KR101054025B1 - Visually impaired walking guidance method and system - Google Patents

Abstract

The present invention extracts the vanishing point and the road outline information based on the edge image of a single image obtained through a camera in an outdoor environment, and outputs the pedestrian's location and the direction of the pedestrian's road as voice information determined based on the extracted information. Therefore, the present invention relates to a visually impaired gait guide method and system for allowing a visually impaired person to walk safely.

To this end, the present invention comprises the steps of obtaining an image from a camera; Extracting vanishing point and outline information of a road using edge information detected in the acquired image; Determining a location of the pedestrian based on the outline information of the road, and determining a direction of the road seen by the pedestrian based on the vanishing point; And providing the converted voice information to the voice output unit by applying the determined position of the pedestrian and the direction of the road to predetermined walking guide information, respectively.

Visually impaired, walking, guidance, vanishing point, outline information, voice information, pedestrian position, direction of road

Description

METHOD FOR GUIDING A BLIND PERSON AND SYSTEM THEREFOR}

The present invention relates to a walking guidance method and a system for the visually impaired, in particular by selecting a gait area that can be visually impaired through the image obtained from the camera and by providing the voice information to guide information required for walking in the gait area The present invention relates to a method and a system for guiding a visually impaired person who can safely walk.

In general, people with visual impairments are very limited in walking because they are difficult to obtain information about the external environment, and it is difficult to properly respond to dangerous factors such as obstacles.

Currently, tool devices and systems for safe walking of visually impaired people are actively researched at home and abroad.

Conventional visually impaired gait guidance systems provided guidance information for assisting the visually impaired gait using ultrasound, tactile, laser, GPS, camera, and the like.

The visually impaired walking guidance system using ultrasonic waves is equipped with a bundle of ultrasonic sensors configured in the form of a belt, and analyzes information on each direction through a computer worn in the form of a bag and informs the visually impaired. Although the visually impaired gait guide system using the ultrasonic sensor has an advantage that the implementation is relatively simple and does not have a large amount of calculation, it can only determine the presence or absence of surrounding obstacles and cannot obtain information about the texture and color of the object. There are many disadvantages in obtaining the movement information of an object.

Another type of visually impaired walking guidance system using ultrasonic sensors is a cane-shaped road guidance system. The road guidance system is equipped with an ultrasonic sensor on a guide wand for general blind people to detect obstacles and guide the road in a non-hazardous direction. Such a road guidance system is simple in structure and easy to apply, but has a disadvantage in that the detection distance is limited because only a small object in front of it or an obstacle in proximity can be detected.

The visually impaired pedestrian guidance system using a laser can grab the laser sensor tool and shake it from side to side to obtain local information that exists in the surrounding area, and analyzes risk factors for movement such as stairs and cliffs in terms of time. There is an advantage in that there is a curve or a single camera can be used to obtain unknown distance information. However, it can be applied only to an adjacent area and has a disadvantage in that the color or texture of the detected object is unknown.

The visually impaired walking guidance system using GPS is a system for guiding a road to pedestrians by using empirically acquired and stored area map information, and provides a sequential path to a destination. This GPS-guided visual aid system has a database that contains local information, and it enables communication between computers and people through voice conversion technology and voice recognition technology, and provides a route to a destination in a specific area. Give convenience. However, there is a drawback that the risk of the immediate pedestrians cannot be analyzed.

The visually impaired walking guidance system using a camera is equipped with a headset, a microphone, and two cameras, and analyzes the front area and constructs three-dimensional information to detect an object in the front area. It is common to use a stereo camera to obtain the size and distance information of objects in the real world space through the camera. The visually impaired pedestrian guidance system using such a camera is mostly performed in a limited space such as a room, and is made in a simple background to obtain the characteristics of an object. In particular, an environment such as a corridor in a building is a very good environment for acquiring three-dimensional information.

As described above, the conventional visually impaired walking guidance system is not only difficult to apply in an environment having an external complex background, but also has a lot of limitations on information provided.

An object of the present invention is to provide visual information on the road so that the visually impaired can walk safely using a single camera even in an outdoor environment having complicated background information, so that the visually impaired person can walk safely in an outdoor environment. In order to provide a visually impaired walking guidance method and system.

In addition, another object of the present invention, by selecting a gait area in the image of the outdoor environment obtained from the camera and, if an obstacle existing in the gait area is detected, by indicating the location of the obstacle and the distance between the obstacle and the pedestrian, The present invention provides a method and a system for guiding a visually impaired pedestrian, which can prevent a collision accident.

