Indoor navigation method and system based on multi-view fusion

Abstract

The application provides an indoor navigation method and system based on multi-view fusion, relating to the technical field of indoor navigation. The method comprises: loading map information of a target indoor space, then deploying and configuring a CCD image acquisition device, then starting the CCD image acquisition device to acquire indoor multi-view information through multi-view acquisition, then performing image processing and extraction to obtain image extraction features, connecting a client to obtain indoor navigation requirement information, adding the features to the map information, performing path planning, and obtaining indoor navigation planning results. The application mainly solves the problem that traditional outdoor positioning technology has obvious limitations in indoor environments, and its accuracy and stability often cannot reach the ideal level. Through multi-view fusion technology, more rich environmental information can be obtained, and the accuracy and reliability of positioning and navigation can be improved. More accurate and reliable positioning and navigation information is provided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of indoor navigation, in particular to an indoor navigation method and system based on multi-view fusion. BACKGROUND

[0002] With the popularity of smart phones and the rapid development of mobile devices, people's demand for navigation in indoor environments is becoming higher and higher. For example, tourists in large indoor places such as shopping malls, airports, and museums need accurate navigation services to find their desired destinations. In addition, indoor navigation technology also has broad application prospects in the fields of industrial manufacturing, logistics distribution, and unmanned driving. However, indoor environments are more complex than outdoor environments, with more occlusions, signal interference, and other problems, so indoor positioning technology faces greater challenges. Currently, mainstream indoor positioning technologies include WiFi positioning, Bluetooth positioning, magnetic field strength positioning, and radio signal positioning, but they all have problems such as insufficient accuracy, insufficient stability, and high cost.

[0003] However, in the process of implementing the technical solutions of the embodiments of the present application, it was found that the above-mentioned technologies at least have the following technical problems:

[0004] Traditional outdoor positioning technology has obvious limitations in indoor environments, and its accuracy and stability often cannot reach the ideal level. SUMMARY

[0005] The present application mainly solves the problem that traditional outdoor positioning technology has obvious limitations in indoor environments, and its accuracy and stability often cannot reach the ideal level.

[0006] In view of the above problems, the present application provides an indoor navigation method and system based on multi-view fusion. In a first aspect, the present application provides an indoor navigation method based on multi-view fusion, which comprises: loading indoor map information of a target indoor space, the indoor map information comprising indoor path information and indoor obstacle position information; based on the indoor map information, arranging CCD image acquisition devices in the target indoor space, the CCD image acquisition devices being uniformly distributed inside the target indoor space; after the CCD image acquisition devices are configured, starting the CCD image acquisition devices to perform multi-view acquisition and obtain indoor multi-view information, wherein the indoor multi-view information comprises target surface size and imaging size of the CCD image acquisition devices; based on the indoor multi-view information, performing image processing extraction to obtain image extraction features, the image extraction features comprising color indicators, texture indicators, and shape indicators; connecting a client to obtain indoor navigation requirement information, the indoor navigation requirement information comprising indoor navigation starting point position coordinates and indoor navigation ending point position coordinates; adding the image extraction features to the indoor map information, and then performing path planning according to the indoor navigation requirement information to obtain indoor navigation planning results, the indoor navigation planning results comprising a dynamically updated path.

[0007] In a second aspect, the present application provides an indoor navigation system based on multi-view fusion, which comprises: an indoor map information acquisition module for loading indoor map information of a target indoor space, the indoor map information comprising indoor path information and indoor obstacle position information; an image acquisition device arrangement module for arranging CCD image acquisition devices in the target indoor space based on the indoor map information, the CCD image acquisition devices being uniformly distributed inside the target indoor space; an indoor multi-view information acquisition module for configuring the CCD image acquisition devices and then starting the CCD image acquisition devices to perform multi-view acquisition and obtain indoor multi-view information, wherein the indoor multi-view information comprises target surface size and imaging size of the CCD image acquisition devices; a graphic feature extraction module for performing image processing extraction based on the indoor multi-view information to obtain image extraction features, the image extraction features comprising color indicators, texture indicators, and shape indicators; a requirement information acquisition module for connecting a client to obtain indoor navigation requirement information, the indoor navigation requirement information comprising indoor navigation starting point position coordinates and indoor navigation ending point position coordinates; and a navigation planning result acquisition module for adding the image extraction features to the indoor map information, and then performing path planning according to the indoor navigation requirement information to obtain indoor navigation planning results, the indoor navigation planning results comprising a dynamically updated path.

