KR101049155B1 - Method for judging obstacle of autonomous moving apparatus and autonomous moving apparatus - Google Patents

Abstract

PURPOSE: An automatic moving device and an obstacle determining method of the device are provided to make more accurate decision if an obstacle is a human. CONSTITUTION: An obstacle determining method of an automatic moving device is as follows. An obstacle is detected from a front area. It is determined whether an obstacle is a moving obstacle or a fixed obstacle. It is determined what object the obstacle is. The contour of the obstacle is detected. The detected contour is changed into a defined shape of a figure. The changed data is compared with a shape of an object predefined as a figure. It is determined whether the changed data corresponds to a specific kind of an obstacle if being matched with the predefined shape of object.

Description

Obstacle Determination Method of Autonomous Mobile Device and Autonomous Mobile Device Using It {METHOD FOR JUDGING OBSTACLE OF AUTONOMOUS MOVING APPARATUS AND AUTONOMOUS MOVING APPARATUS}

The present invention relates to an obstacle determination method of the autonomous mobile device and an autonomous mobile device using the same.

In general, autonomous mobile devices such as autonomous mobile robots are used in various industries.

The autonomous mobile robot assists the disabled, performs logistical transfer operations at the factory, or performs human operations in dangerous environments such as monitoring and alerting, space exploration, nuclear waste treatment plants, or the deep sea.

In addition, autonomous mobile robots can be used as unmanned cleaners, unmanned lawn mowers, and are also used for military purposes.

However, to date, the cognitive ability and reasoning ability of autonomous mobile devices such as autonomous mobile robots are inferior to those of humans, so they have a very limited disadvantage.

Therefore, researches to increase the intelligence of robots in many fields, including movement and object judgment, have been made.

Above all, the most basic function of autonomous mobile devices such as autonomous mobile robots should be able to move freely to the desired target point.

In order for the autonomous mobile robot to move freely, it needs the ability to determine where it is and the ability to identify obstacles around the movement path.

These functions can be performed by the localization technology, obstacle detection and map building technology of the autonomous mobile robot.

Among them, interest has been increased to detect obstacles to the front area to which an autonomous mobile device such as an autonomous mobile robot moves, and to determine whether the obstacle is a human or a simple obstacle.

If a person is a person, it is necessary to develop a technology for avoiding the person or finding a new moving path by modifying the moving path of the autonomous mobile robot, and various research and developments have been in progress.

The present invention has been made in view of the above, and provides an obstacle determination method of an autonomous mobile device that can move autonomously and an autonomous mobile device using the same.

In addition, the present invention provides an obstacle determination method of the autonomous mobile device that can be more accurately determined when the obstacle is a person and an autonomous mobile device using the same.

In addition, the present invention provides a method for determining an obstacle of an autonomous mobile device and an autonomous mobile device using the same.

Obstacle determination method according to an aspect of the present invention, the obstacle detection step for detecting the obstacle to the front area, the obstacle classification step for classifying whether the obstacle is a moving obstacle or a fixed obstacle and the obstacle to determine which object the obstacle Determination step is included.

The obstacle detecting step of the obstacle determining method may include an outline detecting step of detecting the outline of the obstacle, a data converting step of converting the outline detected in the outline detecting step into a prescribed shape, and converting the converted data into a predefined object shape. The method may include a data comparing step for comparing and a data matching step for preparing whether a specific kind of obstacle corresponds.

The obstacle determining method may further include a candidate obstacle selecting step of measuring or estimating the size of the obstacle and selecting the obstacle as an obstacle if the size is larger than a predetermined size, and performing data conversion on the obstacle selected as the candidate obstacle.

The candidate obstacle selection step of the obstacle determination method may include a size determination step of determining the size of the obstacle using a distance to the obstacle and the detected contour, and a size comparison step of comparing whether the size is greater than or equal to a prescribed size. .

