WO2020098532A1

WO2020098532A1 - Method for positioning mobile robot, and mobile robot

Info

Publication number: WO2020098532A1
Application number: PCT/CN2019/115745
Authority: WO
Inventors: 刘干; 苏辉; 蒋海青
Original assignee: 杭州萤石软件有限公司
Priority date: 2018-11-12
Filing date: 2019-11-05
Publication date: 2020-05-22
Also published as: CN111178366B; CN111178366A

Abstract

A method for positioning a mobile robot, and a mobile robot. The mobile robot can (S101) collect a first image in the current field of view, (S102) match the first image with prestored sample data at a specified position, and (S103) determine the specified position where the sample data is successfully matched with the first image to be the current position of the mobile robot.

Description

Mobile robot positioning method and mobile robot

Technical field

The present application relates to the technical field of mobile robots, in particular to a positioning method of mobile robots and mobile robots.

Background technique

Positioning technology is a key technology in the research of mobile robots. For mobile robots, accurate space positioning technology is the prerequisite for its autonomous navigation and obstacle avoidance.

At present, the method of using image analysis to obtain the position of the mobile robot requires the training of a large number of image samples to calculate and determine the relative position of the mobile robot and the environment, and then obtain the accurate position of the mobile robot itself. However, when the above method is used to obtain the accurate position of the mobile robot itself, high-performance computing hardware is required and the cost is high.

Summary of the invention

In view of this, the present application provides a positioning method and mobile robot for a low-cost mobile robot.

A first aspect of the present application provides a positioning method for a mobile robot. The method is performed by a mobile robot. The method includes: acquiring a first image in a current field of view; and combining the first image with a pre-stored sample at a specified location The data is matched; the specified position where the sample data is successfully matched with the first image is determined as the current position of the mobile robot.

A second aspect of the present application provides a mobile robot. The mobile robot includes an acquisition module, a memory, and a processor, wherein the acquisition module is used to acquire a first image in the current field of view; the memory is used to store a designated Sample data of the location; the processor is configured to match the first image with the sample data and determine the specified location where the sample data and the first image match successfully as the current location of the mobile robot .

A third aspect of the present application provides a computer-readable storage medium on which a computer program is stored, which, when executed by a processor in a mobile robot, causes the mobile robot to: acquire a first image in the current field of view; Matching the first image with the pre-stored sample data of the specified position; determining the specified position where the sample data is successfully matched with the first image as the current position of the mobile robot.

According to the embodiments of the present application, the position of the mobile robot can be accurately located, and the cost is low.

BRIEF DESCRIPTION

1 is a flowchart of a positioning method of a mobile robot according to an exemplary embodiment of the present application;

2 is a flowchart of a process of extracting features from an image according to an exemplary embodiment of the present application;

3 is a flowchart of a process of performing de-redundancy processing on a first image according to an exemplary embodiment of the present application;

FIG. 4 exemplarily shows the effect of applying de-redundancy processing to the first image;

5 is a schematic structural diagram of a mobile robot according to an exemplary embodiment of the present application.

detailed description

Exemplary embodiments will be described in detail here, examples of which are shown in the drawings. When the following description refers to the accompanying drawings, unless otherwise indicated, the same reference numerals in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of devices and methods consistent with some aspects of the application as described in the appended claims.

The terminology used in this application is for the purpose of describing particular embodiments only, and is not intended to limit this application. The singular forms "a", "said" and "the" used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates other meanings. It should also be understood that the term "and / or" as used herein refers to and encompasses any one or all possible combinations of one or more associated listed items.

It should be understood that although the terms first, second, third, etc. may be used to describe various information in this application, the information should not be limited by these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present application, the first information may also be referred to as second information, and similarly, the second information may also be referred to as first information. Depending on the context, the word "if" as used herein can be interpreted as "when" or "when" or "responsive".

The following describes several specific embodiments of the present application. These embodiments may be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

FIG. 1 is a flowchart of a mobile robot positioning method according to an exemplary embodiment of the present application. Referring to FIG. 1, the method may include steps S101 to S103.

In step S101, the mobile robot may collect the first image in the current field of view.

It should be noted that when the mobile robot needs to determine its own position, it will move to one of a plurality of designated positions and collect the first image in the current field of view.

In step S102, the mobile robot may match the above-mentioned first image with pre-stored sample data at a specified position.

Specifically, in an embodiment, the sample data is a second image of the specified location that is collected in advance at different shooting angles. This step S102 may be implemented by the following operations.