[0008] The one or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0009] The present application provides an indoor navigation method and system based on multi-view fusion, relating to the technical field of indoor navigation. The method comprises: loading map information of a target indoor space, then deploying a CCD image acquisition device and configuring it, then starting the CCD image acquisition device to acquire indoor multi-view information through multi-view acquisition, then performing graph processing and extraction to obtain image extraction features, connecting a client to obtain indoor navigation requirement information, adding the features to the map information, performing path planning, and obtaining indoor navigation planning results.

[0010] The present application mainly solves the problem that traditional outdoor positioning technology has obvious limitations in indoor environments, and its accuracy and stability often cannot reach the ideal level. Through multi-view fusion technology, more rich environmental information can be obtained to improve the accuracy and reliability of positioning and navigation. More accurate and reliable positioning and navigation information is provided.

[0011] The above description is only a summary of the technical solutions of the present application. In order to more clearly understand the technical means of the present application, the content of the specification can be implemented, and in order to make the above and other purposes, characteristics and advantages of the present application more obvious and easy to understand, the specific embodiments of the present application are described below. BRIEF DESCRIPTION OF DRAWINGS

[0012] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only exemplary, and for those skilled in the art, other drawings can be obtained without creative labor on the basis of the provided drawings.

[0013] Figure 1 A flowchart of an indoor navigation method based on multi-view fusion is provided for the embodiments of the present application;

[0014] Figure 2 A flowchart of a path planning adjustment method in an indoor navigation method based on multi-view fusion is provided for the embodiments of the present application;

[0015] Figure 3 A flowchart of a method for updating effective road segments in a path in an indoor navigation method based on multi-view fusion is provided for the embodiments of the present application;

[0016] Figure 4 A structural diagram of an indoor navigation system based on multi-view fusion is provided for the embodiments of the present application.

[0017] The reference signs are explained as follows: indoor map information acquisition module 10, image acquisition device arrangement module 20, indoor multi-view information acquisition module 30, graphic feature extraction module 40, demand information acquisition module 50, and navigation planning result acquisition module 60. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0019] The present application mainly solves the problem that the traditional outdoor positioning technology has obvious limitations in indoor environments, and its accuracy and stability often cannot reach the ideal level. Through multi-view fusion technology, more rich environmental information can be obtained to improve the accuracy and reliability of positioning and navigation. More accurate and reliable positioning and navigation information is provided.

[0020] In order to better understand the above technical solutions, the above solutions will be described in detail below with reference to the drawings in the specification and specific embodiments:

[0021] Embodiment One

[0022] As shown in the indoor navigation method based on multi-view fusion, the method comprises: Figure 1

[0023] Loading indoor map information of a target indoor space, the indoor map information comprising indoor path information and indoor obstacle position information;

[0024] Specifically, to load the indoor map information of the target indoor space, first, these information is acquired. These information comes from the prior knowledge of indoor map, or may come from the perception and detection of indoor environment. For example, the indoor environment can be modeled by using prior knowledge or machine learning method, so as to acquire the indoor map information. Once the indoor map information is obtained, it can be loaded into the navigation system for realizing indoor navigation. These information can include indoor path information, indoor obstacle position information, etc., which can be used to construct the model of indoor scene, and also can be used for path planning, obstacle avoidance, etc. When loading the indoor map information, the complexity and dynamics of indoor environment need to be considered. For example, the obstacles in indoor environment can move or change, so the indoor map information needs to be updated in real time. In addition, the accuracy and reliability of data need to be considered to ensure the accuracy and stability of navigation.

[0025] ​Based on the indoor map information, CCD image acquisition devices are arranged in the target indoor space, and the CCD image acquisition devices are uniformly distributed inside the target indoor space.