In order to achieve the above object, a visually impaired walking guidance method according to an embodiment of the present invention includes: acquiring an image from a camera; Extracting vanishing point and outline information of a road using edge information detected in the acquired image; Determining a location of the pedestrian based on the outline information of the road, and determining a direction of the road seen by the pedestrian based on the vanishing point; And providing the converted voice information to the voice output unit by applying the determined position of the pedestrian and the direction of the road to predetermined walking guide information, respectively.

In accordance with another aspect of the present invention, there is provided a walking guide method for visually impaired, comprising: determining whether an obstacle is detected in an image in a walking area obtained through outline information of a road; When the obstacle is detected, generating a conversion table using coordinates of the image in the walking area for each predetermined distance; Estimating the position of the obstacle and the distance between the obstacle and the pedestrian based on the conversion table; And providing the converted voice information to the voice output unit by applying the position of the obstacle and the distance between the obstacle and the pedestrian to predetermined obstacle guide information.

The extracting the vanishing point and the outline information of the road may include converting the acquired image into a gray image; Detecting an edge image from the converted gray image; Finding connectivity for neighboring pixels in eight directions of each edge pixel in the detected edge image; Extracting linear components continuously related in a specific direction using the connectivity; Extracting a linear component directed toward the center in each area bisected based on the center of the acquired image; Selecting linear components of a predetermined length or more continuously from the extracted linear components; And taking the average of the selected linear components, extracting the intersection point of the average linear components in each region as a vanishing point, and extracting the average linear components in each region as the outline information of the road.

The determining may include selecting a walking area using the vanishing point and the outline information of the road; Calculating an intersection point of an average linear component in the lower end portion of the selected gait area and the respective areas as the width of the gait area, and calculating a center point in the width as the center of the road; Determining the location of the pedestrian by comparing the center of the road with the center of the image; And comparing the direction of the vanishing point and the center of the image to determine the direction of the pedestrian's view.

The determining of whether the obstacle is detected may include: extracting vertical components by applying an image in the gait area to a Gabor filter; And determining the bottom surface and the width information of the obstacle through the bundle of the extracted vertical components.

In the extracting the vanishing point and the outline information of the road, it is preferable to extract the linear information having a specific direction by using the extended chain code.

In addition, the walk guide method for the visually impaired according to another embodiment of the present invention, obtaining an image from the camera; Extracting vanishing point and outline information of a road using edge information detected in the acquired image; Determining a location of the pedestrian based on the outline information, and determining a direction of a road seen by the pedestrian based on the vanishing point; Determining whether an obstacle is detected among images in the walking area obtained through the outline information; When the obstacle is detected, generating a conversion table using coordinates of the image in the walking area for each predetermined distance; Estimating the position of the obstacle and the distance between the obstacle and the pedestrian based on the conversion table; And applying the determined position of the pedestrian and the direction of the road to predetermined pedestrian guidance information, and applying the estimated position of the obstacle and the distance between the obstacle and the pedestrian to predetermined obstacle guide information to convert the voice information. And providing the voice output unit.

The providing to the voice output unit may convert the completed sentence into voice information by applying the position of the pedestrian and the direction of the road to the predetermined walking guidance information, and convert the estimated sentence into the predetermined disability guide information. It is preferable to convert the completed sentence into voice information by applying the position of the obstacle and the distance between the obstacle and the pedestrian.

On the other hand, a visually impaired walking guidance system according to an embodiment of the present invention, the camera for taking an image; A voice output unit for outputting voice information for walking guidance; And acquiring an image from the camera, extracting vanishing point and road outline information using edge information detected in the obtained image, and based on the extracted pedestrian position and the extracted vanishing point based on the extracted outline information. It is characterized in that it comprises a visually impaired gait guide device for providing the converted voice information to the voice output unit by applying the direction of the road determined by the pedestrian determined to the predetermined gait guide information.

The walk guide apparatus for the visually impaired determines whether an obstacle is detected among the images in the walking area obtained through the outline information. When the obstacle is detected, the coordinates of the image in the walking area for each predetermined distance are determined. The position of the obstacle and the distance between the obstacle and the pedestrian are estimated based on the generated conversion table, and the estimated position of the obstacle and the distance between the obstacle and the pedestrian are applied to predetermined obstacle guide information. It is preferable to provide the voice information to the voice output unit.

