Garden navigation method and system based on augmented reality

Abstract

The invention relates to the technical field of garden guiding, in particular to a garden guiding method and system based on augmented reality, and the method comprises the steps: constructing an interest vector based on the attention popularity of visitors entering a garden to different guiding themes, and constructing a scenic spot attribute vector based on the weight of each scenic spot in the garden in different guiding themes; the matching degree of each scenic spot is calculated based on the interest vector and the scenic spot attribute vector, a plurality of intentional scenic spots are screened out, weighted summation is carried out based on the global matching degree and the local coherence to obtain a recommendation value of each guide route, a route recommendation list is generated for the tourist to preliminarily select, and each scenic spot arrives at the route recommendation list. The virtual image can be retrieved according to the garden image obtained in real time, and the virtual image is presented to the display interface of the guide glasses, so that the convenience and immersion of tourists in sightseeing are enhanced, and immersive guide experience is given to the tourists; meanwhile, along with the real-time position movement of the tourists in the touring process, the recommended route can be adjusted in real time, the touring requirements of the tourists are dynamically met, and personalized touring experience is provided.

Description

Technical Field

[0001] This invention relates to the field of garden tour guidance technology, specifically to a garden tour guidance method and system based on augmented reality. Background Technology

[0002] Currently, the main methods of guiding visitors through gardens include traditional guided tours, text signs set up in the scenic area, and electronic guides based on mobile apps. However, these methods all have their limitations. For example, traditional guided tours are limited by the individual quality of the guides and their time constraints, making it difficult to meet the needs of every visitor in terms of content and pace; text signs offer limited information and are not vivid or intuitive enough; while mobile app guides can provide a certain amount of content, visitors need to hold their phones, which affects the convenience and immersion of the visit.

[0003] With the development of augmented reality (AR) technology, AR navigation glasses can be applied to garden tours. By overlaying virtual information of different landscapes onto real-world scenes, tourists can experience in-depth tours and enjoy an immersive guided tour experience.

[0004] In addition to existing guided tour methods, it is also necessary to consider the changing needs of tourists and provide a guided tour method and system that can provide personalized recommendations for tour routes based on the real-time dynamics of tourists. Summary of the Invention

[0005] To address the aforementioned problems in the background art, the present invention provides a garden navigation method and system based on augmented reality.

[0006] The technical solution of this invention is as follows:

[0007] An augmented reality-based garden navigation method includes the following steps:

[0008] S1. Construct interest vectors based on the popularity of visitors to different guided tour themes, and construct attraction attribute vectors based on the weight of each attraction in the garden for different guided tour themes;

[0009] S2. Calculate the matching degree of each attraction based on the interest vector and attraction attribute vector, filter out attractions with a matching degree greater than the matching degree threshold, and obtain several intended attractions.

[0010] S3. The global matching degree is obtained based on the matching degree of all desired attractions on each tour route. The local coherence degree is obtained based on the tour theme coherence between adjacent attractions and the set dynamic weight of adjacent attraction transfer. The recommended value of each tour route is obtained by weighted summation based on the global matching degree and the local coherence. After sorting the recommended values ​​from largest to smallest, the top-ranked tour routes are generated into a route recommendation list and presented on the tour glasses display interface.

[0011] S4. Based on voice commands, select the current tour route from the route recommendation list, and provide route instructions and voice prompts for several attractions on the current tour route in sequence. When each attraction is reached, extract feature points from the acquired garden image, retrieve the virtual image based on the feature points, and present the virtual image on the tour glasses display interface, while providing voice explanations simultaneously.

[0012] Specifically, in S4, if a tourist deviates from the current guided tour route, a deviation voice prompt will be given to the tourist;

[0013] If the tourist chooses to continue deviating until reaching a deviated spot that is not on the current tour route, this deviated spot is taken as a desired spot, and several desired spots are added to the deviated spot. Step S3 is triggered for these desired spots to obtain the recommendation value of each tour route that includes this deviated spot. The tour route with the highest recommendation value is recommended.

