the vegetation type classification system, Detrended Correspondence Analysis was conducted to observe the trend of succession. Since restoration, vegetation unit 1 and vegetation unit 2 were considered to have developed the most extensive vegetation. In vegetation unit 2 and vegetation unit 4, many of the species found were in the early vegetation development in S. koreensis subgroup. Accordingly, vegetation in the Baramjae area can be categorized as a stepwise succession.

• Journal of Korean Society of Forest Science
• /
• v.79 no.2
• /
• pp.162-172
• /
• 1990

Habit type classification based on the polyclimax and discontinuum concept applied to Mt. Kasan forest. The results obtained were summirized as follows. 1. PIDE/RHSC H.T. : Pizzas densiflora is a topographic climax which is distributed in a ridge and a upper slope. Rhododeudron schlippenbachii is well represented in the undergrowth (average coverage 21%). 2. CALA/CAAR H.T. : Carpinus laxiflora is distributed in a low slope and adjoined to Pinzrs densiflora of a upper slope and a ridge. Calamaglotis aruredinacea is well represented in the undergrowth (average coverage 10%). 3. QUMO/CAHU H.T. : Quercus mongolica is a topographic climax which is distributed in a upper slope and a midslope of the Kasan fort outside and ranged from 600m to 700m, Carex humilis is well represented in the undergrowth (average coverage 21%). 4. QUMO/CAAR H.T. : Qreercus moragolica is distributed in the basin of Mt. Kasan fort inside and ranged from 700m to 870m, Cdlamagrotis arundinacea is abundant in the undergrowth(average coverage 43%). 5. FRRH/IMTE H.T. : Fraxinus rhynchophilla is distributed in a valley. Impatiens textori is well represented in the undergrowth(average coveraged 10%). 6. COCO/VISA H.T. : Cornus controversa is distributed in a valley and aranged from 780m to 830m Viburunum sargentii is well represented in the undergrowth(average coverage 12%).

• Korean Journal of Remote Sensing
• /
• v.22 no.6
• /
• pp.553-563
• /
• 2006

This study aims to produce land-cover maps of Korean peninsula using multi-temporal MODIS (Moderate Resolution Imaging Spectroradiometer) imagery. To solve the low spatial resolution of MODIS data and enhance classification accuracy, Linear Spectral Mixture Analysis (LSMA) was employed. LSMA allowed to determine the fraction of each surface type in a pixel and develop vegetation, soil and water fraction images. To eliminate clouds, MVC (Maximum Value Composite) was utilized for vegetation fraction and MinVC (Minimum Value Composite) for soil fraction image respectively. With these images, using ISODATA unsupervised classifier, southern part of Korean peninsula was classified to low and mid level land-cover classes. The results showed that vegetation and soil fraction images reflected phenological characteristics of Korean peninsula. Paddy fields and forest could be easily detected in spring and summer data of the entire peninsula and arable land in North Korea. Secondly, in low level land-cover classification, overall accuracy was 79.94% and Kappa value was 0.70. Classification accuracy of forest (88.12%) and paddy field (85.45%) was higher than that of barren land (60.71%) and grassland (57.14%). In midlevel classification, forest class was sub-divided into deciduous and conifers and field class was sub-divided into paddy and field classes. In mid level, overall accuracy was 82.02% and Kappa value was 0.6986. Classification accuracy of deciduous (86.96%) and paddy (85.38%) were higher than that of conifers (62.50%) and field (77.08%).

• Journal of Korean Society of Forest Science
• /
• v.112 no.1
• /
• pp.11-22
• /
• 2023

This study was established to provide basic information necessary for ecological management to restore the naturalness of black locust (Robinia pseudoacacia L.) plantations located in the mountains of Gyeongbuk, Korea. Using vegetation data collected from 200 black locust stands, vegetation types were classified using the TWINSPAN method, the spatial arrangement status according to the environmental gradient was identified through DCA analysis, and a synoptic table of communities was prepared based on the diagnostic species determined by determining community fidelity (Φ) for each vegetation type. The vegetation types were classified into seven types, namely, Quercus mongolica-Polygonatum odoratum var. pluriflorum type, Castanea crenata-Smilax china type, Clematis apiifolia-Lonicera japonica type, Rosa multiflora-Artemisia indica type, Quercus variabilis-Lindera glauca type, Ulmus parvifolia-Celtis sinensis type, and Prunus padus-Celastrus flagellaris type. These types usually reflected differences in complex factors such as altitude, moisture regime, successional stage, and disturbance regime. The mean relative importance value of the constituent species was highest for black locust(39.7), but oaks such as Quercus variabilis, Q. serrata, Q. mongolica, Q. acutissima, and Q. aliena were also identified as important constituent species with high relative importance values, indicating their potential for successional trends. In addition, the total percent cover of constituent species by vegetation type, life form composition, species diversity index, and indicator species were compared.

