KR20110008947A - Extraction method for depth of water from ecdis to estimate topography of seafloor - Google Patents

Abstract

PURPOSE: A water depth information extraction method for extracting an electronic navigation chart image is provided to improve the reliability of a collision avoidance system by extracting a submarine topography analogized based on the water depth information. CONSTITUTION: A binary image extracts a predetermined color region from a captured electronic navigation chart image(S100). The text raw section in which the information of water depth is extracted from the binarized image(S200). The boundary of exacting about the character area separates from the extracted character area(S300). A real data value obtains through the letter recognition process of the discrete character image. The submarine topography obtains the coordinates of the character string combination generation(S400).

Description

Extraction method for depth of water from ECDIS to estimate topography of seafloor}

The present invention relates to a method for extracting depth information of an electronic chart for inferring the seabed topography, and more particularly, extracts depth information, which is one of decision factors for collision avoidance, in an collision prevention system of a ship. In addition, the present invention relates to a method for extracting depth information of an electronic chart which stores extracted depth information and coordinates and infers a terrain of a seabed.

At present, about 80% of accidents in Korea are caused by human error, and 70% of all accidents except for fishing boats are caused by collision and grounding. Therefore, in order to prevent marine accidents and enhance the safety of navigation, simulation functions are provided at the time of entering and leaving the port of the ship, when navigating a new port or route, not only to predict the course of the ship but also to prevent accidents such as stranding or collision. There is an urgent need to develop preventive programs.

Recently, various simulations have been developed to improve ship performance. However, it provides essential functions such as the ability to simulate arbitrary routes, the ability to test in conjunction with an Autopilot system, and the ability to receive data during actual navigation and simulate the actual navigation situation. The program is not being developed.

In order to develop a program that satisfies all of the above functions, essential data are required, and one of them is depth data for identifying a seabed topography, that is, depth information.

However, although the current depth can be read using the sounding equipment, the depth information of the surrounding seas can only be obtained by the operator interpreting the paper chart.

Therefore, it was not possible to develop a program with the functions as the depth information of the surrounding sea area could not be obtained, and there was a problem that more sea accidents occurred due to the inability to grasp the surrounding seabed topography.

The present invention has been made to solve the above-mentioned conventional problems, to meet the requirements,

An object of the present invention is to quickly and accurately determine the characters representing the depth in the electronic chart image to extract depth information, and to store the extracted depth information and coordinates to infer the topography of the seabed to improve the reliability of the collision avoidance system of the ship The present invention provides a method for extracting depth information of an electronic chart for inferring the seabed topography.

In order to achieve the above object, according to the present invention, a method for extracting depth information of an electronic chart for inferring the seabed topography includes: a color segmentation step of extracting a specific color region from a captured electronic chart image to obtain a binary image; A character region extraction step of classifying a character string region having depth information in the binarized image; A string separation step of obtaining and separating an accurate boundary for each character area from the extracted character area; Characterizing the individual character image obtained in the character string separation step to obtain an actual data value, and obtaining the center coordinates of the character string set to estimate the seabed topography.

The color dividing step converts the RGB color space of the electronic chart image into an HSI color space, obtains a HUE channel image, and labels the obtained HUE channel image based on color to have pixels having the same label. And binarize the data into sets.

The extracting of the text area finds the position of the string set representing the depth information by filtering based on the size of each masked area in the resultant image of the open operation in which the erosion and expansion operations are sequentially applied using a rectangular structural element. My extract is characterized by.

In the character string separation step, the vertical and horizontal profiles of the extracted text area are obtained, and after the vertical and horizontal alignment, the white and white pixels are not detected in the vertical and horizontal profiles. It is characterized by separating.

The method of obtaining the actual value of the separated character image in the step of estimating the seabed terrain is to create a feature vector by making the individual separated character image as a vector, find a minimum distance from the sample vector consisting of 0 to 9, and It is characterized by finding the actual value.