According to an embodiment of the present invention by providing a voice information on the road so that the visually impaired can walk safely using a single camera even in an outdoor environment having complex background information, it is correct and safe for the blind even in an outdoor environment There is an effect that can guide the walk.

In addition, according to an embodiment of the present invention by selecting a gait area in the image of the outdoor environment obtained from the camera and when the obstacle existing in the gait area is detected, by indicating the location of the obstacle and the distance between the obstacle and the pedestrian, The collision with the obstacles can be prevented in advance, so that it is possible to escape the dangerous situation.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

1 is a view illustrating a visually impaired walking guidance system according to an embodiment of the present invention, Figure 2 is a block diagram of the visually impaired walking guidance system shown in FIG.

Referring to FIG. 1, a visually impaired walking guidance system according to an embodiment of the present invention provides a single camera 10, a voice input / output unit 20, and a visually impaired person to provide guide information for walking to a visually impaired person. It is comprised including the walking guide apparatus 30. FIG. The visually impaired gait guide device 30 obtains an image taken from the camera 10 through a cable 40, but the present invention is not limited thereto, and an image taken from the camera 10 is acquired through a wireless communication method. May be

In the present embodiment, the area in front of which the pedestrian can walk is defined as a walking area B, and the area in front of which is obtained through the camera 10 is defined as the viewing area A that the pedestrian sees.

The visually impaired walking guidance system according to the present invention obtains an image photographed from the camera 10, extracts information necessary for walking by analyzing the acquired image, and converts the result into voice information and outputs the converted information. In this case, the TTS (Text To Speech) engine for converting the pedestrian guide information and the disability guide information required for walking into voice according to a preset voice conversion format is adopted. To. The converted voice information is output through the voice output unit 21, so that the visually impaired person can listen while walking.

Referring to FIG. 2, the voice input / output unit 20 is a voice output unit 21 for outputting voice information for walking guidance, a voice input unit 22 for receiving a voice, and the voice output unit 21 is a voice output unit 21. For example, it may be a headset, and the voice input unit 22 may be a microphone, for example.

The visually impaired gait guide device 30 obtains an image from the camera 10, extracts vanishing point and outline information of the road using edge information detected in the acquired image, and then positions the pedestrian based on the extracted outline information. The first sentence generated by applying the direction of the pedestrian's view determined on the basis of the vanishing point is converted into the voice information and output through the voice output unit 21. In this case, the first sentence is a pedestrian guide information of a pre-configured template structure. For example, the pedestrian guide information is "You are proceeding at the [pedestrian position] of the road. It can be seen, "is a sentence completed by applying the extracted pedestrian position and the direction of the road respectively.

In addition, the visually impaired pedestrian guide device 30 determines whether there is an obstacle in the pedestrian area according to whether the vertical component is extracted from the image in the pedestrian area B obtained through the extracted outline information of the road. When the vertical component is extracted, it is determined that there is an obstacle among the images in the walking area, and the visually impaired walking guide apparatus 30 generates a conversion table by using the preset image coordinates in the walking area for each actual distance. Based on the conversion table, the position of the obstacle and the distance between the obstacle and the pedestrian are estimated, a second sentence is generated by applying the estimated position of the obstacle and the distance between the obstacle and the pedestrian, and the second sentence is converted into voice information. Output via (21). In this case, the second sentence is the disability guide information, and has a template structure which is preconfigured like the walking guide information. For example, the disability information may be "an obstacle in the direction of the front [the location of the obstacle]. There is an obstacle in the distance of [the distance between the obstacle and the pedestrian] m.", And the estimated position of the obstacle, This sentence is completed by applying the distance to the pedestrian.

The detection of whether an obstacle exists is to detect a characteristic of the vertical component of the obstacle existing in the walking area B. Here, the detection target is selected as a representative obstacle, such as people, trees, pillars, etc. having a strong vertical component, but the present invention is not limited thereto. A vertical Gabor filter is applied to detect a strong portion of the vertical component, and the position of the final obstacle is detected through clustering of adjacent vertical components.

A visually impaired walking guide method of the visually impaired walking guidance system having such a configuration will be described with reference to FIGS. 3 and 4.

FIG. 3 is a flowchart illustrating a method for guiding a visually impaired person walking according to an embodiment of the present invention, and FIG. 4 is a flowchart illustrating a subroutine related to extraction of vanishing point and road outline information shown in FIG. 3.

Referring to FIG. 3, the visually impaired gait guide device 30 obtains an image from the camera 10 (S100). The acquired image is as shown in (a) of FIG. 9.