[0026] Specifically, based on the indoor map information, CCD image acquisition devices can be arranged in the target indoor space, and these devices can be uniformly distributed inside the entire target indoor space. The CCD image acquisition device can be a camera or a sensor, which can be used to collect images or data of the target indoor space. By uniformly distributing the CCD image acquisition devices inside the target indoor space, more abundant spatial information can be obtained, thereby improving the positioning accuracy and stability. The number and distribution range of devices: the number and distribution range of CCD image acquisition devices need to be determined according to the size and shape of the target indoor space. The number and distribution range of devices should be able to cover the entire space to ensure that sufficient data information is obtained. The installation height of the device: the installation height of the CCD image acquisition device should be determined according to the characteristics of the target indoor space. In some cases, the device needs to be installed at a certain height to be able to cover the entire space. The accuracy and stability of the device: the accuracy and stability of the CCD image acquisition device are very important for the accuracy and reliability of indoor navigation. Therefore, high-quality devices need to be selected and regularly maintained and calibrated. The power consumption and transmission speed of the device: the power consumption and transmission speed of the device need to be considered to ensure that the speed of data acquisition and processing can meet the needs of real-time navigation. Arranging CCD image acquisition devices in the target indoor space based on indoor map information can improve the positioning accuracy and stability, while considering the number, distribution range, installation height, accuracy and stability of the device.

[0027] After the CCD image acquisition devices are configured, the CCD image acquisition devices are started for multi-view acquisition to obtain indoor multi-view information, wherein the indoor multi-view information includes the target surface size and imaging size of the CCD image acquisition device.

[0028] Specifically, after configuring the CCD image acquisition devices, these devices can be started for multi-view acquisition to obtain indoor multi-view information. This information includes the target surface size and imaging size of the CCD image acquisition devices. The target surface size refers to the size of the photosensitive surface of the CCD image acquisition device, which determines the range and resolution of the images that can be acquired. The imaging size refers to the distance from the object surface to the target surface of the CCD image acquisition device, which determines the magnification and depth of field of the image. When performing multi-view acquisition, the relative positions and angle relationships between the CCD image acquisition devices need to be considered to ensure more accurate spatial information. In addition, the acquired images or data need to be processed and analyzed to extract useful features and information, such as the positions of obstacles, path information, etc. By configuring the CCD image acquisition devices and starting them for multi-view acquisition, indoor multi-view information can be obtained, including target surface size and imaging size, etc. These information can be used to achieve more accurate and stable indoor navigation.

[0029] Based on the indoor multi-view information, perform graph processing extraction to obtain image extraction features, including color indicators, texture indicators, and shape indicators.

[0030] Specifically, based on the indoor multi-view information, graph processing extraction can be performed to obtain image extraction features, including color indicators, texture indicators, and shape indicators, etc. The color indicator refers to the color information in the image, which can be used to determine the material, surface state, and other characteristics of the object. The texture indicator refers to the texture information in the image, which can be used to determine the surface texture, structure, and other characteristics of the object. The shape indicator refers to the contour and shape information of the object in the image, which can be used to determine the shape and size of the object. In performing graph processing extraction, various computer vision and image processing techniques can be used, such as feature extraction, image segmentation, edge detection, etc. Through these techniques, the features in the image can be extracted and further processed and analyzed.

[0031] Connect to the client to obtain indoor navigation demand information, including indoor navigation starting point position coordinates and indoor navigation ending point position coordinates.

[0032] Specifically, first, a server or application needs to be established to receive connection requests from clients. This server or application can be a standalone system or an integral part of the indoor navigation system. Once the connection is established, the user's indoor navigation request information can be obtained through the client. This can be achieved using a web application or mobile application to receive the user's input navigation request information. In this case, the user can input the coordinates of the starting and ending points, as well as other possible navigation request information, such as route preferences and traffic conditions, through a browser or application interface. Alternatively, indoor navigation request information can be received through an API interface. In this case, the client can be another application or system that sends navigation request information to the indoor navigation system by calling the API. The API can be a web-based service interface such as RESTful or SOAP, or other types of interfaces.

