Interaction method of vehicle-mounted robot, related device, electronic equipment and storage medium

Abstract

The invention discloses an interaction method of a vehicle-mounted robot, a related device, electronic equipment and a storage medium, and the interaction method of the vehicle-mounted robot comprises the steps: carrying out the detection of a shot image of a physical robot, and obtaining a detection result, about the wearing of a physical ornament by the physical robot, in the shot image; in response to the detection result, representing that the entity robot wears the entity ornament, and based on the feature information of the entity ornament, obtaining a virtual ornament matched with the entity ornament as a target ornament; and based on the current state of the virtual ornament worn by the virtual robot, whether a driver and passengers are prompted to dress up the target ornament for the virtual robot is determined. According to the scheme, on the premise that the image consistency between the entity robot and the virtual robot is ensured as much as possible, the complexity of the operation needing to be executed for maintaining the image consistency can be reduced.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to an interaction method for an in-vehicle robot and related devices, electronic equipment and storage media. Background Technology

[0002] With the increasing popularity of vehicle intelligence and the rapid improvement of vehicle hardware configuration, more and more vehicles are beginning to be equipped with physical robots to provide interactive services.

[0003] Currently, in addition to physical robots, vehicles often display virtual robots inside, which can also provide interactive services. However, in practical applications, there is often a discrepancy between the physical and virtual robots, especially after customizing their appearance. This leads to cumbersome manual adjustments required by users to maintain consistency. Therefore, reducing the complexity of maintaining consistency between physical and virtual robots while ensuring maximum visual consistency is a pressing issue. Summary of the Invention

[0004] The main technical problem addressed by this application is to provide an interaction method for vehicle-mounted robots, as well as related devices, electronic devices, and storage media, which can reduce the complexity of operations required to maintain consistency between physical and virtual robots while ensuring consistency in appearance as much as possible.

[0005] To address the aforementioned technical problems, the first aspect of this application provides an interaction method for an in-vehicle robot, comprising: detecting, based on an image captured of a physical robot, obtaining a detection result in the captured image regarding the physical robot wearing physical jewelry; wherein, the detection result includes whether the physical robot is wearing physical jewelry, and, if so, the feature information of the physical jewelry worn by the physical robot; in response to the detection result indicating that the physical robot is wearing physical jewelry, obtaining a virtual jewelry matching the physical jewelry as a target jewelry based on the feature information of the physical jewelry; and, based on the current state regarding the virtual robot wearing virtual jewelry, determining whether to prompt the driver or passenger to dress the virtual robot with the target jewelry.

[0006] To address the aforementioned technical problems, a second aspect of this application provides an interactive device for an in-vehicle robot, comprising: an image detection module, an accessory acquisition module, and a dressing prompt module. The image detection module is used to detect, based on a captured image of a physical robot, a detection result of the physical robot wearing physical accessories in the captured image; wherein, the detection result includes whether the physical robot is wearing physical accessories, and, if so, the feature information of the physical accessories worn by the physical robot; the accessory acquisition module is used, in response to the detection result indicating that the physical robot is wearing physical accessories, to obtain a virtual accessory matching the physical accessories as a target accessory based on the feature information of the physical accessories; the dressing prompt module is used, based on the current state of the virtual robot wearing virtual accessories, to determine whether to prompt the driver or passenger to dress the virtual robot with the target accessory.

[0007] To address the aforementioned technical problems, a third aspect of this application provides an electronic device comprising at least a memory and a processor coupled to each other, wherein the memory stores at least program instructions, and the processor executes the program instructions to implement the interaction method of the vehicle-mounted robot described in the first aspect.

[0008] To address the aforementioned technical problems, a fourth aspect of this application provides a computer-readable storage medium storing program instructions executable by a processor, the program instructions being used to implement the interaction method of the vehicle-mounted robot described in the first aspect.