The contour detection step of the obstacle determination method may include a scan step of scanning an area including an obstacle in two dimensions, a data synthesis step of accumulating two-dimensional information of the scan step to create three-dimensional shape information, and a scan direction of the synthesized data The method may include extracting two-dimensional shape information by projecting the plane perpendicular to the plane.

In the obstacle matching step of the obstacle determination method, it may be determined whether at least one of the color or the posture of the obstacle corresponds to a specific type of obstacle.

In the obstacle matching step, the color may use skin color according to race when the obstacle is a human.

In the obstacle matching step, the posture may determine whether the obstacle corresponds to one of a standing posture, a sitting posture, a squat posture, and a lying posture.

In the data comparison step of the obstacle determination method, the predefined object shape may be classified according to the aspect ratio of the person or the height and shoulder width.

In the obstacle detecting step of the obstacle determination method, any one or all of a predetermined object shape detectable by an algorithm such as the distance to the obstacle and the contour of the obstacle and the general hop transform may be applied simultaneously.

The obstacle classification step of the obstacle determination method may be classified according to whether obstacles move over time.

In the obstacle determining step of the obstacle determining method, center points for a plurality of moving obstacles may be tracked.

The obstacle determining step of the obstacle determining method may be classified as a moving obstacle or a fixed obstacle according to the moving speed of the center point with respect to the plurality of moving obstacles.

An autonomous moving device according to an aspect of the present invention corresponds to an obstacle of a certain type by using a matching portion between a photographing unit scanning an obstacle with respect to the front region and the contour of the obstacle obtained from the scan and a shape of a predefined obstacle. A first detection unit determining whether the obstacle is moving, a first detection unit detecting whether the obstacle is a moving obstacle by sensing a moving speed of the center point of the obstacle, and a second detection unit detecting whether the obstacle is an organism by detecting the movement of the obstacle And a third detecting unit detecting a wavelength of the obstacle to detect whether or not a person and a determining unit determining which object is the detected obstacle.

The imaging unit of the autonomous mobile device may be any one of a stereo camera or a laser ranging sensor (LADAR, Flash Lidar, etc.).

The autonomous movement apparatus may further include a second detection unit which extracts a two-dimensional shape of the movement obstacle by performing an additional scan within a range including the movement obstacle.

According to the obstacle determining method of the autonomous mobile device according to the present invention and the autonomous mobile device using the same, it is possible to accurately determine and classify the obstacles to determine what kind of obstacles.

Obstacle determination method of the autonomous mobile device and the autonomous mobile device using the same can be more stably determined when the obstacle is a human.

1 is a schematic diagram showing an obstacle detected by the autonomous platform according to an embodiment of the present invention.
2 is a flowchart illustrating a method of determining an obstacle of an autonomous mobile device according to an embodiment of the present invention.
Figure 3 is a block diagram showing the configuration of an autonomous mobile device according to an embodiment of the present invention.
4 is a flow chart showing in detail the obstacle detection step in the obstacle determination method of FIG.
5 to 6 are views for explaining a criterion for determining an obstacle as a person in the obstacle detection step of FIG.
7 is a conceptual diagram illustrating a method for precisely detecting an obstacle detected in the obstacle detecting step of FIG. 4.
8 is a flowchart illustrating a method of precisely detecting an obstacle of FIG. 7.
9 is a flow chart showing in detail the obstacle classification step in the obstacle determination method of FIG.
10 is a flowchart illustrating an obstacle determination step in detail in the obstacle determination method of FIG. 2.

The above-described features and effects of the present invention will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, and thus, those skilled in the art to which the present invention pertains may easily implement the technical idea of the present invention. Could be. As the inventive concept allows for various changes and numerous embodiments, particular embodiments will be illustrated in the drawings and described in detail in the text. However, this is not intended to limit the present invention to a specific disclosure, it should be understood to include all modifications, equivalents, and substitutes included in the spirit and scope of the present invention. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

Hereinafter, a method of determining an obstacle of an autonomous mobile device and an autonomous mobile device using the same will be described with reference to the accompanying drawings.

