The purpose of this study is to propose an approach to fuse multiscale charts into three-dimensional (3D) electronic navigational chart (ENC) systems based on underwater topography and remote sensing image. This is the first time that the fusion of multiscale standard ENCs in the 3D ENC system has been studied. First, a view-dependent visualization technology is presented for the determination of the display condition of a chart. Second, a map sheet processing method is described for dealing with the map sheet splice problem. A process order called “3D order” is designed to adapt to the characteristics of the chart. A map sheet clipping process is described to deal with the overlap between the adjacent map sheets. And our strategy for map sheet splice is proposed. Third, the rendering method for ENC objects in the 3D ENC system is introduced. Fourth, our picking-up method for ENC objects is proposed. Finally, we implement the above methods in our system: automotive intelligent chart (AIC) 3D electronic chart display and information systems (ECDIS). And our method can handle the fusion problem well.
Since the 1960s, when “geographic information system (GIS)” was issued, the relative theories and technologies, including spatial data processing and management, spatial data service, web GIS, and 3D GIS, have been greatly developed [
With the development of computer science and sensor technology, spatial information can be conveniently obtained in various ways. Up to now, the most advanced and efficient way is the air-to-ground observation (remote sensing). It has many advantages such as large perception range, short repetition period, and high resolution[
Remote sensing image has been widely used in surveying and mapping, navigation, resource survey, meteorology, and so on. Researchers have committed to the combination of GIS and remote sensing image, especially the use of remote sensing images in 3D GIS systems. Numerous methods have been proposed to process and visualize remote sensing images in GIS systems [
There is a huge potential of applications in reach, if we can take advantage of the 3D technology and remote sensing image in marine GIS systems, especially in ENC systems. Many researchers in this area have researched 3D charts to provide chart display systems, aimed to reduce the amount of marine accidents caused by fatigue, mental overload, and limited awareness of the navigational situation. Porathe and Sivertun [
However, most of the researches listed above mainly focus on the description of overwater information. Few researchers study how to fuse multiscale charts into 3D ENC systems based on underwater topography and remote sensing image. The ENC is critical to marine GIS systems. It is an abstraction and generalization of the real world, which focuses on sea data, with less emphasis on land information. The remote sensing image contains rich land information. And the 3D underwater topography can provide more direct and intuitive sounding information. So, these data should be fused together to provide a more visually efficient ENC system.
In this paper, we propose an approach to fuse multiscale charts into 3D ENC systems based on underwater topography and remote sensing image. Our fusion approach is mainly divided into three parts: view-dependent visualization, map sheet processing, and object rendering. Then, we introduce the picking up method for information inquiry. At last, we implement our method in our system AIC 3D ECDIS. The application shows the effectiveness of our approach.
In order to provide more rich and intuitive geospatial information for navigation, we use the remote sensing image to describe the overland information and use the 3D underwater topography to directly describe the sounding data. Artificial water surface objects and entity objects are extracted from the ENC to fuse with them. Artificial water surface objects, such as depth contours, coastlines, depth areas, anchor zones, and prohibited areas, are organized into point, line, and polygon object groups. Entity objects, such as beacon lights, bridges, and port structures, are represented by 3D models. Other objects (e.g., land areas) in the ENC can be ignored. In this paper, we propose an approach to fuse these data together. Fusion effects are shown in Figures
However, the ENC is based on the scheme of subdivision and multiscale. As can be seen from Figure handle the multiscale problem; solve the map sheet splice problem; fuse multiscale charts with the remote sensing image; fuse multiscale charts with the underwater topography.
Based on the above analysis, we propose a method to solve the problems listed above. The method consists of three parts: view-dependent visualization, map sheet processing, and object rendering. Firstly, use the view-dependent visualization technology to display the charts according to their scales. Secondly, use the map sheet processing method to deal with the map sheet splice problem. There are two main steps in this process: map sheet clipping and map sheet splice. Thirdly, use the rendering method to solve the last two problems.
Electronic navigational chart.
Essentially, the 3D ENC system is a 3D computer graphics system, which is based on the 3D rendering pipeline. In order to display multiscale charts in 3D ENC systems, we should deal with the scale problem in the viewpoint observation system. Here, we use view-dependent visualization technology to display multiscale charts; that is, the display process of charts is determined by the change in viewpoint. The detailed steps of the view-dependent visualization are described as follows. Calculate the height Calculate the current display scale ds. ds is calculated by the following equation: Display the chart according to the current display scale and its drawing scale. According to the display requirements of S-52 standard, the ENC should be displayed in its drawing scale, and it can be enlarged or reduced when needed. Zoom factor can be 0.5 to 3 times the drawing scale. However, the ENC can be limitlessly enlarged when needed. Thus, we mainly consider the reduced state. In this step, we design a display condition for each chart. To be specific, if the display scale is larger than 1 in 2 of the drawing scale of the chart, the chart should be displayed; otherwise, the chart should not be displayed. The larger scale chart is prior to the smaller scale chart, that is, the larger scale chart is displayed above the smaller scale chart.
In order to directly and intuitively display underwater topography, artificial water surface objects such as depth contours, coastlines, depth areas, anchor zones, and prohibited areas are displayed in semitransparent mode. However, the multiscale charts overlap and interweave. This will seriously affect the display effect. We propose a method to deal with this problem. Our method of map sheet processing is a dynamic process based on the change in viewpoint. There are three key points in our method: 3D order, map sheet clipping, and map sheet splice. The detailed steps of this process are shown in Figure
The map sheet processing.
In order to get the right result in the map sheet processing, the charts should be processed in order. We design a process order called “3D order” to adapt to the characteristics of the chart. In the third step of the map sheet processing, the “3D order” is shown in Figure
The schematic diagram for 3D order.