[0009] The above scheme detects images of a physical robot to obtain detection results regarding the robot wearing physical accessories. These results include whether the robot is wearing physical accessories and, if so, the characteristic information of the accessories worn by the robot. Responding to the detection results, the scheme identifies the robot as wearing physical accessories. Based on the characteristic information of the physical accessories, a matching virtual accessory is selected as the target accessory. Furthermore, based on the current state of the virtual robot wearing the virtual accessory, the scheme determines whether to prompt the driver / passenger to dress up the virtual robot with the target accessory. Therefore, by detecting accessories on the physical robot and combining this with the current state of the virtual robot wearing accessories, the scheme determines whether to prompt the driver / passenger to dress up the virtual robot with virtual accessories. This eliminates the need for the driver / passenger to manually search for matching virtual accessories to ensure consistency, reducing the complexity of operations required to maintain consistency. On the other hand, by dressing up the virtual robot with the target accessory, the scheme ensures consistency between the physical robot and the virtual robot's appearance as much as possible, even when the physical robot is wearing physical accessories. Therefore, while ensuring the consistency of appearance between physical and virtual robots as much as possible, the complexity of the operations required to maintain this consistency can be reduced. Attached Figure Description

[0010] Figure 1 This is a flowchart illustrating an embodiment of the interaction method for an in-vehicle robot according to this application; Figure 2 This is a schematic diagram of the framework of an embodiment of the interaction system of the vehicle-mounted robot of this application; Figure 3 This is a schematic diagram of the framework of an embodiment of the interactive device for the vehicle-mounted robot of this application; Figure 4 This is a schematic diagram of the framework of an embodiment of the electronic device of this application; Figure 5 This is a schematic diagram of a framework of an embodiment of the computer-readable storage medium of this application. Detailed Implementation

[0011] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0012] In the following description, specific details such as particular system architectures, interfaces, and technologies are presented for illustrative purposes rather than for limiting purposes, in order to provide a thorough understanding of this application.

[0013] In this paper, the terms "system" and "network" are often used interchangeably. The term "and / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the slash " / " generally indicates that the preceding and following related objects have an "or" relationship. Furthermore, "many" in this paper indicates two or more objects.

[0014] Please see Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the interaction method for an in-vehicle robot according to this application. It should be noted that the process operations in this embodiment can be executed by an electronic device with computing capabilities or by related equipment including such an electronic device. The specific structure of the electronic device or related equipment including the electronic device is not limited here. Specifically, this embodiment may include the following steps: Step S11: Based on the captured images of the physical robot, the detection results of the physical robot wearing physical accessories in the captured images are obtained.

[0015] In this embodiment of the disclosure, the detection result may include whether the physical robot is wearing physical jewelry, and if so, the feature information of the physical jewelry worn by the physical robot. As a possible example, the feature information may include the field value of at least one preset field. It should be noted that several preset fields can be preset. For example, the several preset fields may include, but are not limited to: jewelry category (e.g., hat, etc.), jewelry subcategory (e.g., pirate hat, Santa hat, birthday hat, etc.), jewelry color (e.g., black, red, white, etc.), etc. Based on this, when detecting physical jewelry in the captured image, each preset field can be filled to obtain the field value of at least one preset field as feature information. In practical applications, it is possible that some preset fields do not have their field values ​​detected, in which case the preset fields may not be filled (i.e., the field value can be empty). Of course, the above example is only one possible example of feature information in practical applications, and other possible configurations of feature information are not limited here, nor will they be listed one by one.

[0016] In one implementation scenario, an in-vehicle camera can capture images of a physical robot inside the vehicle. For example, the in-vehicle camera can capture images of the robot in real time, obtaining images at various moments. These images can then be continuously analyzed to detect whether the robot is wearing physical jewelry. Alternatively, the in-vehicle camera can capture images of the robot when the vehicle starts (e.g., the instant the vehicle starts, or during a period of time after starting), obtaining images of the robot during vehicle startup. These images can then be analyzed during startup to detect whether the robot is wearing physical jewelry, reducing the computational load of image detection and avoiding frequent alerts related to jewelry that could inconvenience passengers. It should be noted that the above examples are merely a few possible implementations of image capture and jewelry detection for physical robots in practical applications. Other possible methods are not limited here, nor will they be listed in detail.