1 is a schematic diagram showing an obstacle detected by the autonomous platform according to an embodiment of the present invention.

Referring to the drawings, the autonomous mobile device 1 is an autonomous mobile device, and may be, for example, any one of a vehicle and a robot, but is not particularly limited.

The autonomous platform moves while scanning the obstacles 2 and 3 in front, and determines which obstacle corresponds to the obstacle when the obstacle is detected.

2 is a flowchart illustrating an obstacle determining method of an autonomous mobile device according to an embodiment of the present invention.

First, the autonomous platform 1 performs an obstacle detecting step ST100 for scanning an obstacle on a front area to determine an obstacle in front.

The obstacle detection step ST100 may scan an obstacle using any one of a stereo camera or a laser ranging sensor (LADAR, Flash Lidar, etc.).

The obstacle scanning step ST100 detects the distance to the obstacle and the outline of the obstacle, converts the detected outline of the obstacle by an algorithm such as a generalized Hough transform, and compares it with a predefined object shape and detects it accordingly. It is possible to roughly determine which obstacle the object falls under.

Detecting the outline of the obstacle can detect the outline of the obstacle after precisely scanning the obstacle in three dimensions and projecting it in two dimensions.

A more detailed description of the obstacle detection step will be described later.

Next, the autonomous mobile device 1 performs an obstacle classification step ST200 for classifying whether the obstacle is a moving obstacle or a fixed obstacle.

The obstacle classification step ST200 may be classified according to whether obstacles move over time.

In the autonomous mobile device 1, the obstacles 2 and 3 located in the front can be largely classified into moving obstacles and fixed obstacles. The moving obstacles refer to obstacles that can move, such as a person, an animal, or an insect.

Here, the meaning of the movement means not merely a reciprocating movement of a predetermined section repeatedly, but means an obstacle capable of actively moving the position.

Fixed obstacles are obstacles that cannot be actively moved, such as rocks, trees, puddles, pillars and bushes, but are passively moved by wind or physical external force applied from the outside.

In the obstacle classification step ST200, center points for a plurality of moving obstacles may be tracked.

The center point may correspond to a single moving obstacle or to a center point of each moving obstacle.

For example, when the obstacle is a rectangular shape, the inner center of the rectangular shape may be the center point, and in the case of a person, the center point may be near the navel.

The reason for tracking the center point of the obstacle is to determine whether it is a moving obstacle or a fixed obstacle in consideration of the direction and speed of movement over time, and further to determine whether it is a person or another object.

A more detailed description of the obstacle classification step will be described later.

Next, the autonomous platform 1 performs an obstacle determination step (ST300) for determining whether the obstacle is a human being, a non-human creature, or an inanimate object.

In the case of detecting an organism, if a motion such as a posture is detected using a motion detection sensor, it is classified as an organism.

In the case of detecting a person, when a wavelength of a range emitted from the person is detected using a human body sensor, the person is classified as a person.

A more detailed description of the obstacle determination step will be described later.

Next, an autonomous mobile device according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

3 is a view showing an autonomous mobile device according to an embodiment of the present invention.

The autonomous mobile device includes a photographing unit 100, a first detecting unit 200, a second detecting unit 300, a first detecting unit 400, a second detecting unit 500, a third detecting unit 600, and a determination unit. It comprises a part 700.

The photographing unit 100, the autonomous mobile device 1 scans the obstacles (2, 3) to the front area.

The photographing unit 100 may be any one of a stereo camera or a laser ranging sensor (LADAR, Flash Lidar, etc.). Detailed description of the stereo camera or laser distance measuring sensor (LADAR, Flash Lidar, etc.) will be omitted.

When the obstacle is detected, the first detector 200 may detect an outline of the obstacle, and detect a distance from the obstacle and a size of the obstacle.

Accordingly, by using the matching relationship between the contour of the obstacle and the shape of the predefined obstacle, it is possible to determine what kind of obstacle it corresponds to, and determine whether the obstacle is over a certain size to determine whether the object has an obstacle possibility. have.