[0014] In step S3, a recommended value for each tour route is obtained by weighted summation based on global matching degree and local coherence degree, as expressed by the following formula:

[0015] ,

[0016] in, This is the recommended value for the guided tour route. The weighting coefficients are set. For global matching degree, For local coherence; To guide all the intended attractions on the tour route Quantity, For the intended tourist attractions The degree of matching; The dynamic weighting of adjacent attractions along the tour route. For adjacent attractions along the tour route and Similarity of the guided tour themes between them , where n is the number of all attractions along the tour route.

[0017] Furthermore, the dynamic weighting of adjacent attractions The formula is expressed as follows:

[0018] ,

[0019] in, The distance weighting coefficient is set. For adjacent intended attractions arrive distance, This is the average of the sum of distances to all adjacent attractions along the tour route. The attenuation coefficient is set.

[0020] The interest vector constructed in S1 is represented as follows: ,in The interest weight for the j-th tour topic, , m represents the total number of guide topics; The value is calculated by weighting and summing the proportion of visitors selecting the j-th guided tour topic and the proportion of interaction with the j-th guided tour topic to obtain the popularity. The popularity is then normalized to obtain the interest vector, as shown in the following formula:

[0021] ,

[0022] in, In order to keep up with the trending topics, To select the weights for the proportions, The weight of the interaction ratio, To determine the number of visitors selecting the j-th guided tour topic, This refers to the total number of visitors entering the park. The number of interactions between visitors and the j-th guided tour topic. The total number of interactions between visitors and all guided tour topics;

[0023] ,

[0024] in, The interest weight for the j-th tour topic, To measure the popularity of the j-th guided tour topic, To minimize the popularity of all m tour topics, This represents the maximum popularity of all m tour topics.

[0025] The scenic spot attribute vector constructed in S1 is represented as follows: ,in Let i be the weight of attraction i in tour theme j. .

[0026] In step S2, the matching degree of each attraction is calculated based on the interest vector and the attraction attribute vector, and the formula is expressed as follows:

[0027] ,

[0028] in, The matching degree of attraction i, For interest vectors, Let i be the vector of scenic spot attributes. To calculate the interest vector Length, To calculate the length of the attraction attribute vector for attraction i.

[0029] This invention also provides an augmented reality-based garden navigation system, comprising:

[0030] Guide glasses: used to acquire different data in real time: the visitor's current location, real-world garden images, and the visitor's voice commands, and transmit these different data to the controller in real time, which then transmits them to a remote server for processing; used to provide voice prompts and explanations under the controller's instructions, and to present virtual images on the guide glasses' display interface;

[0031] Remote server: Communicates with the controller to receive data from other units acquired by the controller and processes the data, including the following processing modules:

[0032] Vector construction module: used to construct interest vectors based on the popularity of visitors to different guided tour topics, and to construct attraction attribute vectors based on the weight of each attraction in the park under different guided tour topics;

[0033] Intended attractions filtering module: This module calculates the matching degree of each attraction based on the interest vector and the attraction attribute vector, filters out attractions with a matching degree greater than the matching degree threshold, and obtains several intended attractions.

[0034] Route recommendation module: It is used to obtain the global matching degree based on the matching degree of all desired attractions on each tour route, obtain the local coherence degree based on the tour theme coherence between adjacent attractions and the set dynamic weight of adjacent attraction transfer, and obtain the recommendation value of each tour route by weighted summation based on the global matching degree and local coherence. After sorting the recommendation values ​​from largest to smallest, the top-ranked tour routes are generated into a route recommendation list and presented on the tour glasses display interface.

[0035] Augmented Reality Module: Used to select the current guided tour route from the route recommendation list based on voice commands, and to provide route instructions and voice prompts for several attractions on the current guided tour route in sequence. At each attraction location, feature points are extracted from the acquired garden image, virtual images are retrieved based on the feature points, and the virtual images are displayed on the guided tour glasses display interface, with simultaneous voice explanation.

