The ENC manufactured by the standards of the International Hydrographic Organization(IHO, 2000) and includes location information such as water depth, sediment type, buildings, structures, buoys, rivers, and roads. The ENC separates levels according to spatial resolution, and the water depth and sediment type data of level 4, which is the highest resolution(scale 1:30,000-1:90,000), were used in this study. Fig. 1 and 2 show that the entire area of the used ENC includes the Korean Peninsula and annexed the island, ranging from longitude 124°15'36''-132°08'24'' E and latitude 32°54'36''-38°50'24''N on coordinate. Fig. 1 shows the merging results of the water depth for the collected ENC. Water depths in some coastal waters exhibit a dense distribution due to merging process of the ENC. Fig. 2 shows the sediment distribution from multiple ENC. It can be seen that the range or density of data is very insufficient compared to the water depth. Especially in the East Sea, data exist only in areas very close to the shoreline. In addition, there is no sediment data in the Mid-Yellow Sea region.

For the complex Korean coastal waters, the understanding of high-precision maritime environmental information is an essential element for generating optimal ship escape support information. Currently, available ENC provides information with excellent accuracy, but more high-resolution DB is needed to build to provide optimal escape support information in coastal waters. The KHOA regularly conducts high-precision measurement projects on selected areas and limitedly provides high-resolution water depth and sediment measurement data.

The KHOA is acquiring high-resolution water depth data by using a multi-beam echo sounder loaded on survey ships for coastal waters. Due to the limited access like security, the coordinates, and the water depth at the location provided after conducting a security examination on Mokpo, Jeju, Goheung, Wando, Geomundo, and some areas in the West Sea where many passenger ships operated. All the collected coordinates and water depth information was converted to a point-format shapefile using the ArcGIS program. Fig. 3 shows the high-resolution depth data collected by the KHOA observations. Compared to Fig. 1, it can be seen that the water depths of the South Sea area and north of Jeju Island densely distributed.

The KHOA collected sediment information in coastal waters through the air hyperspectral sensor(CASI-1500), diving, sample investigation, and ground-spectrum methods. In this study, based on 2018, Jumunjin-Imwon port(128°49' 48''-129°22'12'' E, 37°12'00''-37°54'36'' N), Uljin(129°20'24''- 129°30'36'' E, 36°39'00''-37°12'00'' N), Donghaesi-Busan port(129°20'24''-129°35'24'' E, 35°24'00''-36°39'00'' N), Busan (129°06'00''-129°10'48'' E, 35°04'48''-35°09'36'' N), Taean (126°15'36''-126°22'12'' E, 36°22'48''-36°40'12'' N), Goheung (127°06'00''-127°33'00'' E, 34°24'00''-34°43'48'' N), the sediment type of these six areas was collected. The sediment type classified into six categories, including pearl, pearl and sand mix, sand, gravel, rock, and submerged dike. Then we provided a shapefile of polygon properties. Also, in order to make areas distinction easier, we sorted layers and assigned colors differently by sediment. Fig. 4 shows the sediment type for six regions collected through observations. In particular, the sediment in Taean area was not included in ENC.

The Marine Managed Areas(MMAs) means a specific zone designated or controlled by the government for pollution, risk management, conservation of resources (Orbach and Karrer, 2010). The MMAs classified into particular management areas for marine pollution, marine training areas for marine risk, environment preservation area for resource, and farm preservation according to the purpose. Table 1 shows that the marine information platform of KHOA(www.khoa.go.kr/oceanmap), the Korea National Spatial Data Infrastructure Portal of the Ministry of Land, Infrastructure, and Transport(http://data.nsdi.go.kr/dataset), and the coastal portals of the Ministry of Maritime Affairs and Fisheries(http://coast.mof.go.kr) used for collecting spatial information of MMAs.

The ENC water depth has a low resolution but covers a large area. On the other hand, surveyed depths have very high resolution, but only in some areas. Besides, both data indicate the depth of the nonuniform grid. The purpose of this study is to construct a DB used to select a place of refuge for ships, and a grid of high-resolution data is efficient for operation and management. Also, the suitability of data capacity must be considered since it involves the purpose of providing real-time information.