Next, the visually impaired gait guide device 30 performs a subroutine related to vanishing point and road outline information extraction for extracting vanishing point and road outline information based on the detected edge image among the acquired images (S200 of FIG. 4). ). At this time, in order to extract the outline information of the road, a linear component having a specific direction is extracted using an extended chain code. The extended chain code utilizes the feature that the outline of the three-dimensional road is directed toward the vanishing point in the two-dimensional image. Finally, the vanishing point is extracted by extracting the outline information of the road.

Referring to FIG. 4, a subroutine related to extraction of vanishing point and road outline information (S200) will be described. The visually impaired walking guide apparatus 30 converts an image acquired from the camera 10 into a gray image (S201).

Next, the visually impaired gait guide device 30 detects an edge image from the converted gray image (S203). An edge image refers to an outline detected in a gray image, as shown in FIG. 9B.

Next, the visually impaired gait guide device 30 finds connectivity for neighboring pixels in eight directions of each edge pixel in the detected edge image (S205).

That is, the visually impaired walking guide device 30 extracts a shape formed by a straight line using, for example, an eight-way chain cord, and calculates an angle thereof to detect a straight line connected in a direction to be found. As shown in (a) of FIG. 5, an eight-way chaincode unit is scanned horizontally from the beginning of an edge image to select an edge pixel first found as a starting point, and as shown in (b) of FIG. Search for pixels to find successive connectivity.

Next, the visually impaired gait guide device 30 extracts the linear components continuously related to the found connectivity in a specific direction (S207).

Next, as shown in (c) of FIG. 9, the visually impaired gait guide device 30 extracts a linear component directed to the center in each area divided by the center of the acquired image (S209). In this embodiment, in consideration of the location of the vanishing point is relatively centered through the extraction of the linear component, as shown in Figure 6 by dividing the image to the left and right to extract the linear component toward the vanishing point.

Next, the visually impaired gait guide device 30 selects and displays a linear component of a predetermined length or more continuously from the extracted linear components (S211). At this time, the visually impaired gait guide device 30 is to remove the linear components outside the allowable angle range on the basis of the central direction component, if there are two linear components in the allowable angle range closer to the central direction angle It is preferable to select the linear component.

Next, the visually impaired gait guide device 30 should calculate the central direction angle, and in consideration of the simplicity of the calculation, sets an angle with respect to the eight-way component of the chain code unit, and accumulates the linear components that find the connectivity and averages them. (S213). At this time, the angle is to give a component of 0 to 360 degrees starting from the zero-direction component of the chain code unit. The chain of pixels finally detected is written in one straight line and is not retrieved again. After detecting one straight line, the starting point is searched again. In this case, pixels belonging to the previously detected straight line are excluded from the search target. The parameters to be determined here are the basic unit length, the allowable angle range, and the direction of the straight line to be detected.

Next, the visually impaired gait guide device 30 extracts the intersection point of the average linear components in each region as a vanishing point, and extracts the average linear components in the respective regions as outline information of the road (S215).

Again, in step 3, the visually impaired walking guide device 30 selects the walking area B as shown in FIG. 9D by combining the extracted outline information of the road and the vanishing point (S300).

Next, the visually impaired gait guide device 30 determines the direction of the road and the position of the pedestrian (S302). Here, the visually impaired pedestrian guide device 30 calculates the intersection of the average linear component in the lower portion of the image of the selected pedestrian area and the width of each walking area as the width of the walking area, and the center point in the width as the center of the road. Calculate The pedestrian's position is determined by judging which direction the pedestrian is biased by comparing the center of the road with the center of the image. The direction of the road for pedestrians is determined.

Next, the visually impaired gait guide device 30 determines whether an obstacle is detected in the image in the gait area B (S304). When an obstacle is detected, a vertical component is extracted in an image.

In this embodiment, in order to detect an obstacle present in the walking area B, feature information of the obstacle must be extracted. Obstacles are selected for objects such as people, columns, and trees that have strong characteristics of vertical components. At this time, the Gabor filter in the vertical direction was applied to extract the strong vertical component. The bottom and width information of the obstacle is determined through the bundle of extracted vertical components. FIG. 7 is a diagram illustrating a kernel of a Gabor filter, and FIGS. 7A and 7B show respective filter images for a case where a variable value representing a scale of a Gabor filter is determined differently.