[0017] In one implementation scenario, an object detection model such as Faster R-CNN can be used to detect objects in the captured image, obtaining a first region of the robot in the captured image. Based on this, a preset image containing only the robot (e.g., a rendered image of the robot without a background or wearing any accessories) can be compared with the first region of the robot in the captured image. This allows for the determination that if a second region of the accessory is identified in the captured image through difference comparison, the detection result includes the robot wearing the accessory; otherwise, it can be determined that the robot is not wearing the accessory. Alternatively, image segmentation (e.g., instance segmentation, semantic segmentation) can be performed on the first region of the robot in the captured image to obtain second regions of each instance within the first region (e.g., the robot itself, the accessory, the background, etc.). This allows for the determination that if a second region of the accessory is identified, the detection result includes the robot wearing the accessory; otherwise, it can be determined that the robot is not wearing the accessory. Of course, the above examples are merely possible illustrations of determining whether a robot is wearing accessories using two different methods: difference comparison and image segmentation. Other possible implementation methods are not limited here, nor will they be listed in detail.

[0018] In one implementation scenario, given that a physical robot is wearing physical jewelry, feature extraction can be performed based on the second region of the physical jewelry (the method of obtaining this region can be found in the aforementioned description) to obtain the field values ​​of various preset fields. As a possible example, a large model instruction can be constructed based on the preset fields and the second region of the physical jewelry. This large model instruction instructs a large language model to analyze the second region of the physical jewelry and the field values ​​of the preset fields related to the jewelry. The output information of the large language model in response to the large model instruction is then obtained to acquire the feature information of the physical jewelry worn by the robot. For example, taking the preset fields including jewelry category, jewelry subcategory, and jewelry color as an example, the following large model instructions can be constructed, including but not limited to: Please input the physical accessories worn by the robot in the image [you can fill in the physical accessories in the image here]. The second area for accessories is filled with the following field values ​​related to accessories (accessory category, accessory subcategory, accessory color).

[0019] It should be noted that the above examples are merely one possible illustration of large model instructions. The specific content of these instructions is not limited, nor will they be listed in detail. Alternatively, detection models corresponding to each preset field can be pre-trained. It should be noted that the detection model corresponding to any preset field is a neural network model specifically designed to detect the field value of the preset field in image data. For example, the detection model corresponding to the preset field "jewelry category" is a neural network model specifically designed to detect the field value of the preset field "jewelry category" in image data. Detection models corresponding to other preset fields can be deduced similarly, and will not be elaborated further here. Based on this, the second region of the physical jewelry can be detected using the detection models corresponding to each preset field, yielding the detection results for each preset field. It should be noted that when the detection results include valid content (e.g., for the preset field "Accessories Category", the detection results include "Hat -90%", "Scarf -5%", "Glasses -5%", then since the accessory category can be distinguished according to the predicted probability, the detection results include the valid content "Hat"), this valid content can be used as the field value of the preset field (e.g., using the aforementioned valid content "Hat" as the field value of the preset field "Accessories Category"). When the detection results do not contain valid content (e.g., for the preset field "Accessories Subcategory", the detection results include "Pirate Hat -33.33%", "Santa Hat -33.33%", "Birthday Hat -33.33%", then since the accessory subcategory cannot be distinguished according to the predicted probability, the detection results do not contain valid content), the field value of the preset field can be set to empty. Of course, the above examples are only a few possible examples of obtaining feature information when the feature information contains the field value of at least one preset field. Other possible methods of obtaining feature information are not limited here, nor will they be listed one by one.

[0020] In another implementation scenario, in practical applications, besides the aforementioned method of first detecting the first region of the physical robot in the captured image, and then detecting the second region of the physical ornament within the first region to detect feature information, the captured image can also be directly detected based on a sample library to obtain the detection result. It should be noted that the sample library can contain several image data of a physical robot wearing physical ornaments, and the image data can be labeled with the field values ​​of various preset fields for the physical ornaments worn by the physical robot. For example, in image data A in the sample library, the physical robot is wearing physical ornament A, and image data A can be labeled with the field values ​​of various preset fields such as "ornament category," "ornament subcategory," and "ornament color" for physical ornament A, and so on. Further details about the various image data in the sample library will not be elaborated here. In this way, the captured image can be directly compared with each image data in the sample library to determine the image data most similar to the captured image. If the similarity is higher than the set threshold, the detection result can be directly determined to include physical robots wearing physical ornaments, and the feature information of the physical ornaments worn by the physical robots is the information labeled by the image data most similar to the captured image.