• Animal Systematics, Evolution and Diversity
• /
• v.31 no.3
• /
• pp.160-175
• /
• 2015

The echolocation calls of bats can provide useful information about species that are generally difficult to observe in the field. In many cases characteristics of call structure can be used to identify species and also to obtain information about aspects of the bat's ecology. We describe and compare the echolocation call structure of 14 of the 21 bat species found in Korea, for most of which the ecology and behavior are poorly understood. In total, 1,129 pulses were analyzed from 93 echolocation call sequences of 14 species. Analyzed pulses could be classified into three types according to the pulse shape: FM/CF/FM type, FM type and FM/QCF type. Pulse structures of all species were consistent with previous studies, although geographic variation may be indicated in some species. Overall classification rate provided by the canonical discriminant analysis was relatively low. Especially in the genera Myotis and Murina, there are large overlaps in spectral and temporal parameters between species. On the other hand, classification rates for the FM/QCF type species were relatively high. The results show that acoustic monitoring could be a powerful tool for assessing bat activity and distribution in Korea, at least for FM/QCF and FM/CF/FM species.

• Korean Journal of Environment and Ecology
• /
• v.32 no.3
• /
• pp.289-302
• /
• 2018

The study investigated the forest vegetation in 59 plots between June 2017 and August 2017 to understand the forest vegetation structure of the protected zone for forest genetic resource conservation (forest genetic resource reserve area) in Heuksando Island. We classified the vegetation using the Z-M phytosociological method analyzed the importance value and species diversity of each vegetation classification. The analysis showed the Camellia japonica community group at a top level of forest vegetation hierarchy. In the level of community, it was classified into Dendropanax morbiferus community (Vegetation unit 1; VU 1), Carpinus turczaninowii community, and C. japonica typical community (VU 6). C. turczaninowii community was subdivided into Buxus koreana group (VU 2), Rhododendron mucronulatum group (VU 3), Vitis amurensis group (VU 4) and C. turczaninowii typical group (VU 5). Therefore, it was classified into a total of six vegetation units (one community group, three communities, and four groups). The analysis of the mean codominant value of each VU show that Quercus acuta was the highest in VU 1, C. turczaninowii in VU 2, Pinus thunbergii in VU 3, Pinus densiflora in VU 4, and Castanopsis sieboldii in VU 5 and VU 6. The analysis of species diversity showed that VU 2 was the highest among six units in species richness index, species diversity index, and species evenness index. VU 6 showed the highest among six units in species dominance index. In conclusion, a synecology approach to manage six units and twelve species groups was needed for the forest vegetation of Heuksando Island protected area for forest genetic resource conservation.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2022.06a
• /
• pp.106-107
• /
• 2022

The ship type identification of the surrounding ship is important information for navigators and VTS officers since they can estimate the maneuverability and near-future route of the ships. However, it is more than frequent that the information is not provided due to transmission trouble and seafarers' unfamiliarity with AIS. Thus, this study suggests predicting ship types through the Random Forest classifier after preparing a training and test dataset that contains ship features and types. The AIS data for Ulsan coast in 2018 was used for this study. The method may provide the effect that many navigators and VTS officers discuss and share the experience of predicting ship types.

• Journal of agriculture & life science
• /
• v.43 no.2
• /
• pp.1-8
• /
• 2009

The limestone area covers the narrow range of the Korean Peninsula, and it consists of a peculiar ecosystem and topography. Therefore, this study on limestone area was carried out classification of vegetation type, physicochemical properties of forest soils and correlation between environment factors and vegetation types in order to furnish fundamental data for the forest management of limestone area. Forest vegetation was classified into two community groups such as Quercus variabilis community group and Quercus mongolica community group, and it was classified into eight vegetation units. Soil texture of survey sites showed largely silt loam and soil pH indicated the value of mean 7.55 in the A layer of soil profile. Content of exchangeable cation such as calcium ion ($Ca^{2+}$) and magnesium ion ($Mg^{2+}$) showed $26.04cmol_{c}/kg$ and $2.93cmol_{c}/kg$, respectively, which was about ten times higher than average of other regions of Korea. According to corelation between environmental factors and vegetation units, Q. variabilis community group was positively correlated to soil pH, slope degree and the rate of bare rock, and content of calcium ion ($Ca^{2+}$), and then Q. mongolica community group was positively correlated to altitude, respectively.