[0033] The extracted features from the image are added to the indoor map information. Then, based on the indoor navigation demand information, path planning is performed to obtain the indoor navigation planning result, which includes dynamically updated paths.

[0034] Specifically, by adding image-extracted features to indoor map information, path planning can be performed based on indoor navigation requirements, resulting in an indoor navigation plan, including dynamically updated paths. During path planning, image processing techniques and algorithms can be used to calculate the optimal path from the starting point coordinates to the ending point coordinates. These algorithms can include Dijkstra's algorithm, A* algorithm, etc., which can calculate the optimal path based on information such as distances between nodes and obstacles. When calculating the path, image-extracted features and other indoor map information need to be considered. For example, color indicators can be used to determine the presence of obstacles of a specific color, or shape indicators can be used to determine the presence of passages of a specific shape. This information can be used to optimize path planning and avoid obstacles. After obtaining the indoor navigation plan result, its dynamically updated path needs to be added to the indoor map information. This can be achieved by updating the path information in the indoor map. When updating the path, actual conditions and changes need to be considered, such as the movement or change of obstacles; therefore, the indoor map information needs to be updated in real time. Adding image-extracted features to indoor map information and performing path planning based on indoor navigation requirements yields indoor navigation plan results, including dynamically updated paths. This can provide important support and assistance for achieving more accurate and stable indoor navigation.

[0035] Furthermore, such as Figure 2 As shown, in the method of this application, the CCD image acquisition devices are uniformly distributed within the target indoor space, and the method further includes:

[0036] Based on the CCD image acquisition device, download historical indoor multi-view information, the historical indoor multi-view information includes acquisition time mark;

[0037] Compare the historical indoor multi-view information with the indoor multi-view information to obtain an information comparison result, the information comparison result includes an indoor path comparison result and an indoor obstacle position comparison result;

[0038] If the information comparison result is, erase the dynamic update path in the indoor navigation planning result, and re-plan the path.

[0039] Specifically, based on the CCD image acquisition device, historical indoor multi-view information can be downloaded, which includes acquisition time mark. By comparing historical indoor multi-view information with current indoor multi-view information, information comparison result can be obtained, which includes indoor path comparison result and indoor obstacle position comparison result. If the information comparison result shows that the indoor path or obstacle position has changed, the dynamic update path in the indoor navigation planning result can be erased, and the path planning adjustment can be re-performed. This can be achieved by comparing historical indoor multi-view information and current indoor multi-view information. If the indoor path comparison result is changed, the original path can be erased, and the path planning can be re-performed. If the indoor obstacle position comparison result is changed, the original path can also be erased, and the path planning can be re-performed. When re-planning the path, graphical processing techniques and algorithms can be used to calculate the new optimal path. By downloading historical indoor multi-view information and comparing, information comparison result can be obtained. If the result shows that the path or obstacle position has changed, the dynamic update path in the indoor navigation planning result can be erased and the path planning adjustment can be re-performed. This can provide important support and help for more accurate and stable indoor navigation.

[0040] Further, as shown in the method of the present application, the dynamic update path in the indoor navigation planning result is erased, and the path planning adjustment is re-performed, the method comprising: Figure 3

[0041] Determine the indoor path change position area and the indoor obstacle change position area through the information comparison result;

[0042] Place the dynamic update path in the bottom layer, place the indoor obstacle change position area in the middle layer, and place the indoor path change position area in the top layer to determine the invalid section of the dynamic update path;

[0043] Erase the invalid section of the dynamic update path in the indoor navigation planning result, retain the effective section of the dynamic update path, and re-plan the path planning adjustment using the path planning algorithm. ​

[0044] Specifically, by the information comparison result, the indoor path change position area and the indoor obstacle change position area can be determined. These areas can be used to determine the invalid road segments in the dynamic updated path. The dynamic updated path can be placed in the bottom layer, the indoor obstacle change position area can be placed in the middle layer, and the indoor path change position area can be placed in the top layer. In this way, the changes of the path and the positions of the obstacles can be more clearly seen. Then, the invalid road segments in the dynamic updated path can be determined, which can be due to the movement of the obstacles or other changes and become unfeasible or impassable. After determining the invalid road segments, these road segments can be erased, and the valid road segments in the dynamic updated path can be retained. Then, the path planning algorithm can be used to re-adjust the path planning. This can be achieved by calculating a new optimal path or re-planning a known path. By the information comparison result, the indoor path change position area and the indoor obstacle change position area can be determined, and the invalid road segments in the dynamic updated path can be determined. Then, these invalid road segments can be erased, the valid road segments can be retained, and the path planning adjustment can be re-performed. This can provide important support and help for achieving more accurate and stable indoor navigation.

