Blind Path Recognition Method, System, Device and Storage Medium Based on Machine Vision

Abstract

The present invention discloses a blind path recognition method, system, device and storage medium based on machine vision. The method includes: establishing a database of blind path bricks; combining the blind path brick patterns and geographical location coordinates collected in real time with the database of blind path bricks to obtain corresponding blind path information; and broadcasting the obtained blind path information. By establishing a database of blind path bricks in advance and obtaining the blind path brick patterns and geographical location coordinates in real time, the present invention can obtain blind path information containing surrounding environment information, enabling visually impaired people to better perceive the surrounding environment information and greatly facilitating the travel of visually impaired people. Moreover, the implementation of the present invention is simple and convenient, without the need to change any original road foundation. It only requires spraying customized patterns on specific original blind path bricks, and construction personnel can collect and upload data, without considering other additional issues. The present invention can be widely applied in daily life.

Description

Technical Field

[0001] The present invention relates to the technical field of image recognition, and particularly to a blind path recognition method, system, device and storage medium based on machine vision. Background Art

[0002] Currently, blind paths are basically formed by laying ground bricks with fixed shapes on sidewalks (divided into two types: walking strip blind path bricks and prompt dot-shaped blind path bricks), which can generate different foot sensations for visually impaired people, guiding them to walk forward, distinguish directions and reach their destinations.

[0003] However, the information that the current blind path can provide is extremely poor, and it can only prompt the walking information of whether the current blind path is a straight path or a turning fork point. During straight walking, where has one reached; when reaching a fork point, whether to turn left or right, where one is currently located, and where to go after stopping? All these make it still extremely difficult for blind people to travel in reality.

[0004] At the same time, for products such as electronic blind paths and intelligent blind paths that appear in the current market, chips need to be installed under the road and receivers need to be installed on the blind sticks so that they interact to emit signals. Products like this require huge costs to implement, a large amount of manpower to comprehensively transform the current blind paths, and it is very inconvenient to popularize them among blind people. It is also unrealistic to distribute such high-cost blind sticks to blind people. Summary of the Invention

[0005] In order to solve the above technical problems, the purpose of the present invention is to provide a blind path recognition method, system, device and storage medium based on machine vision.

[0006] The technical solution adopted by the present invention is as follows:

[0007] A blind path recognition method based on machine vision includes the following steps:

[0008] Establish a blind path brick database;

[0009] Combine the blind path brick patterns and geographical position coordinates collected in real time with the blind path brick database to obtain corresponding blind path information;

[0010] According to the obtained blind path information, broadcast the blind path information.

[0011] As a further improvement of the above-mentioned blind path recognition method based on machine vision, the step of establishing the blind path brick database specifically includes:

[0012] Collect various blind path brick patterns and size parameters of each blind path brick pattern;

[0013] Bind the geographical location coordinates to be broadcasted, the corresponding blind path brick patterns, the size parameters corresponding to the blind path brick patterns, and the preset blind path information to obtain blind path binding information, and upload the blind path binding information;

[0014] Generate a blind path brick database according to the uploaded blind path binding information.

[0015] As a further improvement of the above-mentioned blind path recognition method based on machine vision, the step of obtaining the corresponding blind path information by combining the real-time acquired blind path brick patterns and geographical location coordinates with the blind path brick database specifically includes:

[0016] Turn on the camera to obtain a video stream;

[0017] Detect and recognize each frame of the image in the video stream to obtain an image of the blind path brick to be detected;

[0018] Call the machine vision tool to recognize the image of the blind path brick to be detected and obtain the corresponding blind path brick pattern;

[0019] Obtain the GPS signal to obtain the current geographical location coordinates;

[0020] According to the obtained blind path brick pattern and the current geographical location coordinates, find the corresponding blind path information in the blind path brick database.

[0021] As a further improvement of the above-mentioned blind path recognition method based on machine vision, the blind path information includes geographical location information corresponding to the geographical location coordinates and blind path brick information corresponding to the blind path brick pattern.

[0022] Another technical solution adopted by the present invention is:

[0023] A blind path recognition system based on machine vision, including:

[0024] A database establishment unit for establishing a blind path brick database;

[0025] A real-time recognition unit for combining the real-time acquired blind path brick patterns and geographical location coordinates with the blind path brick database to obtain the corresponding blind path information;

[0026] A broadcast unit for broadcasting the blind path information according to the obtained blind path information.