(1) Perform feature extraction on the first image and the second image respectively to obtain a first feature descriptor of the first image and a second feature descriptor of the second image. Feature extraction can be performed based on sift feature extraction algorithm, surf feature extraction algorithm, hog feature extraction algorithm, haar feature extraction algorithm, shape context algorithm, etc., to obtain feature descriptors. For the specific implementation principle and implementation process of each feature extraction algorithm, please refer to the description in the related art, which will not be repeated here.

(2) Calculate the similarity between the first feature descriptor and the second feature descriptor. For the method of calculating the similarity between the two feature descriptors, please refer to the description in the related art, which will not be described here. If the similarity between the first feature descriptor and the second feature descriptor is greater than a preset threshold, it can be considered that the first feature descriptor and the second feature descriptor match.

In addition, in another embodiment, the sample data is a second feature descriptor of a second image, and the second image is an image of the specified position pre-collected at different shooting angles, this step S102 may be This is achieved through the following operations.

(1) Perform feature extraction on the first image to obtain a first feature descriptor of the first image.

(2) Calculate the similarity between the first feature descriptor and the second feature descriptor. If the similarity between the first feature descriptor and the second feature descriptor is greater than a preset threshold, it can be considered that the first feature descriptor and the second feature descriptor match.

For example, if the room where the mobile robot is located includes six corners, the six corners may be the specified location, and the different shooting angles may include 20 degrees, 50 degrees, 80 degrees, and so on.

In step S103, the mobile robot may determine the specified location where the sample data matches the first image successfully as the current location of the mobile robot.

2 is a flowchart of a process of extracting features from an image according to an exemplary embodiment of the present application. Referring to FIG. 2, the process of extracting features from an image may include steps S201 to S204.

In step S201, non-maximum suppression processing is performed on the image to obtain the feature points of the image.

For the specific implementation principle and implementation process of this step, please refer to the description in the related art, which will not be repeated here. For example, Table 1 and Table 2 both show the gray values of 9 neighboring pixels. In the example shown in Table 1, the gray value of a pixel with a gray value of 87 is greater than the gray values of other surrounding pixels. At this time, the pixel with a gray value of 87 is considered as a feature point. In the example shown in Table 2, for pixels with a gray value of 40, the gray values of surrounding pixels are larger than 40 and smaller than 40. At this time, the pixels with a gray value of 40 are not Feature points.

Table 1

4343	2626	3232
4343	8787	2626
6565	7777	4545

Table 2

4343	2626	3232
4343	4040	2626
6565	7777	4545

In step S202, for each feature point, the designated neighborhood of the feature point is divided into multiple sub-regions, and for each sub-region, the gradient amplitude and gradient direction of each pixel in the sub-region are calculated.

For example, the designated neighborhood may be a 16 * 16 neighborhood. The 16 * 16 neighborhood of each feature point can be divided into 16 4 * 4 sub-regions.

For a specific method of calculating the gradient amplitude and gradient direction of each pixel point, please refer to the description in the related art, which will not be repeated here.

In step S203, the gradient direction of each pixel is corrected so that the corrected gradient direction is within a specified range.

The gradient direction of each pixel obtained by the above calculation may be in the range of 0 ° to 360 °, and the corrected gradient direction is in the specified range (for example, 0 ° to 180 °).

For example, when correcting the gradient direction of each pixel, if the gradient direction of the pixel is greater than 180 °, the gradient direction of the pixel can be rotated 180 ° in a counterclockwise direction to make it in a plane rectangular coordinate system In the first or second quadrant, the corrected gradient direction is obtained. On the other hand, if the gradient direction of the pixel is less than 180 °, the gradient direction of the pixel can be directly determined as the corrected gradient direction.

In step S204, the feature vector corresponding to the sub-region is obtained according to the gradient amplitude of each pixel in the sub-region and the corrected gradient direction, and the feature corresponding to the feature point is determined according to the feature vector corresponding to each sub-region Vector and the feature vector corresponding to each feature point to determine the feature descriptor of the above image.