[0045] Further, the method of the present application further comprises:

[0046] using a path planning algorithm to obtain M planning paths, wherein M≥2 and M is a positive integer;

[0047] taking the time consumption of each path as a first evaluation index, evaluating the M planning paths to obtain M first evaluation indexes;

[0048] taking the length of each path as a second evaluation index, evaluating the M planning paths to obtain M second evaluation indexes;

[0049] taking the comfort level of each path as a third evaluation index, evaluating the M planning paths to obtain M third evaluation indexes;

[0050] comprehensively evaluating the M first evaluation indexes, the M second evaluation indexes, and the M third evaluation indexes to obtain an optimal updated path with the highest comprehensive evaluation index.

[0051] Specifically, M planning paths can be obtained by using a path planning algorithm, where M is greater than or equal to 2 and M is a positive integer. These paths can be calculated based on different starting points, ending points or path preference conditions. For each path, the time consumption can be taken as the first evaluation index, the path length can be taken as the second evaluation index, and the comfort degree can be taken as the third evaluation index. These indexes can be used for comprehensive evaluation of the paths. The first evaluation index can be obtained by calculating the time consumption of each path. The time consumption can be calculated by measuring the distance and speed in the path. Similarly, the length of each path can be calculated as the second evaluation index. For the comfort degree index, the turning, speed change and the like in the path can be evaluated. After obtaining the evaluation index of each index, the optimal update path can be determined by a certain comprehensive evaluation method. For example, the indices of each index can be weighted and summed, or other comprehensive evaluation methods can be used to determine the optimal path. Finally, the optimal update path with the highest comprehensive evaluation index is obtained. By obtaining multiple planning paths by using a path planning algorithm and using multiple indexes for comprehensive evaluation, the optimal update path can be obtained. This can provide important support and help for realizing more accurate and stable indoor navigation.

[0052] Further, in the method of the present application, the comfort degree of each path is taken as the third evaluation index to evaluate the M planning paths to obtain M third evaluation indices, the method comprising:

[0053] obtaining a comfort degree evaluation index of the client, the comfort degree evaluation index comprising a path width index and a path smoothness index;

[0054] on each path in the M planning paths, obtaining M sample point sets according to a preset sampling frequency, the sample point sets comprising a first road sampling point set and a second road sampling point set;

[0055] evaluating the M sample point sets in sequence by using the comfort degree evaluation index to obtain M third evaluation indices.

[0056] Specifically, after obtaining M planning paths, comfort evaluation indexes of the client can be acquired, including path width indexes and path smoothness indexes. These indexes can be used to evaluate the comfort degree of the paths. Next, M sample point sets can be acquired on each planning path according to a preset sampling frequency. These sample point sets can include a first road following sample point set and a second road following sample point set. These sample point sets can be used to evaluate the characteristics and attributes of the paths. Then, comfort evaluations can be performed on the M sample point sets in sequence through the comfort evaluation indexes, obtaining M third evaluation indexes. These indexes can be used to comprehensively evaluate the comfort degree of the paths. During the comfort evaluation process, the path width indexes and the path smoothness indexes need to be considered. The path width indexes can reflect the spaciousness and safety of the paths, while the path smoothness indexes can reflect the flatness and driving comfort of the paths. Through the evaluation of these indexes, more accurate and reliable comfort evaluation results can be obtained. By acquiring the comfort evaluation indexes of the client and performing comfort evaluations on the M sample point sets, more accurate and reliable comfort evaluation results can be obtained. This can provide important support and help for realizing more accurate and stable indoor navigation.