[0036] In addition, the present invention provides an augmented reality-based garden tour guide device, including a processor and a memory, wherein the processor executes a computer program stored in the memory to implement the augmented reality-based garden tour guide method as described above.

[0037] The present invention also provides a computer-readable storage medium containing a computer program that, when executed by a processor, can implement the steps of the augmented reality-based garden tour method described above.

[0038] The beneficial effects of this invention are as follows:

[0039] By constructing interest vectors and attraction attribute vectors, several potential attractions are first selected. Then, a weighted sum is performed based on global matching degree and local coherence to obtain the recommended value for each guided tour route, generating a route recommendation list for tourists to initially select. During the tour, at each attraction, a virtual image can be retrieved based on real-time acquired garden images and displayed on the guided glasses' interface, enhancing the convenience and immersion of the tour and providing tourists with an immersive guided tour experience. At the same time, as the tourist's real-time location moves during the tour, the recommended route can be adjusted in real time to dynamically adapt to the tourist's tour needs and provide a personalized tour experience. Detailed Implementation

[0040] Exemplary embodiments of this disclosure will now be described in more detail.

[0041] Example

[0042] This embodiment provides a garden navigation method based on augmented reality, including the following steps:

[0043] S1. Construct interest vectors based on the popularity of visitors to different guided tour themes, and construct attraction attribute vectors based on the weight of each attraction in the garden for different guided tour themes.

[0044] Specifically, in step S1, an interest vector is constructed based on the level of interest visitors show to different guided tour topics. The constructed interest vector is represented as follows: ,in The interest weight for the j-th tour topic, , m represents the total number of guide topics; The value is calculated by weighting and summing the proportion of visitors selecting the j-th guided tour topic and the proportion of interaction with the j-th guided tour topic to obtain the popularity. The popularity is then normalized to obtain the interest vector, as shown in the following formula:

[0045] ,

[0046] in, In order to keep up with the trending topics, To select the weights for the proportions, The weight of the interaction ratio, To determine the number of visitors selecting the j-th guided tour topic, This refers to the total number of visitors entering the park. The number of interactions between visitors and the j-th guided tour topic. The total number of interactions between visitors and all guided tour topics;

[0047] ,

[0048] in, The interest weight for the j-th tour topic, To measure the popularity of the j-th guided tour topic, To minimize the popularity of all m tour topics, This represents the maximum popularity of all m tour topics.

[0049] In this embodiment, taking into account the focus of tourists visiting the garden landscape, different tour themes can be categorized as: plants, architecture, history, and ecology.

[0050] Step S1 also includes constructing a scenic spot attribute vector based on the weights of each scenic spot i in the garden under different tour themes. The constructed scenic spot attribute vector Represented as: ,in Let i be the weight of attraction i in tour theme j. The weighting of each attraction within different guided tour themes is fixed, primarily based on its characteristics and value in different thematic areas such as history, architectural style, plant species and ecology, culture, and art. Taking Suzhou gardens as an example, the weighting of attractions in the Humble Administrator's Garden across four different guided tour themes—plants, architecture, history, and ecology—can be determined according to its landscape features. , , , Then the scenic spot attribute vector can be represented as: .

[0051] S2. Calculate the matching degree of each attraction based on the interest vector and attraction attribute vector, filter out attractions with a matching degree greater than the matching degree threshold, and obtain several intended attractions.

[0052] In step S2, the matching degree of each attraction i is calculated based on the interest vector and the attraction attribute vector, as shown in the following formula:

[0053] ,

[0054] in, The matching degree of attraction i, For interest vectors, Let i be the vector of scenic spot attributes. To calculate the interest vector Length, To calculate the length of the attraction attribute vector for attraction i.