For example, for a 4 * 4 subregion, the gradient amplitude and gradient direction of each pixel in the subregion are calculated as shown in Table 3 (where the left side of the slash indicates the gradient amplitude and the right side Indicates the gradient direction):

table 3

10/45°10/45 °	50/270°50/270 °	90/180°90/180 °	150/220°150/220 °
50/60°50/60 °	60/70°60/70 °	100/80°100/80 °	120/130°120/130 °
80/350°80/350 °	90/200°90/200 °	50/30°50/30 °	30/80°30/80 °
130/160°130/160 °	110/50°110/50 °	40/70°40/70 °	90/160°90/160 °

After correcting the gradient direction in step S203, the gradient amplitude and corrected gradient direction of each pixel in the sub-region are shown in Table 4:

Table 4

10/45°10/45 °	50/90°50/90 °	90/180°90/180 °	150/40°150/40 °
50/60°50/60 °	60/70°60/70 °	100/80°100/80 °	120/130°120/130 °
80/170°80/170 °	90/20°90/20 °	50/30°50/30 °	30/80°30/80 °
130/160°130/160 °	110/50°110/50 °	40/70°40/70 °	90/160°90/160 °

According to Table 4, the feature vector corresponding to the sub-region can be obtained. As shown in Table 5, the feature vector of the sub-region can be a 4-dimensional feature vector. The first dimension of the feature vector is the dimension corresponding to 0 °, the second dimension is the dimension corresponding to 45 °, and the third dimension is The corresponding dimension of 90 °, the fourth dimension is the corresponding dimension of 135 °. The feature vector can be calculated by the following method: For a pixel, if the gradient direction of the pixel after correction falls exactly at the boundary point (ie, 0 °, 45 °, 90 °, or 135 °), then directly The gradient amplitude of the pixel is added to the dimension corresponding to the boundary point. Refer to Table 5, for example, for the first pixel, the gradient direction of the pixel after correction is 45 °, which just falls on the boundary point, so the gradient amplitude of the pixel is directly added to the second dimension of the feature vector on. For another example, for a pixel with a gradient amplitude of 90 and a gradient direction of 180 °, the gradient amplitude of the pixel is added to the dimension corresponding to 0 °, that is, to the first dimension of the feature vector.

For a pixel, if the corrected gradient direction of the pixel falls within the interval between two adjacent boundary points, the first distance between the corrected gradient direction of the pixel and the starting point of the interval is calculated first (Ie, the first angle difference) and the second distance between the corrected gradient direction of the pixel and the end point of the interval (ie, the second angle difference), and then assign the pixel according to the ratio of the second distance to the first distance The gradient amplitude of the point, so that the ratio of the gradient amplitude component assigned to the dimension corresponding to the beginning of the interval to the gradient amplitude component assigned to the dimension corresponding to the end of the interval is equal to the second distance and The ratio of the first distance. For example, for a pixel with a gradient amplitude of 150 and a gradient direction of 40 °, the corrected gradient direction of the pixel falls within the interval of 0 ° to 45 °, and the first distance from the starting point of the interval 0 ° is 40 °, the second distance of 45 ° from the end of the interval is 5 °. The ratio of the second distance to the first distance can be calculated as 1: 8. Therefore, the gradient amplitude of the pixel can be divided into 9 equal parts, of which 1 part is added to the dimension corresponding to 0 °, and 8 parts are added to the dimension corresponding to 45 °. That is, the gradient amplitude component corresponding to 8 copies with a value of 133.33 is added to the dimension corresponding to 45 °, and the gradient amplitude component corresponding to 1 share with a value of 16.67 is added to the dimension corresponding to 0 °.

It should be noted that when the gradient direction of a pixel after correction falls between 135 ° and 180 °, the gradient amplitude component corresponding to the second distance corresponding to 180 ° is added to the dimension corresponding to 0 °.

table 5

In this way, through the above method, the feature vector corresponding to each sub-region can be obtained, and then the feature vector corresponding to each sub-region can be combined together (for example, after the feature vector corresponding to the first sub-region is arranged in sequence The feature vector corresponding to the region), that is, the feature vector corresponding to the feature point is obtained. For example, the feature vector corresponding to the feature point may be a 64-dimensional feature vector. The feature vector corresponding to each feature point can be combined together to obtain the feature descriptor of the image.

Through the above image feature extraction process, while ensuring the robustness, the number of dimensions of the calculated image feature descriptor can be reduced, thereby reducing the calculation cost.

Optionally, according to an embodiment of the present application, before step S102, the mobile robot may also perform filter processing, enhancement processing, and / or de-redundancy processing on the first image. For example, the mobile robot may use relevant filtering algorithms and enhancement algorithms to perform filtering processing and enhancement processing on the first image.

FIG. 3 is a flowchart of a process of performing de-redundancy processing on a first image according to an exemplary embodiment of the present application. Referring to FIG. 3, the process may include steps S301 to S302.

In step S301, a redundant area in the first image may be determined.