The second detector 300 performs an additional scan within a range including the obstacle and detects the obstacle by performing 3D shape restoration and 2D projection on the obstacles 2 and 3. Range sensor may be used as the second detector 300.

Accordingly, the contour of the obstacle can be detected more accurately, and it is possible to more accurately determine what kind of obstacle corresponds to using a matching relationship with the shape of the predefined obstacle.

The first detection unit 400 detects whether the obstacle is a moving obstacle by detecting a moving speed of the center point of the obstacle.

The second detecting unit 500 detects a movement over time when the obstacle 2 is a moving obstacle. The second detection unit 500 may use a motion detection sensor.

At this time, if the motion detection sensor detects movements such as posture of obstacles, it is classified as a living organism.

The third sensing unit 600 detects whether the obstacle is a person or another object. The third detection unit 600 may use a human detection sensor.

The determination unit 700 is input through the first detection unit 200, the second detection unit 300, the first detection unit 400, the second detection unit 500 and the third detection unit 600. Determine whether the detected obstacle is a person or other object.

The determination unit 700 includes the photographing unit 100, the first detection unit 200, the second detection unit 300, the first detection unit 400, the second detection unit 500, and the third detection unit ( Various signals and information input through 600 may be determined and may be connected to various navigation devices.

Next, the obstacle detection step in the obstacle detection method according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

4 is a flowchart illustrating an obstacle detection step in detail in the obstacle determination method of FIG. 2.

The obstacle detecting step is performed through the first detector 200.

First, the autonomous platform 1 obtains image information by performing a scan (ST100) of the front area while driving on a road or a plain.

At this time, the stereo camera provided in the photographing unit 100 scans the front area and transmits the scanned image information to the first detection unit 200.

Since the scanned image information contains a lot of information, the edge is detected through edge detection on the imaging screen, and the obstacle is detected by classifying the object forming the closed loop as an object having a potential obstacle. .

Next, depth information is calculated from the scanned image information (ST110), and the distance D to the obstacles 2 and 3 is calculated (ST120). Accordingly, the size of the object can be estimated.

It can be represented by the distance D = f (focal length, horizontal and vertical pixels, image detection device size, depth). Here, the actual horizontal size of the object may be represented by the horizontal actual size ㅧ pixels of 1 pixel, the actual vertical size of the object may be represented by the vertical actual size ㅧ pixels of 1 pixel.

Next, a plurality of objects located in front of the autonomous platform 1 are set whether or not they correspond to an object having a possibility of obstacle (hereinafter referred to as a candidate obstacle) (ST130).

That is, in the case of a very small object, it is not necessary to recognize it as an obstacle, but there is a possibility that it may be wrongly determined as an obstacle. Therefore, the object is classified as a candidate obstacle having a horizontal size and a vertical size larger than a predetermined size from the estimated size of the object (ST130).

Next, to detect whether the object corresponds to a predefined object shape among candidate obstacles, the object is compared with a predefined object shape (ST140).

To do this, a generalized Hough transform is applied to candidate obstacles and it is determined whether the shape is similar to a predefined object shape stored in the object definition DB. General hop transform transforms candidate obstacles into the shape of a figure.

For example, if the shape to which the normal hop transform is applied to the candidate obstacle is a rectangle and the predefined object shape stored in the object definition DB includes a rectangle, the candidate obstacle is determined to have the shape defined in the predefined object shape. do.

Next, when the shape to which the general hop transform is applied has a shape defined in the predefined object shape, it is determined whether the shape to which the general hop transform is applied corresponds to a specific obstacle.

For example, if the specific obstacle is a person, it is determined whether the shape to which the general hop transform is applied has a characteristic shape of the person.

The following example illustrates an example of determining whether a candidate obstacle corresponds to a person when a rectangle is included in a predefined object shape and a shape in which a general hop transform is applied to the candidate obstacle corresponds to a rectangle.

5 to 6 are diagrams for describing a criterion for determining a candidate obstacle as a person in the obstacle detecting step of FIG. 4.