In order to facilitate description, here we give a schematic diagram for multiscale charts in Figure
The schematic diagram for multiscale charts.
Map sheet clipping is used to deal with the overlap between the adjacent map sheets. When two charts overlap, we should clip the other map sheet based on one map sheet in 3D order described above and update the geographic scope of each map sheet. Figure
The schematic diagram for the map sheet clipping.
This operation can be divided into two parts: map frame clipping and object clipping. The purpose of object clipping is to use the geographic scope of one chart to clip artificial water surface objects and entity objects in another chart. The geographic scope of the chart (map frame), whether regular or irregular, is a polygon. Artificial water surface objects and entity objects are point, line, or polygon objects in the chart. So, map frame clipping can be treated as a special case of object clipping. Here, we mainly describe the object clipping specifically. For a point object, we just need to judge if the object lies inside the geographic scope of the reference chart and then do the clip. For line and polygon objects, the clipping process is shown in Figure
The clipping process of line and polygon objects.
Overwater view effect of the fusion.
Water surface view effect of the fusion.
Underwater view effect of the fusion.
It should be noted that, according to the data transfer standard S-57, line and polygon objects in ENC are segmented, and each part is marked with ordinal number, direction, display attribute, and so on. We sort the boundary points of each object into counterclockwise order to simplify the following calculations and then process each segment in counterclockwise order.
Suppose segment
In this clipping process, the intersection calculation and the clip operation are based on the counterclockwise order. And the calculation of each intersection point is followed by a clip operation. Boundary points that lie inside the frame of the reference chart should be removed. Frame points between two intersection points should be added to the boundary point set of the object in counterclockwise order. When all intersection calculations are finished, the object is also clipped.
Map sheet splice is used to deal with the splice point after the map sheet clipping. Splice points of the same object in two adjacent charts may have different coordinates. To handle this problem, there are three commonly used methods: enforcement method, averaging method, and optimization method. For two splice points, enforcement method directly assigns coordinates of one point to the other point. This method is mainly used in the case of knowing which point is more accurate and reliable. Averaging method uses the average values of the coordinates of the two points to replace their coordinates. Optimization method is relatively complicated and needs lots of intersection calculations.
In order to improve the efficiency of the map sheet splice, we adopt the following strategy. Firstly, compare the scale of the two adjacent charts. Secondly, choose the splice method according to the comparison result. If the two charts have the same scale, we use the averaging method to deal with the splice points. If one chart has larger scale than the other one, we choose the enforcement method. The reason is because larger scale chart is more accurate and reliable.
There are two classes of ENC objects we use: artificial water surface object and entity object. Entity objects are represented by 3D models. Artificial water surface objects are organized into point, line, and polygon object groups. Essentially, artificial water surface objects are vector data. Most methods map the vector data to the terrain in 3D GIS systems [
side length of the minimum bounding box of the object, measured in meters (m); side length of the minimum bounding box of the object, measured in meters (m); side length of the minimum bounding box of the object, measured in meters (m); field-of-view angle, measured in radians (rad); width of the viewport, measured in pixels (px); distance between the viewpoint and the center of the minimum bounding box of the object, measured in meters (m).
For entity objects, we use the distance between the viewpoint and the center of the minimum bounding box of the entity model to determine whether the model should be displayed. This method is also used in the rendering of the point object.
In order to query the information of ENC objects, we need to solve a problem first: pick up the object in the 3D ENC system. There are two classes of ENC objects that we use: artificial water surface object and entity object. For entity objects, we use the ray tracing method [ Convert the pick-up point to a pick-up range. The pick-up point is a screen pixel point. We extend several pixels in the Clear the color buffer of the pick-up range. This operation is executed in the back buffer, so it does not affect users’ visual effect. Then, draw all objects with specified picking up flag colors instead of the original colors. The same object uses the same picking up flag color. This process is also executed in the back buffer. At last, read the value of the color buffer of the pick-up range to judge whether an object should be picked up. After all objects are judged, we can get a list of picked up objects. It should be noted that objects may overlap. So, we have to use the final value of the color buffer of the pick-up range to make the judgment; that is, the judgment must be made after the rendering of all objects is finished.
This method can avoid complicated spatial intersection calculations. It is very efficient.
We implement our methods in our system AIC 3D ECDIS [
There are two ways of implementing our method. One is to use the method on the server side and publish the data to the client side. The other is to use the method in the client program. In order to simplify the client’s workload, we choose to implement the method on the server side. When the client program needs data, it sends the request, which includes state parameters, such as viewpoint, viewport, world reference point, geographic scope, and display scale, to the server side. On the server side, spatial analysis servlet components and spatial calculation servlet components are responsible for the map sheet processing. These servlet components receive parameters from data service module and store the processed data in the cache memory. Then, the data service module publishes the data to the client. The client program receives the data and uses the view-dependent visualization technology and the rendering method to visualize the data. Here, we give the fusion effects in the client program. The fusion effects are shown in Figures
This paper provides a solution for the fusion of multiscale charts in 3D ENC systems based on underwater topography and remote sensing image. It is the first time that this subject has been studied.
The remote sensing image, normally used in 3D GIS, is extended to enhance the understanding of the land environments in 3D ENC systems. The underwater topography gives an intuitive cognition of the underwater depth. The ENC data has important applications in the analysis and management of overwater spatial environments. Combination of these data helps in understanding the navigation environment. Our method based on the characteristics of the ENC is proposed to fuse these data together in the 3D ENC system. And we present the picking up method for ENC objects in the 3D ENC system.
At present, we implement the method on the server side. Further work has to be done in terms of data service, for example, server side concurrency calculation.
The authors declare that there is no conflict of interests regarding the publication of this paper.
This work was supported by the Fundamental Research Funds for the Central Universities (Grant no. 3132013303).