In this study, in order to improve the accuracy of the designation of the ship refuge place, we separately constructed a DB of high-resolution depth for the coastal area where ship accidents occur frequently. The coastal waters are assumed to be 0-30m depths around the Korean peninsula. The spatial resolution of the depth of the coastal waters DB set at 50m, and the ENC and the KHOA data used. In the region where the two data overlap, data processing is performed to remove the ENC data. The data of the adjacent location in the surveyed water depth classified into the same object and the polygon generated, and the electronic chart water depth data included in the polygon removed. Fig. 5 shows that the remaining depths after the removal of the ENC and the surveyed depths were merged to create high-resolution depth DB. Compared with Fig. 1, it can be seen that the depth data of the south-west sea and the north sea of Jeju have changed.

ArcGIS Inverse Distance Weighted(IDW) Toolbox was used to generate 50m grid information. The IDW based on the assumption that the relevance and similarity between the values of a particular position and a surrounding position are inversely proportional to distance(Setianto and Triandini, 2013). In order to input the IDW, the depth data converted into the UTM coordinate system, and the number of neighborhoods included in the influence range for a selected point set to 15. Fig. 6 shows that the raster file produced using IDWwas separated using a polygon file of a 0-30m depth area. The separated raster file was saved as the shapefile of the point attribute and entered into the DB.

This section described the process of integrating the sediment of ENC and measured data. First of all, because the notation of the ENC and the measured data are different, we made it uniform using the notation name and number of the electronic chart, as shown in Table 2. In the electronic chart, because there is no number assigned, 'Submerged breakwater' is numbered as 100 in the DB. In generally, submerged breakwaters are not included in the sediment classification, but included because they affect the safety of ship navigation. In order to increase the calculation efficiency when calculating the ship support information, the grid of the integrated sediment and the depth information is set to be the same. The merged sediment information was generated by inputting in the same grid with the ENC and the measured location information. Furthermore, the sediment information at the 'NaN' grid generated by the spatial interpolation method with the 'Nearest' option. Fig. 7 shows the re-sampled sediment data in coastal water.

In order to select a refuge place of ship, it is necessary to set the area where escape is impossible to prevent secondary damage. On the other hand, information about the surrounding ports is needed to restore the vessel's and reduce the risk of navigation, and it may be considered a refuge place. Therefore, in this study, it was classified into a non-escape and escape candidate area considering the characteristics of MMAs, as shown in Table 1. The non-escape areas include fishing zones, environmental protection waters, military training waters, etc., and the escape candidate areas include port information. The spatial information classified into layers according to the classification criteria and the shapefile of the polygon attributes generated by using ArcGIS. Fig. 8 shows the classification result considering the characteristics of MMAs.

The built database can be used to extract a suitable place for the ship to stay in an emergency situation. We collected the AIS information of the collision accident between the Oriental Pearl II and the PILION ship at 06:28 on 05 January 2011 and defined it as an incident corresponding to escape level 1 by ‘crash’ regarding the contents of the written judgment. The navigation environment parameters included in a regular radius around the location of the target vessel was extracted and is used to calculate the candidate of grounding zone, contingency anchorage and exclusion area, respectively.

Fig. 9 is the output screen of the results of the calculation of the grounding zone, contingency anchorage, and the candidate exclusion area. The green and yellow circles indicate the possible location of the grounding zone and contingency anchorage, respectively. The blue polygon indicates the escape exclusion zone. If a candidate area for the grounding zone or contingency anchorage included in the escape exclusion zone, it will be changed to red and be output to the candidate exclusion areas. When the ship's location is changed, the information contained in the radius is automatically changed and updated. Then, the extracted navigation environment parameters used as input data to the escape route selection algorithm proposed in Park et al.(2018). This algorithm estimates the most suitable evacuation site considering the ship's condition and surrounding environmental factors. In the situation in Fig. 9, the grounding zone was guided to the excellent contingency place(red line), and the surrounding coast guard ship guided to the second contingency(yellow line).