If a candidate obstacle is detected, it is checked whether the obstacle has a color corresponding to the human skin color. The skin color can be divided into three types according to race (white race, black race, yellow race).

Also, check whether the figure detected through the normal hop transform corresponds to the aspect ratio of the person and whether the standing posture, sitting posture, lying posture, or squatting posture by checking the posture of the obstacle. You can check if this is the case.

The posture may be other than the above-mentioned posture, but is limited to some for convenience of description.

However, it is possible to change to the above-described other postures, and it is clear that various postures can be mapped in advance and stored in the autonomous mobile device 1.

This series of procedures detects whether the object can be classified as skin color, corresponds to the aspect ratio of a person, and what posture is taken and completes the detection (ST150).

In this case, classifying the figure detected by the general hop transform into a person may be performed as in the following example.

5 to 6, when a person stands among the plurality of candidate obstacles detected, a range of 10% to 90% of the height A according to the body classification standard and 10% to 90% of the shoulder width B are shown. By selecting a range, objects within the ratio range can be classified as humans.

For example, if the height (A) / width (B) ratio of an object belongs to A ₁₀ / B ₉₀ to A ₉₀ / B ₁₀ , it can be considered as a person.

If an obstacle is sitting, the obstacles included in the ratio range by selecting the height corresponding to 1/2 of the 10% to 90% of the height and the 10% to 90% of the shoulder width in the body classification standard. You can classify as a person.

For example, when the vertical (height, A) / width (shoulder width, B) a ratio of an obstacle _{(A 10/2) / B} 90 ~ (A 90/2) belonging to a / B ₁₀ may be classified as a person .

If an obstacle is lying, the standing case is converted into a lying case and judged.If the vertical (B) / horizontal (A) ratio of the obstacle belongs to B ₁₀ / A ₉₀ to B ₉₀ / A ₁₀ , Can be classified.

If the obstacle is curled up on the ground, it is judged by converting the sitting case into a lying down, and the vertical (B) / horizontal (A) ratio of the obstacle is (B _10/2 ) / A ₉₀ ~ If they belong to (B _90/2 ) / A ₁₀ , they can be classified as human.

It is possible to detect which object the candidate obstacle detected by the above-described process corresponds to, and thus it is possible to determine it as an obstacle.

The obstacle detection step as described above may be precisely performed in the following manner. This is performed by the second detector 300. The operator may perform the obstacle detection step by the first detection unit 200 and the obstacle detection step by the second detection unit 300 simultaneously or selectively.

7 is a conceptual diagram illustrating a method of precisely detecting an obstacle detected in the obstacle detecting step of FIG. 4.

When the candidate obstacle is detected, the second detector 300 performs additional scanning on the peripheral area where the candidate obstacle is located.

The second detector 300 is a range sensor including a laser scanner and performs a scan for obtaining range data on a plane perpendicular to the ground.

The range sensor may be mounted on a separate driving unit, and the image sensor is obtained by scanning a front region of the autonomous platform 1 including the object with the range sensor.

The image data obtained at this time is a three-dimensional image. That is, referring to the drawing, when a spherical object exists at a predetermined distance as shown, the scanned image data corresponds to data representing a surface of one side of the sphere, which is a three-dimensional image.

At this time, the portion existing as a space is a case in which there is no range value because the range sensor signal of the obstacle 2 is not returned, and when there is a return value reflected by a sphere, color is given according to the range. A video image can be generated.

The three-dimensional image of the scanned obstacle 2 can be made into a two-dimensional image by projecting onto the projection plane.

Edge detection and transformation are performed on the two-dimensional image generated by the projection, and the matching process is performed to compare with the predefined object shape through the predefined object shape. Matching is performed.

FIG. 8 is a flowchart illustrating a method of precisely detecting an obstacle of FIG. 7 and illustrates a method of precisely detecting an obstacle using a first detector and a second detector.

 First, the photographing unit 100 scans image information on an obstacle (ST100), and the first detection unit 200 detects a plurality of obstacle candidates (ST150).