[0057] Further, the method of the present application, based on the indoor multi-view information, performs graph processing extraction to obtain image extraction features, including color indexes, texture indexes, and shape indexes. The method includes:

[0058] The indoor multi-view information is preprocessed to obtain indoor multi-view preprocessing information.

[0059] Based on the indoor multi-view preprocessing information, color indexes are extracted using a color histogram method to obtain color index extraction results. Based on the indoor multi-view preprocessing information, texture indexes are extracted using a gray level co-occurrence matrix method to obtain texture index extraction results. Based on the indoor multi-view preprocessing information, shape indexes are extracted using an edge detection algorithm to obtain shape index extraction results.

[0060] The color index extraction results, texture index extraction results, and shape index extraction results are fused to obtain the image extraction features.

[0061] Specifically, indoor multi-view information can be pre-processed to obtain indoor multi-view pre-processed information. Pre-processing can include noise removal, image enhancement, image restoration, etc. to improve the quality and reliability of the image. Based on the indoor multi-view pre-processed information, different methods can be used to extract the features of the image. For the extraction of color indicators, the color histogram method can be used. This method can count the frequency of different colors of pixel points in the image, thereby obtaining the feature extraction result of the color distribution. For the extraction of texture indicators, the gray level co-occurrence matrix method can be used. This method can count the distribution of gray values of adjacent pixel points in the image, thereby obtaining the texture feature information. For the extraction of shape indicators, an edge detection algorithm can be used. The edge detection algorithm can detect the edge information in the image, i.e. the contour of the object and the boundary between different regions. Through processing and analysis of the edge information, the shape feature information can be obtained. Finally, the color indicator extraction result, the texture indicator extraction result and the shape indicator extraction result can be fused to obtain the image extraction feature. Feature fusion can combine and integrate the information of different features to obtain a more comprehensive and accurate image feature representation. Through pre-processing and feature extraction of indoor multi-view information, image extraction features can be obtained. These features can be used to implement various tasks such as target detection, object recognition, scene understanding, etc. The process of feature extraction and fusion can improve the accuracy and reliability of image processing, providing important support and help for realizing more accurate and stable indoor navigation.

[0062] Embodiment Two

[0063] Based on the same inventive concept as the aforementioned embodiment of the indoor navigation method based on multi-view fusion, as shown in Figure 4 The present application provides an indoor navigation system based on multi-view fusion, which comprises:

[0064] An indoor map information acquisition module 10 is used to load the indoor map information of the target indoor space, which includes indoor path information and indoor obstacle position information.

[0065] An image acquisition device arrangement module 20 is used to arrange CCD image acquisition devices in the target indoor space based on the indoor map information, and the CCD image acquisition devices are uniformly distributed inside the target indoor space.

[0066] An indoor multi-view information acquisition module 30 is used to configure the CCD image acquisition devices and start the CCD image acquisition devices for multi-view acquisition to obtain indoor multi-view information, wherein the indoor multi-view information includes the target surface size and imaging size of the CCD image acquisition devices.

[0067] a graphics feature extraction module 40 configured to perform graphics processing extraction based on the indoor multi-view information to obtain image extraction features, the image extraction features including color indicators, texture indicators, and shape indicators;

[0068] a demand information acquisition module 50 configured to connect to a client to acquire indoor navigation demand information, the indoor navigation demand including indoor navigation starting point coordinates and indoor navigation ending point coordinates;

[0069] a navigation planning result acquisition module 60 configured to add the image extraction features to the indoor map information, and then perform path planning according to the indoor navigation demand information to obtain an indoor navigation planning result, the indoor navigation planning result including a dynamically updated path.

[0070] Further, the system further includes:

[0071] a path planning adjustment module configured to download historical indoor multi-view information based on the CCD image acquisition device, the historical indoor multi-view information including acquisition time markers; compare the historical indoor multi-view information with the indoor multi-view information to obtain an information comparison result, the information comparison result including indoor path comparison results and indoor obstacle position comparison results; and if the information comparison result is that the dynamically updated path in the indoor navigation planning result is deleted and path planning adjustment is re-performed.