As described above, according to the method for extracting the depth information of the electronic chart for inferring the seabed topography of the present invention, it is possible to quickly and accurately identify and extract information corresponding to the depth from various elements represented on the electronic chart, and the depth information By extracting the seabed topographical information from the electronic chart image using the system, it is possible to simulate the arbitrary route, the ability to test in conjunction with the Autopilot system, and receive the actual sailing data to receive the actual sailing situation. It has the effect of providing the necessary data needed to develop programs that provide essential functions, such as the ability to simulate.

In addition, since it can be used as additional data in determining the route, there is an effect of improving the reliability of the collision avoidance system of the ship.

Hereinafter, with reference to the accompanying drawings will be described in detail a preferred embodiment of the method for extracting the depth information of the electronic chart for inferring the seabed terrain according to the present invention.

1 is a flowchart illustrating a step of extracting depth information of an electronic chart for inferring the seabed terrain according to the present invention, and FIG. 2 is an electronic chart in a method for extracting depth information of an electronic chart for inferring the seabed terrain according to the present invention. FIG. 3 is a diagram illustrating a color splitting process of extracting a specific color region from the image of the electronic chart of FIG. 2 to subdivide colors, and FIG. 4 includes depth division information of FIG. 3. FIG. 5 is a diagram illustrating a process of extracting a string region from the image of FIG. 4, and FIG. 6 is a diagram illustrating a process of separating the extracted string region into individual characters. 7 is a diagram showing the results of recognizing the depth information, respectively.

As shown in the figure, in order to extract the seabed topographical information from the image, it is necessary to first extract the depth information.

Here, the depth information extraction is based on the standard format for the electronic chart representation described in International Hydrographic Organization (IHO) S-52, "Color & Symbol Specification For ECDIS".

As shown in FIG. 1, the depth information extraction is largely performed on the individual character image obtained in the color segmentation step S100, the character area extraction step S200, the character string separation step S300, and the character string separation step S300. By performing a character recognition process by applying a character recognition algorithm to obtain the actual data value and to obtain the center coordinates of the string set to estimate the seabed topography (S400).

In the color dividing step S100, after converting an RGB color space into an HSI color space for a color representing depth data displayed on the electronic chart image of FIG. 2, a Hue channel image is obtained. Hue) The channel image is labeled based on color and binarized by dividing pixels having the same label into a set. Therefore, the results obtained for the three labels can be confirmed as shown in FIG. 3, and the color of the general characters is the same in addition to the longitude line and the depth information in the electronic chart (ECDIS) used as shown in FIG. 4. The algorithm that removes lines is then applied to obtain a binarized image that finally contains depth information in the segmentation result.

In the character region extracting step (S200), after obtaining the binarization image as shown in FIG. 5, size filtering is performed after masking the string region using a morphology operation to obtain an accurate depth information string set. Go through the process. Specifically, each masking in the image of the result of the opening operation is applied by sequentially applying the erode operation and the dilation operation using a 5 by 5 rectangular structural element. Filtering is based on the size of the specified region to distinguish whether or not a string is set. That is, by obtaining information such as the maximum height, the maximum width, the width of the masked area, and filtering based on a boundary value corresponding to the depth information string, the position of the string set representing the depth information can be finally found and extracted.

As shown in FIG. 6, the character string separating step S300 is a step of separating the character strings by unit characters in order to recognize the characters based on the extracted character region. In other words, the vertical and horizontal profiles of the selected character string area are obtained, and then vertically and horizontally aligned. Then, the position at which the white pixel is not detected in the vertical and horizontal profiles is set as the reference coordinate, and the individual characters are considered in consideration of the width of the character. To separate it.

Each of the above steps uses digital image processing-related algorithms as a series of processes for efficiently separating characters from images to enable individual character recognition in the electronic chart image. Using the depth information obtained through the above process, it is possible to extract the seabed topographic information from the electronic chart image. It can also be used as additional data in route determination, which can improve the reliability of the ship's collision avoidance system.

The estimating of the seabed terrain (S400) obtains the actual values of the individual characters separated in the character string separating step (S300), recognizes them on the image, obtains the center coordinates of the string set, and stores them together with the recognized depth information. Infer the terrain.