[0027] As a further improvement of the above-mentioned blind path recognition system based on machine vision, the database establishment unit specifically includes:

[0028] A data acquisition unit for acquiring various blind path brick patterns and the size parameters of each blind path brick pattern;

[0029] An information binding unit, configured to bind the geographical location coordinates to be reported, the corresponding blind path brick pattern, the size parameters corresponding to the blind path brick pattern, and the preset blind path information to obtain blind path binding information, and upload the blind path binding information;

[0030] A database generation unit, configured to generate a blind path brick database according to the uploaded blind path binding information.

[0031] As a further improvement of the described blind path recognition system based on machine vision, the real-time recognition unit specifically includes:

[0032] A video acquisition unit, configured to turn on a camera to acquire a video stream;

[0033] A video detection unit, configured to perform detection and recognition on each frame of the image in the video stream to obtain an image of a blind path brick to be detected;

[0034] A pattern recognition unit, configured to call machine vision tools to recognize the image of the blind path brick to be detected to obtain the corresponding blind path brick pattern;

[0035] A coordinate acquisition unit, configured to acquire a GPS signal to obtain the current geographical location coordinates;

[0036] A search unit, configured to search for the corresponding blind path information in the blind path brick database according to the obtained blind path brick pattern and the current geographical location coordinates.

[0037] As a further improvement of the described blind path recognition system based on machine vision, the blind path information includes geographical location information corresponding to the geographical location coordinates and blind path brick information corresponding to the blind path brick pattern.

[0038] Another technical solution adopted by the present invention is:

[0039] A blind path recognition device based on machine vision, including:

[0040] A memory, configured to store a program;

[0041] A processor, configured to execute the program, and the program causes the processor to execute the described blind path recognition method based on machine vision.

[0042] Another technical solution adopted by the present invention is:

[0043] A computer storage medium, including a computer program, which when running on a computer, causes the described method to be executed.

[0044] The beneficial effects of the present invention are:

[0045] A method, system, device and storage medium for blind path recognition based on machine vision according to the present invention can obtain blind path information containing surrounding environment information by establishing a blind path brick database in advance and acquiring blind path brick patterns and geographical location coordinates in real time, enabling visually impaired persons to better perceive the surrounding environment information and greatly facilitating the travel of visually impaired persons. Moreover, the implementation of the present invention is simple and convenient, without the need to change any original road foundation. It only requires spraying customized patterns on the original specific blind path bricks, and the construction personnel can collect and upload data, without considering other additional issues. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 is a flowchart of the steps of a method for blind path recognition based on machine vision according to the present invention;

[0047] Figure 2 is a block diagram of the modules of a system for blind path recognition based on machine vision according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] The following further describes the specific embodiments of the present invention with reference to the accompanying drawings:

[0049] Refer to Figure 1 , a method for blind path recognition based on machine vision according to the present invention includes the following steps:

[0050] S1. Establish a blind path brick database;

[0051] S2. Combine the blind path brick patterns and geographical location coordinates collected in real time with the blind path brick database to obtain corresponding blind path information;

[0052] S3. Broadcast the obtained blind path information.

[0053] In this embodiment, the blind path brick patterns can be sprayed on the original blind path bricks or the blind path bricks can be redesigned, so that blind path bricks including blind path brick patterns are provided at the required geographical location coordinates.

[0054] Further as a preferred embodiment, the step of establishing the blind path brick database specifically includes:

[0055] S11. Collect various blind path brick patterns and size parameters of each blind path brick pattern;

[0056] S12. Bind the geographical location coordinates to be broadcast, the corresponding blind path brick pattern, the size parameters corresponding to the blind path brick pattern, and the preset blind path information to obtain blind path binding information, and upload the blind path binding information;

[0057] S13. Generate a blind path brick database according to the uploaded blind path binding information.

[0058] In this embodiment, the implementer can collect and upload the patterns and geographical location coordinates of the blind sidewalk bricks on-site through a mobile phone app. After the blind sidewalk binding information uploaded generates a blind sidewalk brick database, it can be downloaded and updated for use through the relevant interfaces of the app.

[0059] Further as a preferred implementation manner, the blind sidewalk brick patterns and geographical location coordinates collected in real time are combined with the blind sidewalk brick database to obtain corresponding blind sidewalk information. This step specifically includes:

[0060] S21. Turn on the camera to obtain a video stream;

[0061] S22. Detect and identify each frame of the image in the video stream to obtain an image of the blind sidewalk brick to be detected;

[0062] S23. Invoke a machine vision tool to identify the image of the blind sidewalk brick to be detected and obtain the corresponding blind sidewalk brick pattern;

[0063] S24. Obtain the GPS signal to obtain the current geographical location coordinates;

[0064] S25. According to the obtained blind sidewalk brick pattern and the current geographical location coordinates, search in the blind sidewalk brick database to obtain the corresponding blind sidewalk information.