[0072] Further, the system further includes:

[0073] an effective path segment update module configured to determine indoor path change position areas and indoor obstacle change position areas through the information comparison result; place the dynamically updated path in a bottom layer, place the indoor obstacle change position areas in a middle layer, place the indoor path change position areas in a top layer, and determine invalid path segments of the dynamically updated path; delete the invalid path segments of the dynamically updated path in the indoor navigation planning result, retain effective path segments in the dynamically updated path, and re-perform path planning adjustment using a path planning algorithm.

[0074] Further, the system further includes:

[0075] The comprehensive evaluation module is configured to: obtain M planning paths by using a path planning algorithm, where M is greater than or equal to 2 and M is a positive integer; take time consumption of each path as a first evaluation index, and evaluate the M planning paths to obtain M first evaluation indexes; take length of each path as a second evaluation index, and evaluate the M planning paths to obtain M second evaluation indexes; take comfort level of each path as a third evaluation index, and evaluate the M planning paths to obtain M third evaluation indexes; and perform comprehensive evaluation by using the M first evaluation indexes, the M second evaluation indexes, and the M third evaluation indexes to obtain an optimal update path with the highest comprehensive evaluation index.

[0076] Further, the system further comprises:

[0077] The comfort level evaluation module is configured to: obtain a comfort level evaluation index of the client, the comfort level evaluation index including a path width index and a path smoothness index; obtain M sample point sets on each path in the M planning paths according to a preset sampling frequency, the sample point sets including a first path direction sample point set and a second path direction sample point set; and perform comfort level evaluation on the M sample point sets in sequence by using the comfort level evaluation index to obtain M third evaluation indexes.

[0078] Further, the system further comprises:

[0079] The image extraction feature obtaining module is configured to: pre-process the indoor multi-view information to obtain indoor multi-view pre-processed information; extract a color index by using a color histogram based on the indoor multi-view pre-processed information to obtain a color index extraction result; extract a texture index by using a gray level co-occurrence matrix based on the indoor multi-view pre-processed information to obtain a texture index extraction result; extract a shape index by using an edge detection algorithm based on the indoor multi-view pre-processed information to obtain a shape index extraction result; and perform feature fusion on the color index extraction result, the texture index extraction result, and the shape index extraction result to obtain the image extraction feature.

[0080] The person skilled in the art can clearly know the indoor navigation protection system based on multi-view fusion in the embodiment according to the foregoing detailed description of the indoor navigation method based on multi-view fusion. For the system disclosed in the embodiment, the description is relatively simple because the system corresponds to the device disclosed in the embodiment, and the related parts are described in the method part.

[0081] The foregoing description of the disclosed embodiments enables a person skilled in the art to make or use the application. Modifications of these embodiments will occur to persons of skill in the art, and that the generic principles defined herein can be applied to other embodiments without departing from the spirit or scope of the application. Therefore, the present application is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for indoor navigation based on multi-view fusion, characterized in that, The method comprises: loading indoor map information of a target indoor space, the indoor map information comprising indoor path information and indoor obstacle position information; based on the indoor map information, deploying CCD image acquisition devices in the target indoor space, the CCD image acquisition devices being uniformly distributed inside the target indoor space; after the CCD image acquisition devices are configured, starting the CCD image acquisition devices to perform multi-view acquisition and obtaining indoor multi-view information, wherein the indoor multi-view information comprises target surface size and imaging size of the CCD image acquisition devices; based on the indoor multi-view information, performing graph processing extraction to obtain image extraction features, the image extraction features comprising color indicators, texture indicators, and shape indicators; connecting a client to obtain indoor navigation demand information, the indoor navigation demand information comprising indoor navigation starting point position coordinates and indoor navigation ending point position coordinates; adding the image extraction features to the indoor map information, and then performing path planning according to the indoor navigation demand information to obtain indoor navigation planning results, the indoor navigation planning results comprising a dynamically updated path; erasing the dynamically updated path in the indoor navigation planning results and re-performing path planning adjustment, the method comprising: determining indoor path variation position areas and indoor obstacle variation position areas through information comparison results; placing the dynamically updated path at a bottom layer, placing the indoor obstacle variation position areas at a middle layer, placing the indoor path variation position areas at a top layer, and determining invalid path segments of the dynamically updated path; erasing the invalid path segments in the dynamically updated path in the indoor navigation planning results, retaining valid path segments in the dynamically updated path, and re-performing path planning adjustment using a path planning algorithm; the method further comprises: obtaining M planning paths using a path planning algorithm, wherein M≥2 and M is a positive integer; taking time consumption of each path as a first evaluation indicator to evaluate the M planning paths to obtain M first evaluation indexes; taking length of each path as a second evaluation indicator to evaluate the M planning paths to obtain M second evaluation indexes; taking comfort level of each path as a third evaluation indicator to evaluate the M planning paths to obtain M third evaluation indexes; performing comprehensive evaluation through the M first evaluation indexes, the M second evaluation indexes, and the M third evaluation indexes to obtain an optimal updated path with the highest comprehensive evaluation index; taking comfort level of each path as a third evaluation indicator to evaluate the M planning paths to obtain M third evaluation indexes, the method comprising: obtaining a comfort level evaluation indicator of a client, the comfort level evaluation indicator comprising a path width indicator and a path smoothness indicator; on each path in the M planning paths, obtaining M sample point sets according to a preset sampling frequency, the sample point sets comprising a first path following sample point set and a second path following sample point set; sequentially performing comfort level evaluation on the M sample point sets through the comfort level evaluation indicator to obtain M third evaluation indexes.