In some cases, the first image captured by the mobile robot may include a ground image part, and the ground image part is usually weak in texture, and some may be smooth, and may contain a large amount of redundant information. For example, when the mobile robot is a sweeping robot, the sweeping robot works in a home scene, and the floor is a tile or a wooden floor, which is easy to reflect light and has a weak texture. Therefore, when feature extraction is performed on the floor image part included in the first image captured by the cleaning robot, there are often no feature points or fewer feature points. Even if the feature points are extracted, the similarity of the feature points of the floor image part is very high, and it is easy to cause mismatches when matching. If the floor image is partially removed in image processing, the efficiency of single frame processing will be improved, and the reliability of matching will be improved to a certain extent.

As an example, this step S301 can be implemented by the following operations.

(1) Calculate the average value of the gray value of each pixel in the first designated area in the first image.

Referring to the foregoing description, if the first image captured by the mobile robot includes a ground image portion, the proportion of the ground image portion in the height direction of the first image is generally not less than 5%. Therefore, the lowermost part of the first image with a proportion of 5% can be designated as the first designated area. Then, the average value of the gray value of each pixel in the first designated area can be calculated.

(2) The first updated image is obtained by subtracting the average value of the gray value of each pixel in the second designated area in the first image.

The second designated area may be set according to actual needs. For example, the lower half of the first image may be designated as the second designated area. Then, the first updated image can be obtained by subtracting the above average value from the gray value of each pixel in the lower half of the first image.

(3) Update the gray value of each pixel in the second designated area in the first updated image to a gray value greater than the first preset threshold to 255, and update the first updated image The gray value of each pixel in the second designated area in the above is less than or equal to the first preset threshold and the gray value is updated to 0 to obtain a second updated image.

(4) For each row area of the second designated area in the second updated image, the proportion of pixels in the row area with a gray value of 255 is counted to obtain the proportion corresponding to the row area . It should be noted that the proportion of pixels with a gray value of 255 in each row area refers to the ratio of the number of pixels with a gray value of 255 in the row area to the image width.

(5) For the second specified area in the second updated image, when the proportion of the specified number of continuous line areas from top to bottom is greater than the second preset threshold, determine the specified number of continuous lines The last line in the area is the target line number in the second updated image above. The specified quantity can be set according to actual needs, for example, it can be 2. In addition, the second preset threshold may also be set according to actual needs, for example, it may be 50%.

(6) Determine the area between the line indicated by the target line number and the last line in the first image as a redundant area.

In step S302, the gray value of each pixel in the determined redundant area may be updated to 0 to obtain an image after de-redundancy processing.

FIG. 4 exemplarily shows the effects of applying the above-described de-redundancy processing to the first image through images a to e. Image a is an example of the first image collected by the mobile robot.

After performing the above operations (1)-(3) on the image a, the image b (ie, the second updated image) is obtained. Among them, the gray value of most pixels in the floor image part has been set to 0, and there are still sporadic white points. For this reason, before performing the above operation (4), the image b may be subjected to a morphological operation to remove sporadic white dots. (For the specific implementation principle and implementation process of the morphological operation, please refer to the description in the related art It will not be repeated here), and the image c is obtained.

Further, by performing the above operations (4)-(6) on the image c, the redundant area in the image a can be determined, as shown in the image d. Then, the gray value of each pixel in the redundant area can be updated to 0 to obtain the de-redundant processed image e.

According to the above-mentioned embodiment, redundant regions in the first image can be removed, thereby improving the efficiency of subsequent processing and improving the accuracy of matching.

5 is a schematic structural diagram of a mobile robot according to an exemplary embodiment of the present application. Referring to FIG. 5, the mobile robot may include an acquisition module 510, a memory 520, and a processor 530. The collection module 510 can collect the first image in the current field of view. The memory 520 may store sample data at a specified location. The processor 530 may match the first image with the sample data, and determine the specified position where the sample data is successfully matched with the first image as the current location of the mobile robot.

According to an embodiment, the sample data may be a second image of the specified location that is collected in advance at different shooting angles. The processor 530 may match the first image with the sample data by performing the following operations: performing feature extraction on the first image and the second image respectively to obtain the first feature of the first image A descriptor and a second feature descriptor of the second image, and calculating the similarity between the first feature descriptor and the second feature descriptor.

According to another embodiment, the sample data may be a second feature descriptor of a second image, and the second image is an image of the specified position that is pre-collected at different shooting angles. The processor 530 may match the first image with the sample data by performing the following operations: performing feature extraction on the first image, obtaining a first feature descriptor of the first image, and calculating the The similarity between the first feature descriptor and the second feature descriptor.