[0065] In this embodiment, to achieve the unique real geographical location positioning when the image recognition is successful, a certain number of blind sidewalk bricks of different styles are required. As long as the GPS signal (the accuracy is generally within 100 meters) is combined to ensure that there are no duplicate blind sidewalk brick pictures within a certain distance range (such as 100 meters), the real geographical location corresponding to the picture can be uniquely determined.

[0066] When used on the iOS platform, the image recognition on the iOS platform is based on the relevant image recognition classes in the Apple system machine vision Kit 1.3 library (running requirements: iOS 11.3+ and iPhone with A9+ chip / iPhone released after 6s). Therefore, in this embodiment, the machine vision tool adopts the machine vision Kit. After pre-installing or collecting photo materials into the code project, when the APP runs, the rear camera is turned on, and then by invoking the methods of relevant classes, the real-time comparison detection of each frame of the picture in the video stream is realized. When a photo is successfully recognized, the name of the picture can be obtained through the relevant classes of the machine vision Kit. At this time, combined with the mobile phone GPS signal, by querying the relevant database, the unique geographical location coordinates corresponding to the current photo in the real world can be obtained, and the current position of the person corresponding to the blind sidewalk brick pattern can be displayed in combination with the map. After the image recognition and positioning are successful, the relevant functions such as searching for surrounding information and navigation of the current position can be realized by combining the mobile phone compass and the map relevant interfaces.

[0067] When applied to the Android platform, the machine vision tool in this embodiment adopts the machine vision Core. With the enhanced image function in the machine vision Core, an application that can respond to 2D images in the user environment is constructed. A set of reference images is provided. When an image is detected in the camera's field of view, the machine vision Core tracks the physical position of these images in the machine vision session and combines the actual geographical location of the images to determine the geographical location of the user (GPS determines the large range, and the precise range is determined according to different images). After obtaining the geographical location of the user, relevant nearby information is pushed (for example: there is a bus stop xx ahead, and the xxx bus is about to arrive...).

[0068] Further as a preferred embodiment, the blind path information includes geographical location information corresponding to geographical location coordinates and blind path brick information corresponding to the blind path brick pattern.

[0069] Reference Figure 2 , a blind path recognition system based on machine vision according to the present invention includes:

[0070] A database establishment unit for establishing a blind path brick database;

[0071] A real-time recognition unit for combining the blind path brick pattern and geographical location coordinates collected in real time with the blind path brick database to obtain corresponding blind path information;

[0072] A broadcast unit for broadcasting the blind path information according to the obtained blind path information.

[0073] Further as a preferred embodiment, the database establishment unit specifically includes:

[0074] A data collection unit for collecting various blind path brick patterns and the size parameters of each blind path brick pattern;

[0075] An information binding unit for binding the geographical location coordinates to be broadcast, the corresponding blind path brick pattern, the size parameters corresponding to the blind path brick pattern, and the preset blind path information to obtain blind path binding information and uploading the blind path binding information;

[0076] A database generation unit for generating a blind path brick database according to the uploaded blind path binding information.

[0077] Further as a preferred embodiment, the real-time recognition unit specifically includes:

[0078] A video acquisition unit for turning on the camera to acquire a video stream;

[0079] A video detection unit for detecting and recognizing each frame of the image in the video stream to obtain an image of the blind path brick to be detected;

[0080] A pattern recognition unit for calling machine vision tools to recognize the image of the blind sidewalk brick to be detected and obtaining the corresponding blind sidewalk brick pattern;

[0081] A coordinate acquisition unit for acquiring GPS signals and obtaining the current geographical location coordinates;

[0082] A search unit for searching for corresponding blind sidewalk information in the blind sidewalk brick database according to the obtained blind sidewalk brick pattern and the current geographical location coordinates.

[0083] As a further preferred embodiment, the blind sidewalk information includes geographical location information corresponding to the geographical location coordinates and blind sidewalk brick information corresponding to the blind sidewalk brick pattern.

[0084] The present invention further includes a blind sidewalk recognition device based on machine vision, which includes:

[0085] A memory for storing programs;

[0086] A processor for executing the program, and the program causes the processor to execute the blind sidewalk recognition method based on machine vision.

[0087] The present invention further includes a computer storage medium, which includes a computer program, and when it runs on a computer, the method is executed.