FIG. 10 is a diagram illustrating a process of estimating a position of an obstacle. FIG. 10A illustrates an image obtained from the camera 10, and FIG. 10B illustrates FIG. 10A. The edge image detected from the image shown in FIG. 10 (c) is a diagram showing a linear component toward the center in each region bisected with respect to the center of the edge image. d) shows a pedestrian area selected using vanishing point and road outline information, and FIG. 10 (e) shows vertical components extracted by applying an image of the selected pedestrian area to a Gabor filter. FIG. 10 (f) is a view showing a strong vertical component in FIG. 10 (e), and FIG. 10 (g) is a view showing a clustered state of FIG. 10 (f). 10 (h) is a view showing an image in which the position of the last obstacle is displayed.

As a result of the determination in step S304, when no obstacle is detected in the image in the walking area B, the visually impaired walking guide device 30 applies the determined direction of the road and the position of the pedestrian to predetermined walking guide information. To convert the voice information into voice information and output the voice information to the voice output unit 21 (S306).

As a result of the determination in step S304, when an obstacle is detected in the image in the walking area B, that is, when there is an obstacle in front of the image, the visually impaired walking guide device 30 is located in the walking area at a predetermined distance. A conversion table is generated using the image coordinates (S305).

Next, the visually impaired gait guide device 30 estimates the position of the obstacle and the distance between the obstacle and the pedestrian using the conversion table (S307).

In more detail, the visually impaired gait guide device 30 estimates the actual distance of the obstacle projected on the image using a conversion table representing the coordinates of the image for each actual distance. Indeed, when using a single image for the three-dimensional information, the depth information is not known, but in the case of obtaining the vanishing point information on the road as described above, the depth information can be known by considering some limitations. The limitations are that the road outlines are parallel, the extension line meets the vanishing point in the image, and the location of the vanishing point is close to the center of the image. Therefore, the actual distance of the position coordinates projected on the image can be estimated using the conversion table for the actual distance and coordinates in the image as shown in FIG. 8.

Next, the visually impaired gait guide device 30 is controlled to convert the sentence made by applying the estimated position of the obstacle and the distance between the obstacle and the pedestrian to the predetermined disability guide information to voice information to be output to the voice output unit 21. (S309), the process moves to step S306 described above.

In this embodiment, a new sentence structure is used to inform the pedestrian of the information of the road, that is, the location and direction of the pedestrian, and the information of the obstacle, that is, the location of the obstacle and the distance between the obstacle and the pedestrian. Voice conversion process is required. In order to construct a sentence, a method of inputting a word suitable for a predetermined template-based sentence is used. 11 is to detect the position and gaze information in the walking area that the pedestrian is currently walking and obstacles present in the front and then extract the information about the position and input the corresponding word in a predefined sentence and provide it to the pedestrian Show the result of the final sentence you want to do.

By doing so, it is possible to help the blind person to walk safely even in an outdoor environment having a complicated background.

The present embodiment can further improve the walking stability of the visually impaired by performing the process from the above-described step S100 to step S309 at a predetermined cycle, for example, 1 minute, 5 minutes, 7 minutes.

The invention being thus described, it will be obvious that the same way may be varied in many ways. Such modifications are intended to be within the spirit and scope of the invention as defined by the appended claims.

1 is a view for explaining a walking guidance system for the visually impaired according to an embodiment of the present invention.

2 is a block diagram of the visually impaired walking guidance system shown in FIG.

3 is an operation flowchart for explaining a walking guidance method for the visually impaired according to an embodiment of the present invention.

FIG. 4 is a flowchart illustrating a subroutine related to extraction of vanishing point and road outline information of FIG. 3.

5 is a view for explaining a chain code and a linear component detection process used in the present invention.

FIG. 6 is a diagram illustrating the application of a region of an image in order to obtain a vanishing point and an outline of a road applied to the present invention; FIG.

FIG. 7 illustrates a Gabor filter kernel used for obstacle detection in the present invention. FIG.

8 is a view showing the relationship between the actual distance and the position in the image used in the distance estimation of the obstacle in the present invention.

9 is a diagram illustrating a detection result of a walking area in an image acquired from a camera.

10 is a diagram illustrating an obstacle detection result among images acquired from a camera.

11 is a view showing the voice information to be provided to the visually impaired in the form of sentences through the results obtained in the present invention.