When the similarity between the first feature descriptor and the second feature descriptor is greater than a preset threshold, the processor 530 may determine that the first feature descriptor and the second feature descriptor match.

Specifically, the processor 530 may perform feature extraction on the image through the following operations:

Performing non-maximum suppression processing on the image to obtain feature points of the image;

For each feature point, divide the specified neighborhood of the feature point into multiple sub-regions, and for each sub-region, calculate the gradient amplitude and gradient direction of each pixel in the sub-region;

Correct the gradient direction of each pixel, so that the corrected gradient direction is within the specified range;

Obtain the feature vector corresponding to the sub-region according to the gradient amplitude of each pixel in the sub-region and the corrected gradient direction, and determine the feature vector corresponding to the feature point according to the feature vector corresponding to each sub-region, and based on The feature vector corresponding to each feature point determines the feature descriptor of the image.

In addition, before matching the first image with the sample data, the processor 530 may perform filter processing, enhancement processing, and / or de-redundancy processing on the first image.

Specifically, the de-redundancy processing performed by the processor 530 on the first image may include: determining a redundant area in the first image; and converting the gray value of each pixel in the redundant area Update to 0 to get the de-redundant image.

Specifically, the processor 530 may determine the redundant area in the first image through the following operations:

Calculating the average value of the gray value of each pixel in the first designated area in the first image;

Obtain the first updated image by subtracting the average value of the gray value of each pixel in the second designated area in the first image;

Updating the gray value of each pixel in the second designated area in the first updated image with a gray value greater than the first preset threshold to 255, and updating the first updated image The gray value of each pixel in the second specified area whose gray value is less than or equal to the first preset threshold is updated to 0, to obtain a second updated image;

For each row area of the second designated area in the second updated image, the proportion of pixels in the row area with a gray value of 255 is counted to obtain the proportion corresponding to the row area;

For the second designated area in the second updated image, when the proportion corresponding to the designated number of consecutive row areas from top to bottom is greater than the second preset threshold, the designated number of consecutive rows is determined The target line number of the last line in the area in the second updated image;

The area between the line indicated by the target line number and the last line in the first image is determined as a redundant area.

In addition, the present application may also provide a computer-readable storage medium on which a computer program is stored, which, when executed by a processor in the mobile robot, causes the mobile robot to realize the above according to the embodiments of the present application Either way.

The above are only some embodiments of the present application, and are not intended to limit the present application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of this application shall be included within the scope of protection of this application.

Claims

A positioning method for a mobile robot, the method is performed by the mobile robot, and includes:

Collect the first image in the current field of view;

Matching the first image with pre-stored sample data at a specified location;

The designated position where the sample data is successfully matched with the first image is determined as the current position of the mobile robot.
The method according to claim 1, wherein

The sample data is a second image of the specified position collected in advance at different shooting angles;

Matching the first image with pre-stored sample data at a specified location includes:

Performing feature extraction on the first image and the second image respectively to obtain a first feature descriptor of the first image and a second feature descriptor of the second image;

The similarity between the first feature descriptor and the second feature descriptor is calculated.
The method according to claim 1, wherein

The sample data is a second feature descriptor of a second image, and the second image is an image of the specified location collected in advance at different shooting angles;

Matching the first image with pre-stored sample data at a specified location includes:

Performing feature extraction on the first image to obtain a first feature descriptor of the first image;

The similarity between the first feature descriptor and the second feature descriptor is calculated.
The method according to claim 2 or 3, wherein the mobile robot performs feature extraction on the first image or the second image through the following operations:

Perform non-maximum suppression processing on the image to obtain the feature points of the image;

For each of the feature points, the designated neighborhood of the feature point is divided into multiple sub-regions, and for each of the sub-regions, the gradient amplitude and gradient direction of each pixel in the sub-region are calculated;

Correct the gradient direction of each pixel point so that the corrected gradient direction is within a specified range;

Obtain the feature vector corresponding to the sub-region according to the gradient amplitude of each pixel in the sub-region and the corrected gradient direction, and determine the feature vector corresponding to the feature point according to the feature vector corresponding to each sub-region, and based on The feature vector corresponding to each feature point determines the feature descriptor of the image.
The method according to any one of claims 1 to 3, further comprising:

Before matching the first image with the pre-stored sample data at the specified position, at least one of filtering processing, enhancement processing, and de-redundancy processing is performed on the first image.
The method according to claim 5, wherein the de-redundancy processing of the first image comprises:

Determine the redundant area in the first image;

The gray value of each pixel in the redundant area is updated to 0 to obtain a de-redundant processed image.
The method according to claim 6, wherein determining the redundant area in the first image comprises:

Calculating the average value of the gray value of each pixel in the first designated area in the first image;

Obtain the first updated image by subtracting the average value of the gray value of each pixel in the second designated area in the first image;

Updating the gray value of each pixel in the second designated area in the first updated image with a gray value greater than the first preset threshold to 255, and updating the first updated image The gray value of each pixel in the second specified area whose gray value is less than or equal to the first preset threshold is updated to 0, to obtain a second updated image;

For each row area of the second designated area in the second updated image, the proportion of pixels in the row area with a gray value of 255 is counted to obtain the proportion corresponding to the row area;

For the second designated area in the second updated image, when the proportion corresponding to the designated number of consecutive row areas from top to bottom is greater than the second preset threshold, the designated number of consecutive rows is determined The target line number of the last line in the area in the second updated image;

The area between the line indicated by the target line number and the last line in the first image is determined as the redundant area.
A mobile robot, the mobile robot includes a collection module, a memory and a processor, wherein,

The collection module is used to collect the first image in the current field of view;

The memory is used to store sample data at a specified location;

The processor is configured to match the first image with the sample data, and determine a specified position where the sample data successfully matches the first image as the current location of the mobile robot.
The mobile robot according to claim 8, wherein

The sample data is a second image of the specified position collected in advance at different shooting angles;

The processor matches the first image with the sample data through the following operations:

Performing feature extraction on the first image and the second image respectively to obtain a first feature descriptor of the first image and a second feature descriptor of the second image;

The similarity between the first feature descriptor and the second feature descriptor is calculated.
The mobile robot according to claim 8, wherein

The sample data is a second feature descriptor of a second image, and the second image is an image of the specified location collected in advance at different shooting angles;

The processor matches the first image with the sample data through the following operations:

Performing feature extraction on the first image to obtain a first feature descriptor of the first image;

The similarity between the first feature descriptor and the second feature descriptor is calculated.
The mobile robot according to claim 9 or 10, wherein the processor performs feature extraction on the first image or the second image through the following operations:

Perform non-maximum suppression processing on the image to obtain the feature points of the image;

For each of the feature points, the designated neighborhood of the feature point is divided into multiple sub-regions, and for each of the sub-regions, the gradient amplitude and gradient direction of each pixel in the sub-region are calculated;

Correct the gradient direction of each pixel point so that the corrected gradient direction is within a specified range;

Obtain the feature vector corresponding to the sub-region according to the gradient amplitude of each pixel in the sub-region and the corrected gradient direction, and determine the feature vector corresponding to the feature point according to the feature vector corresponding to each sub-region, and based on The feature vector corresponding to each feature point determines the feature descriptor of the image.
The mobile robot according to any one of claims 8 to 10, wherein the processor is further used to:

Before matching the first image with the sample data, at least one of filter processing, enhancement processing, and de-redundancy processing is performed on the first image.
The mobile robot according to claim 12, wherein the de-redundancy processing of the first image includes:

Determine the redundant area in the first image;

The gray value of each pixel in the redundant area is updated to 0 to obtain a de-redundant processed image.
The mobile robot according to claim 13, wherein determining the redundant area in the first image includes:

Calculating the average value of the gray value of each pixel in the first designated area in the first image;

Obtain the first updated image by subtracting the average value of the gray value of each pixel in the second designated area in the first image;

Updating the gray value of each pixel in the second designated area in the first updated image with a gray value greater than the first preset threshold to 255, and updating the first updated image The gray value of each pixel in the second specified area whose gray value is less than or equal to the first preset threshold is updated to 0, to obtain a second updated image;

For each row area of the second designated area in the second updated image, the proportion of pixels in the row area with a gray value of 255 is counted to obtain the proportion corresponding to the row area;

For the second designated area in the second updated image, when the proportion corresponding to the designated number of consecutive row areas from top to bottom is greater than the second preset threshold, the designated number of consecutive rows is determined The target line number of the last line in the area in the second updated image;

The area between the line indicated by the target line number and the last line in the first image is determined as the redundant area.
A computer-readable storage medium on which a computer program is stored, which when executed by a processor in a mobile robot, causes the mobile robot to:

Collect the first image in the current field of view;

Matching the first image with pre-stored sample data at a specified location;

The designated position where the sample data is successfully matched with the first image is determined as the current position of the mobile robot.