[0088] As can be seen from the above, the present invention can obtain blind sidewalk information containing surrounding environment information by establishing a blind sidewalk brick database in advance and real-time acquiring the blind sidewalk brick pattern and geographical location coordinates, which can enable visually impaired people to better perceive the surrounding environment information and greatly facilitate the travel of visually impaired people. Moreover, the implementation of the present invention is simple and convenient, without changing any original road foundation. Only need to spray custom patterns on the original specific blind sidewalk bricks, and the construction personnel can collect and upload data, without considering other additional problems.

[0089] For visually impaired individuals, the present invention does not require purchasing any accessories. They only need to download an app on their mobile phones to use it. The operation is simple, and the learning cost and time cost are relatively low, enabling them to obtain effective information at important positions and key nodes.

[0090] For the development of cities and society, the present invention can form a unified national or urban blind brick standard, formulate urbanization spray patterns, and form a unified and beautiful digital blind sidewalk customization.

[0091] The above is a specific description of the preferred embodiment of the present invention. However, the present invention is not limited to the described embodiment. Those skilled in the art can make various equivalent deformations or substitutions without departing from the spirit of the present invention, and these equivalent deformations or substitutions are all included within the scope defined by the claims of this application.

Claims

1. A method for blind path recognition based on machine vision, characterized in that, it includes the following steps: Establish a database of blind path bricks; Combine the blind path brick patterns and geographical location coordinates collected in real time with the database of blind path bricks to obtain corresponding blind path information; According to the obtained blind path information, broadcast the blind path information; The step of combining the blind path brick patterns and geographical location coordinates collected in real time with the database of blind path bricks to obtain corresponding blind path information specifically includes: Turn on the camera to obtain a video stream; Detect and recognize each frame of the image in the video stream to obtain an image of the blind path brick to be detected; Call the machine vision tool to recognize the image of the blind path brick to be detected and obtain the corresponding blind path brick pattern; Obtain the GPS signal to obtain the current geographical location coordinates; According to the obtained blind path brick pattern and the current geographical location coordinates, search in the database of blind path bricks to obtain the corresponding blind path information; The step of establishing the database of blind path bricks specifically includes: Collect various blind path brick patterns and the size parameters of each blind path brick pattern; Bind the geographical location coordinates to be broadcast, the corresponding blind path brick pattern, the size parameters corresponding to the blind path brick pattern, and the preset blind path information to obtain blind path binding information, and upload the blind path binding information; Generate a database of blind path bricks according to the uploaded blind path binding information.

2. A method for blind path recognition based on machine vision according to claim 1, characterized in that: The blind path information includes geographical location information corresponding to the geographical location coordinates and blind path brick information corresponding to the blind path brick pattern.

3. A blind path recognition system based on machine vision, characterized in that, it includes: A database establishment unit for establishing a database of blind path bricks; A real-time recognition unit for combining the blind path brick patterns and geographical location coordinates collected in real time with the database of blind path bricks to obtain corresponding blind path information; A broadcast unit for broadcasting the blind path information according to the obtained blind path information; The real-time recognition unit specifically includes: A video acquisition unit for turning on the camera to obtain a video stream; A video detection unit for detecting and recognizing each frame of the image in the video stream to obtain an image of the blind path brick to be detected; a pattern recognition unit for calling the machine vision tool to recognize the image of the blind path brick to be detected and obtain the corresponding blind path brick pattern; A coordinate acquisition unit for obtaining the GPS signal to obtain the current geographical location coordinates; A search unit for searching in the database of blind path bricks according to the obtained blind path brick pattern and the current geographical location coordinates to obtain the corresponding blind path information; The database establishment unit specifically includes: A data acquisition unit for collecting various blind path brick patterns and the size parameters of each blind path brick pattern; An information binding unit for binding the geographical location coordinates to be broadcast, the corresponding blind path brick pattern, the size parameters corresponding to the blind path brick pattern, and the preset blind path information to obtain blind path binding information, and uploading the blind path binding information; A database generation unit for generating a database of blind path bricks according to the uploaded blind path binding information.

4. A blind path recognition system based on machine vision according to claim 3, characterized in that: The blind path information includes geographical location information corresponding to geographical location coordinates and blind path brick information corresponding to the blind path brick pattern.

5. A blind path recognition device based on machine vision, characterized in that, it includes: a memory for storing programs; a processor for executing the programs, and the programs cause the processor to execute the blind path recognition method based on machine vision as described in claim 1 or 2.

6. A computer storage medium, characterized in that, it includes a computer program, and when it runs on a computer, it causes the method as described in claim 1 or 2 to be executed.

Images