[0055] S3. The global matching degree is obtained based on the matching degree of all intended attractions on each tour route. The local coherence degree is obtained based on the similarity of tour themes between adjacent attractions and the set dynamic weight of adjacent attraction transfer. The recommended value of each tour route is obtained by weighted summation based on the global matching degree and the local coherence degree. After sorting the recommended values ​​from largest to smallest, the top-ranked tour routes are generated into a route recommendation list and presented on the tour glasses display interface.

[0056] The recommended value for each tour route is obtained by weighted summation based on global matching degree and local coherence degree, as shown in the formula below:

[0057] ,

[0058] in, This is the recommended value for the guided tour route. The weighting coefficients are set. For global matching degree, For local coherence; To guide all the intended attractions on the tour route Quantity, For the intended tourist attractions The degree of matching; The dynamic weighting of adjacent attractions along the tour route. For adjacent attractions along the tour route and Similarity of the guided tour themes between them , where n is the number of all attractions along the tour route.

[0059] in, , For tourist attractions The attribute vector of the scenic spot. For tourist attractions +1 is the attraction attribute vector.

[0060] Dynamic weighting of adjacent attractions The formula is expressed as follows:

[0061] ,

[0062] in, Adjacent attractions arrive distance, This is the average of the sum of distances to all adjacent attractions along the tour route. The attenuation coefficient is set.

[0063] S4. Based on voice commands, select the current tour route from the route recommendation list, and provide route instructions and voice prompts for several attractions on the current tour route in sequence. When each attraction is reached, extract feature points from the acquired garden image, retrieve the virtual image based on the feature points, and present the virtual image on the tour glasses display interface, while providing voice explanations simultaneously.

[0064] In step S4, the tourist issues a voice command to select the most interesting tour route from the recommended routes list as the current tour route and then tours along the current tour route.

[0065] The guide glasses used in this embodiment are reflective AR guide glasses. The operations in steps S1-S4 are all based on the controller built into the AR guide glasses, which transmits the acquired data from other units to a remote server for processing. The remote server is communicatively connected to the controller. The AR guide glasses also include a positioning unit, an image acquisition unit, a voice unit, and a display unit, all electrically connected to the controller. The positioning unit can be a GPS positioning chip used to obtain the visitor's current location. The image acquisition unit can be a high-resolution camera, positioned on one side of the front of the AR guide glasses frame, used to acquire real-time images of the garden. The voice unit includes an audio acquisition device and an audio output device, both electrically connected to the controller, located on the inside of the temples of the guide glasses, close to the ears. The audio acquisition device can be a microphone used to collect the visitor's voice commands and transmit them to the controller; the audio output device is a bone conduction headset used to provide voice prompts at the controller's command. The bone conduction headset transmits sound through the skull, ensuring clarity of the voice prompts without affecting the visitor's ability to hear ambient sounds.

[0066] The display unit of the AR navigation glasses can be an existing BirdBath optical display module, including a microdisplay, optical components, and optical lenses electrically connected to the controller. The controller transmits the retrieved virtual images to the microdisplay, which can be a Micro-OLED display. The optical components are reflective optical components used to receive the virtual images displayed on the microdisplay and guide the virtual images to the optical lenses. The optical lenses serve as the display interface of the navigation glasses, presenting the virtual images and garden images to the user simultaneously.

[0067] As tourists tour along the current guided route, the positioning unit acquires the tourist's current location in real time and sends the tourist's current location information to the controller. Based on the tourist's current location and the location of each attraction on the current guided route, the controller sends instructions to the voice unit to provide route guidance and voice prompts for several attractions on the current guided route in sequence. For example, the voice prompt "Current location is..., we are about to reach the next attraction, the next attraction is..." and simultaneously generates directional arrows on the display interface of the guide glasses to indicate the route.