In the step of estimating the seabed topography (S400), a method of obtaining an actual value of a separated text image includes generating a feature vector by using individual separated text images as a vector, and then forming a feature vector. Find the minimum distance from the vector) to find the actual value. That is, a character of the depth data is formed into a 5 by 7 matrix, converted into a 35-dimensional feature vector, and the sample vector having the minimum value is selected by adding the squares of the difference between the obtained feature vector and the sample vector.

Therefore, as shown in FIG. 7, after the depth information is recognized on the image, the center coordinates of the string set are obtained and stored together with the recognized depth information, which is a seed point for estimating the seabed topography. Based on these origins, the seabed topography can be inferred by interpolating and estimating the area between each origin.

As described above, the present invention extracts depth information through a fast and efficient image processing process in an electronic island system in an environment where a user cannot obtain depth information, and estimates the seabed topography through the recognition process of the extracted depth information. Essential functions such as the ability to simulate arbitrary routes, the ability to test in conjunction with an Autopilot system, and the ability to receive real-time navigation data and simulate actual navigation conditions. You get the data you need to develop your programs.

Although the preferred embodiments of the present invention have been disclosed for illustrative purposes, those skilled in the art will appreciate that various modifications, additions and substitutions are possible, without departing from the scope and spirit of the invention as disclosed in the accompanying claims. Of course.

Therefore, the scope of the present invention should not be limited to the described embodiments, but should be defined by the claims below and equivalents thereof.

1 is a flow chart showing the depth information extraction step of the electronic chart for inferring the seabed terrain according to the present invention.

2 is a view showing an image of the electronic chart in the depth information extraction method of the electronic chart for inferring the seabed terrain according to the present invention.

3 is a diagram illustrating a color segmentation process of extracting a specific color region from the image of the electronic chart of FIG.

FIG. 4 is a diagram illustrating a binarization image including color separation result depth information of FIG. 3. FIG.

5 is a diagram illustrating a process of extracting a string region from the image of FIG. 4.

6 is a diagram illustrating a process of separating the extracted character string region into individual characters.

7 is a diagram illustrating a recognition result of depth information.

Explanation of symbols on the main parts of the drawings

S100: color division step

S200: Extract Character Area

S300: String Separation Step

S400: estimating the seabed terrain

Claims (5)

A color segmentation step of extracting a specific color region from the captured electronic chart image to obtain a binarized image; A character region extraction step of classifying a character string region having depth information in the binarized image; A string separation step of obtaining and separating an accurate boundary for each character area from the extracted character area; Performing a character recognition process on the individual character image obtained in the character string separating step to obtain a real data value, and obtaining a center coordinate of the character string set, and estimating the seabed topography. Depth information extraction method of electronic chart. The method of claim 1, The color dividing step converts the RGB color space of the electronic chart image into the HSI color space, obtains a HUE channel image, and then labels the obtained HUE channel image based on the color, thereby having pixels having the same label. Depth information extraction method of the electronic chart for inferring the seabed topography, characterized in that by dividing the data into sets. The method of claim 1, The extracting of the text area finds the location of the string set representing the depth information by filtering based on the size of each masked area in the resultant image of the open operation in which the erosion and expansion operations are sequentially applied using the rectangular component. Depth information extraction method of the electronic chart for inferring the seabed terrain, characterized in that extracted. The method of claim 1, In the character string separation step, the vertical and horizontal profiles of the extracted text area are obtained, and after the vertical and horizontal alignment, the white and white pixels are not detected in the vertical and horizontal profiles. Depth information extraction method of the electronic chart for inferring the seabed terrain characterized in that separated. The method of claim 1, The method for obtaining the actual value of the separated character image is characterized by finding the actual value by finding the minimum distance from the sample vector consisting of 0 to 9 after generating the feature vector by making the individual separated character image into a vector. Depth information extraction method of the electronic chart for inferring the seabed topography.

Images