<Description of the symbols for the main parts of the drawings>

10: camera 21: audio output unit

22: voice input unit 30: visually impaired walking guidance device

40: cable

Claims (10)

As a visually impaired walking guidance method of the visually impaired gait guide system including a camera for photographing an image, a voice output unit for outputting audio information, and a visually impaired gait guide device, Acquiring an image from the camera; Extracting vanishing point and outline information of a road using edge information detected in the acquired image; Determining a location of the pedestrian based on the outline information of the road, and determining a direction of the road seen by the pedestrian based on the vanishing point; And And providing the converted voice information to the voice output unit by applying the determined position of the pedestrian and the direction of the road to predetermined walking guidance information, respectively. Determining whether an obstacle is detected among images in the walking area obtained through the outline information of the road; When the obstacle is detected, generating a conversion table using coordinates of the image in the walking area for each predetermined distance; Estimating the position of the obstacle and the distance between the obstacle and the pedestrian based on the conversion table; And And providing the converted voice information to the voice output unit by applying the position of the obstacle and the distance between the obstacle and the pedestrian to predetermined obstacle guide information. delete The method according to claim 1, Extracting the vanishing point and the outline information of the road, Converting the obtained image into a gray image; Detecting an edge image from the converted gray image; Finding connectivity for neighboring pixels in eight directions of each edge pixel in the detected edge image; Extracting linear components continuously related in a specific direction using the connectivity; Extracting a linear component directed toward the center in each area bisected based on the center of the acquired image; Selecting linear components of a predetermined length or more continuously from the extracted linear components; And The visually impaired pedestrian guide includes extracting an intersection of the average linear components in each region as a vanishing point by taking the average of the selected linear components, and extracting the average linear components in each region as outline information of a road. Way. The method of claim 3, The determining step, Selecting a walking area using the vanishing point and the outline information of the road; Calculating an intersection point of an average linear component in the lower end portion of the selected gait area and the respective areas as the width of the gait area, and calculating a center point in the width as the center of the road; Determining the location of the pedestrian by comparing the center of the road with the center of the image; And And comparing the direction of the vanishing point and the center of the image to determine the direction of the road for pedestrians. The method according to claim 1, Determining whether the obstacle is detected, Extracting vertical components by applying the image in the gait area to a Gabor filter; And And determining the bottom surface and width information of the obstacle through the bundle of the extracted vertical components. The method of claim 3, Extracting the vanishing point and the outline information of the road, A walk guide method for the visually impaired, characterized by extracting straight line information having a specific direction using an extended chain cord. As a visually impaired walking guidance method of the visually impaired gait guide system including a camera for photographing an image, a voice output unit for outputting audio information, and a visually impaired gait guide device, Acquiring an image from the camera; Extracting vanishing point and outline information of a road using edge information detected in the acquired image; Determining a location of the pedestrian based on the outline information, and determining a direction of a road seen by the pedestrian based on the vanishing point; Determining whether an obstacle is detected among images in the walking area obtained through the outline information; When the obstacle is detected, generating a conversion table using coordinates of the image in the walking area for each predetermined distance; Estimating the position of the obstacle and the distance between the obstacle and the pedestrian based on the conversion table; And The determined position of the pedestrian and the direction of the road are applied to predetermined pedestrian guide information, and the estimated voice position and the distance between the obstacle and the pedestrian are applied to predetermined obstacle guide information to convert the voice information. And a step of providing the voice output unit. The method of claim 7, The providing to the voice output unit, Converting the completed sentence into voice information by applying the determined pedestrian position and the direction of the road to the predetermined walking guidance information, And a visually impaired pedestrian guide method, by applying the estimated position of the obstacle and the distance between the obstacle and the pedestrian to the predetermined obstacle guide information. A camera for taking an image; A voice output unit for outputting voice information for walking guidance; And Acquire an image from the camera, extract vanishing point and road outline information using edge information detected in the acquired image, and based on the extracted pedestrian position and the vanishing point determined based on the extracted outline information. And a visually impaired gait guide device for providing the converted voice information to the voice output unit by applying the direction of the determined pedestrian's view to predetermined gait guide information. The visually impaired gait guide device, It is determined whether an obstacle is detected among the images in the walking area obtained through the outline information, and when the obstacle is detected, based on the conversion table generated using the coordinates of the image in the walking area for each predetermined distance. The voice output unit may be configured to estimate the location of the obstacle and the distance between the obstacle and the pedestrian, and apply the estimated voice location and the distance between the obstacle and the pedestrian to predetermined obstacle guide information. Walking aid system for the visually impaired, characterized in that provided in the. delete

Images