[0068] As tourists arrive at each scenic spot, a high-resolution camera automatically captures an image of the garden. This camera, positioned on one side in front of the guide glasses' frame, can automatically focus and adjust exposure parameters even in dimly lit corners of the garden to obtain a clear image. This image is then transmitted in real-time to the controller. The controller transmits the captured image to a remote server for processing, extracting feature points, and retrieving a virtual image from the server's database based on these points. This virtual image is then displayed on the guide glasses' screen.

[0069] The specific processing steps are as follows: First, the obtained garden images are converted to grayscale to reduce the complexity of subsequent image processing. Then, scale-invariant feature transformation is applied to the grayscale images to extract feature points, including edges and corners. Simultaneously, a feature descriptor containing neighborhood information is generated for each feature point. The extracted feature points are then matched with scenic spot feature points stored in a database. The database pre-stores the scenic spot feature points and their corresponding virtual images. Virtual images with successful feature matching are obtained. Finally, the projection transformation matrix from the virtual image to the garden image is calculated.

[0070] ,

[0071] Where H is the projection transformation matrix. The x-coordinate of the feature point The ordinate of the feature point. The x-coordinate of the scenic spot's feature points. The vertical coordinates represent the feature points of the scenic spot.

[0072] Based on the projection transformation matrix H, each pixel of the virtual image is projected onto the coordinate system of the garden image, determining the position and orientation of the virtual image within the garden image. Finally, according to the position and orientation of the virtual image in the garden image, the virtual image is presented on the guide glasses display interface. By adjusting the transparency coefficient of the virtual image, both the virtual image and the real garden image can be clearly displayed simultaneously.

[0073] Step S4 also includes providing a voice prompt to tourists if they deviate from the current guided tour route. Specifically, this involves obtaining the tourist's current location P in real time. , ), retrieve the location T of the next attraction that should be reached along the current guided route. , ) and the location of the previous attraction C ( ) , If the direction vector between attraction location C and attraction location T on the current tour route is: =( - , - The direction vector between the attraction's location C and the tourist's current location P is: =( - , - );

[0074] Further obtain the direction vector In direction vector The projection length t on the surface is expressed by the following formula:

[0075] ,

[0076] in, To calculate the dot product between two vectors.

[0077] Furthermore, based on the projection length t, the projection point Q of the tourist's current position P on the line CT between scenic spot position C and scenic spot position T is obtained. , The formula is expressed as follows:

[0078] = +t( - ),

[0079] = +t( - ),

[0080] Finally, based on the distance between the tourist's current position P and the projection point position Q, the deviation distance d is obtained, expressed by the following formula:

[0081] ,

[0082] If the deviation distance d is greater than the set deviation threshold, for example, the deviation threshold can be set to 5 meters, then it is determined that the tourist has deviated from the current guided tour route, and a voice prompt will be given: "You have deviated from the current guided tour route. Do you want to continue the tour along the current guided tour route?"

[0083] In step S4, the determination of whether a tourist has deviated from the current tour route can also be made in the following way: the coverage area of ​​all attractions on the current tour route is obtained, and the new coverage area obtained by expanding each point of the coverage area outward by a preset distance is used as the coverage area boundary of the current tour route. When the tourist is visiting attractions along the current tour route, according to the tourist's position fed back by the positioning unit in real time, if the tourist's position exceeds the coverage area boundary of the current tour route, it is determined that the tourist has deviated from the current tour route, and a voice prompt is given: "You have deviated from the current tour route. Do you want to continue your tour along the current tour route?"

[0084] If the tourist chooses to continue deviating until reaching a deviated spot that is not on the current tour route, this deviated spot is taken as a desired spot, and several desired spots are added to the deviated spot. Step S3 is triggered for these desired spots to obtain the recommendation value of each tour route that includes this deviated spot. The tour route with the highest recommendation value is recommended.