2. The multi-view fusion based indoor navigation method of claim 1, wherein, The CCD image acquisition devices are uniformly distributed inside the target indoor space, and the method further comprises: Based on the CCD image acquisition device, download historical indoor multi-view information, the historical indoor multi-view information includes acquisition time mark; Compare the historical indoor multi-view information with the indoor multi-view information to obtain an information comparison result, the information comparison result includes an indoor path comparison result and an indoor obstacle position comparison result; If the information comparison result is, delete the dynamic update path in the indoor navigation planning result and re-plan the path. 3.The multi-view fusion based indoor navigation method of claim 1, wherein, Based on the indoor multi-view information, perform image processing extraction to obtain image extraction features, the image extraction features include color indicators, texture indicators, and shape indicators, and the method comprises: Pretreat the indoor multi-view information to obtain indoor multi-view pretreated information; Based on the indoor multi-view pretreated information, extract color indicators in the form of a color histogram to obtain a color indicator extraction result; based on the indoor multi-view pretreated information, extract texture indicators in the form of a gray level co-occurrence matrix to obtain a texture indicator extraction result; and based on the indoor multi-view pretreated information, extract shape indicators by using an edge detection algorithm to obtain a shape indicator extraction result; Fuse the color indicator extraction result, the texture indicator extraction result, and the shape indicator extraction result to obtain the image extraction features.

4. An indoor navigation system based on multi-view fusion for performing the method of claim 1, characterized by The system comprises: An indoor map information acquisition module, which is configured to load indoor map information of a target indoor space, the indoor map information including indoor path information and indoor obstacle position information; An image acquisition device arrangement module, which is configured to arrange CCD image acquisition devices in the target indoor space based on the indoor map information, the CCD image acquisition devices being uniformly distributed inside the target indoor space; An indoor multi-view information acquisition module, which is configured to start the CCD image acquisition devices to perform multi-view acquisition after the CCD image acquisition devices are configured, and acquire indoor multi-view information, wherein the indoor multi-view information includes target surface size and imaging size of the CCD image acquisition devices; A graphic feature extraction module, which is configured to perform image processing extraction based on the indoor multi-view information to obtain image extraction features, the image extraction features including color indicators, texture indicators, and shape indicators; A demand information acquisition module, which is configured to connect a client to acquire indoor navigation demand information, the indoor navigation demand including indoor navigation starting point position coordinates and indoor navigation ending point position coordinates; A navigation planning result acquisition module, which is configured to add the image extraction features to the indoor map information, and then perform path planning according to the indoor navigation demand information to obtain an indoor navigation planning result, the indoor navigation planning result including a dynamic update path.

Images