[0021] It should be noted that the above examples are only a few possible implementation methods for detecting captured images to obtain detection results in practical applications. Other possible implementation methods are not limited here, nor will they be listed one by one.

[0022] Step S12: In response to the detection result being characterized as a physical robot wearing physical jewelry, a virtual jewelry matching the physical jewelry is obtained as the target jewelry based on the feature information of the physical jewelry.

[0023] In one implementation scenario, as a possible example, the target accessory can be retrieved from a decoration material library using feature information. The virtual accessories in the decoration material library can have several preset field values ​​defined. The feature information can include at least one preset field value. If the feature information does not lack any preset field values, virtual accessories that match the values ​​of all preset fields in the feature information can be directly selected as the target accessory. For example, assuming the preset fields include accessory category, accessory subcategory, and accessory color, if the preset field "accessory category" has the value "hat," the preset field "accessory subcategory" has the value "pirate hat," and the preset field "accessory color" has the value "black," then field matching can be performed on each virtual accessory based on the aforementioned feature information. If a virtual accessory also has the preset field "hat" for "accessory category," the preset field "pirate hat" for "accessory subcategory," and the preset field "accessory color" for "black," then that virtual accessory can be directly selected as the target accessory matching the physical accessory. Of course, the above example is only one possible case when the feature information does not lack any preset field values ​​in actual application. Other possible cases are not limited here, nor will they be listed one by one.

[0024] In another implementation scenario, differing from the aforementioned implementation, as another possible example, as mentioned earlier, the virtual ornaments in the decorative material library can be defined with several preset field values. The feature information can include the value of at least one preset field. If the feature information is missing some preset field values, virtual ornaments whose field values ​​match the feature information can be selected as candidate ornaments. Then, based on the preset fields in the feature information that are missing values ​​compared to the candidate ornaments, a query statement is constructed. This query statement can be used to prompt the driver / passenger to provide feedback on the preset fields containing the missing values. Based on this, the feature information can be supplemented based on the driver / passenger's feedback on the query statement, and the process of selecting virtual ornaments whose field values ​​match the feature information as candidate ornaments can be repeated until no feature information is missing. Finally, the virtual ornament whose field values ​​in all preset fields match the feature information is selected as the target ornament. The above method, when the field values ​​of the preset fields are missing in the feature information, first filters out matching candidate ornaments from the decoration material library based on the existing field values. Then, it constructs a query statement based on the preset fields where the feature information is missing field values ​​compared to the candidate ornaments, and interacts with the driver and passengers based on the driver and passengers' feedback on the query statement. This process is repeated, which can supplement and improve the feature information through human-computer interaction, thereby improving the completeness and accuracy of the feature information as much as possible, and also enhancing the fun of human-computer interaction.

[0025] In a specific implementation scenario, as long as a preset field with a value in the feature information can match any virtual accessory in the decoration material library, that virtual accessory can be selected as a candidate accessory. For ease of understanding, let's take the example where each preset field includes accessory category, accessory subcategory, and accessory color. If the preset field "Accessory Category" has the value "Hat," the "Accessory Subcategory" has the value "None," and the "Accessory Color" has the value "Black," then for any virtual accessory in the decoration material library, as long as the virtual accessory's preset field "Accessory Category" value is also "Hat" and its preset field "Accessory Color" value is also "Black," this virtual accessory can be selected as a candidate accessory. Of course, the above example is only one possible example of selecting candidate accessories in practical applications; other possible scenarios are not limited here, nor will they be listed in detail.