[0085] This invention also provides an augmented reality-based garden navigation system, comprising:

[0086] Guide glasses: used to acquire different data in real time: the visitor's current location, real-world garden images, and the visitor's voice commands, and transmit these different data to the controller in real time, which then transmits them to a remote server for processing; used to provide voice prompts and explanations under the controller's instructions, and to present virtual images on the guide glasses' display interface;

[0087] Remote server: Communicates with the controller to receive data from other units acquired by the controller and processes the data, including the following processing modules:

[0088] Vector construction module: used to construct interest vectors based on the popularity of visitors to different guided tour topics, and to construct attraction attribute vectors based on the weight of each attraction in the park under different guided tour topics.

[0089] Intended attractions filtering module: This module calculates the matching degree of each attraction based on the interest vector and the attraction attribute vector, filters out attractions with a matching degree greater than the matching degree threshold, and obtains several intended attractions.

[0090] The route recommendation module is used to obtain the global matching degree based on the matching degree of all desired attractions on each tour route, and the local coherence degree based on the tour theme coherence between adjacent attractions and the set dynamic weight of adjacent attraction transfer. The recommendation value of each tour route is obtained by weighted summation based on the global matching degree and the local coherence degree. After sorting the recommendation values ​​from largest to smallest, the top-ranked tour routes are generated into a route recommendation list and presented on the tour glasses display interface.

[0091] Augmented Reality Module: Used to select the current guided tour route from the route recommendation list based on voice commands, and to provide route instructions and voice prompts for several attractions on the current guided tour route in sequence. At each attraction location, feature points are extracted from the acquired garden image, virtual images are retrieved based on the feature points, and the virtual images are displayed on the guided tour glasses display interface, with simultaneous voice explanation.

[0092] Furthermore, the present invention also provides an augmented reality-based garden tour guide device, including a processor and a memory, wherein the processor executes a computer program stored in the memory to implement the augmented reality-based garden tour guide method as described above.

[0093] Finally, the present invention provides a computer-readable storage medium containing a computer program that, when executed by a processor, can implement the augmented reality-based garden navigation method described above.

Claims

1. A garden navigation method based on augmented reality, characterized in that, Includes the following steps: S1. Construct interest vectors based on the popularity of visitors to different guided tour themes, and construct attraction attribute vectors based on the weight of each attraction in the garden for different guided tour themes; S2. Calculate the matching degree of each attraction based on the interest vector and attraction attribute vector, filter out attractions with a matching degree greater than the matching degree threshold, and obtain several intended attractions. S3. The global matching degree is obtained based on the matching degree of all desired attractions on each tour route. The local coherence degree is obtained based on the tour theme coherence between adjacent attractions and the set dynamic weight of adjacent attraction transfer. The recommended value of each tour route is obtained by weighted summation based on the global matching degree and the local coherence. After sorting the recommended values ​​from largest to smallest, the top-ranked tour routes are generated into a route recommendation list and presented on the tour glasses display interface. S4. Based on voice commands, select the current tour route from the route recommendation list, and provide route instructions and voice prompts for several attractions on the current tour route in sequence. When each attraction is reached, extract feature points from the acquired garden image, retrieve the virtual image based on the feature points, and present the virtual image on the tour glasses display interface, while providing voice explanations simultaneously.

2. The augmented reality-based garden navigation method according to claim 1, characterized in that, If a tourist deviates from the current guided tour route in S4, a deviation voice prompt will be given to the tourist. If the tourist chooses to continue deviating until reaching a deviated spot that is not on the current tour route, this deviated spot is taken as a desired spot, and several desired spots are added to the deviated spot. Step S3 is triggered for these desired spots to obtain the recommendation value of each tour route that includes this deviated spot. The tour route with the highest recommendation value is recommended.

3. The augmented reality-based garden navigation method according to claim 1, characterized in that, In step S3, a recommended value for each tour route is obtained by weighted summation based on global matching degree and local coherence degree, as expressed by the following formula: ， in, This is the recommended value for the guided tour route. The weighting coefficients are set. For global matching degree, For local coherence; To guide all the intended attractions on the tour route Quantity, For the intended tourist attractions The degree of matching; The dynamic weighting of adjacent attractions along the tour route. For adjacent attractions along the tour route and Similarity of the guided tour themes between them , where n is the number of all attractions along the tour route.