Next, the center point and the size of the obstacle of interest among the plurality of obstacle candidates are calculated (ST200), and the range sensor provided in the second detector 300 detects an angle and an existence range of the obstacle.

Next, a calculation is performed on the input obstacle angle and the existence range (ST281), and the area data is obtained and accumulated (ST282). This is to obtain a more accurate shape for the obstacle.

After the color mapping of the obstacle of interest (ST283) is performed, projection (ST284) is performed on the plane, and a two-dimensional virtual image based on the area is obtained (ST285).

Then, the shape of the obstacle of interest is extracted (ST286), and the predefined object shape of the object definition DB and the extracted shape is compared (ST287) through a general hop transform. In addition, matching is performed whether the extracted obstacle shape corresponds to a specific obstacle (ST288).

Next, the obstacle classification step in the obstacle detection method according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 9 is a flowchart illustrating an obstacle classification step in detail in the obstacle determining method of FIG. 2.

First, image information of the front region is obtained through an obstacle detection step (ST100), and an obstacle is detected (ST150).

If an obstacle is detected, the center point of each detected obstacle is stored (ST210). The reason for storing the center point for the obstacle is to classify whether the plurality of obstacle candidates correspond to moving obstacles or fixed obstacles.

Next, the image of the obstacle is continuously taken to store and store the center point of the obstacle (ST220), and the moving speed of the center point of the obstacle is calculated from the accumulated image (ST230).

It is determined whether the movement speed of the obstacle is equal to or greater than a predetermined threshold (constant value) that can be classified as a movement obstacle (ST240). If the movement speed is greater than or equal to the predetermined value, the obstacle is classified as a movement obstacle (ST250).

If the moving speed of the obstacle is below a certain value, it is classified as a fixed obstacle (ST260).

Here, the meaning of the constant threshold refers to a reference time at which a time for moving to another location for t hours from the current location is set, and the obstacle may be classified as a moving or fixed obstacle based on the reference time.

Next, the obstacle determination step in the obstacle detection method according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 10 is a flowchart illustrating an obstacle determining step in detail in the obstacle determining method of FIG. 2.

First, when an obstacle is detected (ST150), the moving speed of the center point is determined according to time (ST310). Accordingly, it is possible to determine whether it is a moving obstacle or a fixed obstacle.

If the obstacle is a movement speed of the center point over time is a predetermined value or more to operate the motion detection sensor (ST320). When the motion detection sensor detects a movement such as a posture, it is determined as a moving obstacle, and a human detection sensor is operated to indicate whether the moving obstacle is a person (ST322).

When a signal (wavelength of a person) is detected by the human detection sensor, the moving obstacle is determined to be a person. If a separate signal is not detected by the human detection sensor, it is determined as an organism rather than a human.

Even when the obstacle has a movement speed of the center point over time, the motion detection sensor is operated (ST330).

If a separate signal is not detected by the motion detection sensor, it is determined as a fixed obstacle, and when a signal is detected, the human detection sensor ST332 is operated.

If a signal is detected, it is determined that the moving obstacle corresponds to a person but remains in a fixed state, and if it is not detected, it is determined that the moving obstacle is in a fixed state as other living things.

Accordingly, it is possible to determine whether the obstacle is a human, a living creature, or a non-living object for both the moving obstacle and the fixed obstacle.

As mentioned above, although an embodiment of the present invention has been described, those of ordinary skill in the art may add, change, delete or add components within the scope not departing from the spirit of the present invention described in the claims. The present invention may be modified and changed in various ways, etc., which will also be included within the scope of the present invention.