4. The augmented reality-based garden navigation method according to claim 3, characterized in that, The dynamic weight of adjacent attractions The formula is expressed as follows: ， in, The distance weighting coefficient is set. For adjacent intended attractions arrive distance, This is the average of the sum of distances to all adjacent attractions along the tour route. The attenuation coefficient is set.

5. The augmented reality-based garden navigation method according to claim 1, characterized in that, The interest vector constructed in S1 is represented as follows: ,in The interest weight for the j-th tour topic, , m represents the total number of guide topics; The value is calculated by weighting and summing the proportion of visitors selecting the j-th guided tour topic and the proportion of interaction with the j-th guided tour topic to obtain the popularity. The popularity is then normalized to obtain the interest vector, as shown in the following formula: ， in, In order to keep up with the trending topics, To select the weights for the proportions, The weight of the interaction ratio, To determine the number of visitors selecting the j-th guided tour topic, This refers to the total number of visitors entering the park. The number of interactions between visitors and the j-th guided tour topic. The total number of interactions between visitors and all guided tour topics; ， in, The interest weight for the j-th tour topic, To measure the popularity of the j-th guided tour topic, To minimize the popularity of all m tour topics, This represents the maximum popularity of all m tour topics.

6. The augmented reality-based garden navigation method according to claim 1, characterized in that, The scenic spot attribute vector constructed in S1 is represented as follows: ,in Let i be the weight of attraction i in tour theme j. .

7. The augmented reality-based garden navigation method according to claim 1, characterized in that, In step S2, the matching degree of each attraction is calculated based on the interest vector and the attraction attribute vector, and the formula is expressed as follows: ， in, The matching degree of attraction i, For interest vectors, Let i be the vector of scenic spot attributes. To calculate the interest vector Length, To calculate the length of the attraction attribute vector for attraction i.

8. A garden navigation system based on augmented reality, characterized in that, include: Guide glasses: used to acquire different data in real time: the visitor's current location, real-world garden images, and the visitor's voice commands, and transmit the different data to the controller in real time, and then to the remote server for processing; Used to provide voice prompts and explanations under the control of the controller, and to present virtual images on the display interface of the guide glasses; Remote server: Communicates with the controller to receive data from other units acquired by the controller and processes the data, including the following processing modules: Vector construction module: used to construct interest vectors based on the popularity of visitors to different guided tour topics, and to construct attraction attribute vectors based on the weight of each attraction in the park under different guided tour topics; Intended attractions filtering module: This module calculates the matching degree of each attraction based on the interest vector and the attraction attribute vector, filters out attractions with a matching degree greater than the matching degree threshold, and obtains several intended attractions. Route recommendation module: It is used to obtain the global matching degree based on the matching degree of all desired attractions on each tour route, obtain the local coherence degree based on the tour theme coherence between adjacent attractions and the set dynamic weight of adjacent attraction transfer, and obtain the recommendation value of each tour route by weighted summation based on the global matching degree and local coherence. After sorting the recommendation values ​​from largest to smallest, the top-ranked tour routes are generated into a route recommendation list and presented on the tour glasses display interface. Augmented Reality Module: Used to select the current guided tour route from the route recommendation list based on voice commands, and to provide route instructions and voice prompts for several attractions on the current guided tour route in sequence. At each attraction location, feature points are extracted from the acquired garden image, virtual images are retrieved based on the feature points, and the virtual images are displayed on the guided tour glasses display interface, with simultaneous voice explanation.

9. A garden navigation device based on augmented reality, characterized in that, It includes a processor and a memory, wherein the processor executes a computer program stored in the memory to implement the augmented reality-based garden navigation method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, Used to store a computer program, wherein the computer program, when executed by a processor, implements the augmented reality-based garden navigation method as described in any one of claims 1-7.

Images