Claims (16)

An obstacle detecting step of detecting an obstacle with respect to the front region;
An obstacle classification step of classifying whether the obstacle is a moving obstacle or a fixed obstacle; And
An obstacle determination step of determining which object the obstacle corresponds to,
The obstacle detection step,
Contour detection step of detecting the contour of the obstacle;
A data conversion step of converting the contour lines detected in the contour detection step into a prescribed shape;
A data comparison step of comparing the converted data with a predefined object shape defined in the form of a figure; and
An obstacle matching step of determining whether the converted data correspond to a specific kind of obstacle when the converted data is in the predefined object shape
Obstacle determination method of the autonomous mobile device comprising a. delete The method of claim 1,
The obstacle detecting step may further include a candidate obstacle selecting step of measuring or estimating the size of the obstacle and selecting the obstacle as an obstacle if the size is larger than a predetermined size, and performing data conversion on the obstacle selected as the candidate obstacle. How to determine obstacles on your device. The method of claim 3,
The candidate obstacle selection step,
A size determining step of determining the size of the obstacle using the distance to the obstacle and the detected contour;
A size comparison step of comparing whether the size is greater than or equal to a prescribed size
Obstacle determination method of the autonomous mobile device comprising a. The method of claim 1,
The contour detection step,
A scan step of scanning an area including an obstacle in two dimensions;
A data synthesis step of accumulating the two-dimensional information of the scanning step to create three-dimensional shape information; and
Extracting two-dimensional shape information by projecting the synthesized data onto a plane perpendicular to a scan direction
Obstacle determination method of the autonomous mobile device comprising a. The method of claim 1,
The obstacle matching step,
Obstacle determination method of the autonomous mobile device, characterized in that it determines whether it corresponds to a specific type of obstacle by at least one of the color or posture of the obstacle. The method of claim 6,
The color obstacle method of the autonomous mobile device, characterized in that using the skin color (Skin Color) according to the race when the obstacle is a human. The method of claim 6,
The posture is an obstacle determination method of the autonomous mobile device, characterized in that it is determined whether any one of the standing posture, sitting posture, squat on the ground, lying position when the obstacle is a human. The method of claim 1,
In the data comparison step,
The predefined object shape is an obstacle determination method of an autonomous mobile device, characterized in that classified according to the aspect ratio of the person or height and shoulder width. The method of claim 1,
The obstacle detection step,
A method for determining an obstacle of an autonomous mobile device, characterized in that any one or all of a predefined object shape detectable by an algorithm such as a distance to an obstacle, an outline of an obstacle, and a general hop transform can be simultaneously applied. The method of claim 1,
The obstacle classification step,
Obstacle determination method of the autonomous mobile device, characterized in that classified according to the movement of the obstacle over time. The method of claim 1,
The obstacle determination step,
Obstacle determination method of the autonomous mobile device, characterized in that the center point for the plurality of moving obstacles are tracked. The method of claim 12,
The obstacle determining step
Obstacle determination method of the autonomous mobile device, characterized in that classified as a moving obstacle or a fixed obstacle according to the moving speed of the center point for the plurality of moving obstacles. A photographing unit scanning an obstacle for the front region;
A first detection unit that determines what kind of obstacle corresponds to a matching relationship between an outline of an obstacle obtained from the scan and a shape of a predefined obstacle;
A first detecting unit detecting whether the obstacle is a moving obstacle by sensing a moving speed of a center point of the obstacle;
A second sensing unit which senses whether the obstacle is an organism by sensing the movement of the obstacle;
A third sensing unit which senses whether or not a person is detected by sensing the wavelength of the obstacle; And
It includes a determination unit for determining what object the obstacle is detected,
The first detection unit,
An outline detection module detecting an outline of the obstacle;
A data conversion module for converting the contour detected by the contour detection module into the form of a prescribed figure;
A data comparison module for comparing the converted data with a predefined object shape defined in the form of a figure;
Obstacle matching module for determining whether the converted data corresponds to a specific kind of obstacle when the transformed data is in the predefined object shape
Autonomous mobile device comprising a. The method of claim 14,
Wherein,
Autonomous mobile device, characterized in that any one of a stereo camera or a laser ranging sensor is used. 16. The method of claim 15,
And a second detection unit configured to extract a two-dimensional shape of the moving obstacle by performing an additional scan within a range including the moving obstacle.

Images