[0026] In a specific implementation scenario, after screening out candidate ornaments, inquiry statements can be constructed based on this. Specifically, the missing frequencies of each preset field can be obtained by comparing the characteristic information with the preset fields of the missing field values of each candidate ornament respectively. It should be noted that although the field values of each preset field can be predefined for each virtual ornament in the decoration material library, it is not excluded that in actual applications, the field values of individual virtual ornaments in individual preset fields are null values (that is, there may also be cases where individual preset fields of individual virtual ornaments in the decoration material library lack information). Exemplarily, taking the four fields A, B, C, and D as the preset fields, the virtual ornament "Jia" may only have valid field values in the three preset fields A, B, and C, and the virtual ornaments "Yi" and "Bing" may only have valid field values in the three preset fields A, B, and D. Of course, the virtual ornament "Ding" may have valid field values in all four preset fields A, B, C, and D. Based on this, if the characteristic information only has valid field values in the two preset fields A and B, then compared with the virtual ornament "Jia", the preset field C is missing, compared with the virtual ornaments "Yi" and "Bing", the preset field D is missing, and compared with the virtual ornament "Ding", the preset fields C and D are missing. Therefore, the missing frequency of the preset field C is 2, and the missing frequency of the preset field D is 3. Of course, the above example is only a possible example in actual applications, and other possible situations are not limited here. After obtaining the missing frequencies of each preset field, the preset fields can be selected as target fields in the order of the missing frequencies from high to low, and inquiry statements can be constructed based on the target fields, and the inquiry statements can be used to ask the driver and passengers about the field values of the target fields. Still taking the above example, since the missing frequency of the preset field C is 2 and the missing frequency of the preset field D is 3, the preset field D can be selected as the target field, and an inquiry statement can be constructed based on the target field D, such as "Wow, let me guess what [fill in the target field D here] is. Is it [fill in a random value of the target field D here]" and so on. The specific expression of the inquiry statement is not limited here, and no more examples will be given. The above method, by comparing the characteristic information with the preset fields of the missing field values of each candidate ornament respectively to obtain the missing frequencies of each preset field, selecting the preset fields as target fields in the order of the missing frequencies from high to low, and then constructing inquiry statements based on the target fields, and the inquiry statements are used to ask the driver and passengers about the field values of the target fields, can interact with the driver and passengers preferentially on the preset fields with higher missing frequencies, which helps to complete the information supplement of the preset fields with higher missing frequencies first, and then can further screen the pending candidate ornaments as quickly and efficiently as possible in the subsequent process of returning to the previous screening of candidate ornaments, which helps to improve the accuracy and efficiency of determining the target ornament.

[0027] In a specific implementation scenario, as mentioned earlier, after constructing the query statement, the query statement can be output, awaiting feedback from the driver or passenger. Taking the aforementioned query statement "Wow, let me guess what [the target field D is here] is, is it [the random value of target field D]?" as an example, in practical applications, as a possible example, the driver or passenger's feedback statement could be "It's XXX, do you like it?" Of course, the above example is merely one possible example in practical applications; the specific expression of the feedback statement is not limited here, nor will it be listed in detail. Based on this, feature information can be supplemented based on the feedback statement. Again using the above example, the preset field D value in the feature information can be supplemented to "XXX" in the feedback statement. After this, the virtual accessories whose filter field values ​​match the feature information can be returned as candidate accessories. Taking the previous example, since the field value of the preset field D has been supplemented, a new batch of candidate ornaments can be selected from the decoration material library based on the new feature information (i.e., the field value of the preset field A, the field value of the preset field B, and the field value of the preset field C). This process is repeated until the target ornament that perfectly matches the latest feature information can be selected from the decoration material library.

[0028] In a specific implementation scenario, as mentioned earlier, to detect wearable accessories on a robot within a captured image, detection can be performed on the captured image based on a sample library. If, during the iterative process of the aforementioned steps, feature information is found to be no longer missing, the sample library can be expanded based on captured images with labeled feature information (i.e., those with complete feature information). This increases the likelihood of detecting complete feature information when detecting new captured images based on the sample library in the future.

[0029] In one implementation scenario, regardless of the aforementioned situation (i.e., regardless of whether the feature information lacks the values ​​of preset fields), if no virtual accessory matching the physical accessory is found in the decoration material library, a virtual accessory matching the physical accessory can be directly generated as the target accessory based on the feature information. For example, a large model instruction can be constructed based on the feature information, and this instruction can be used to instruct the multimodal large model to generate a virtual accessory matching the feature information. The output of the multimodal response to the large model instruction can then be obtained as the target accessory matching the physical accessory. Of course, in practical applications, other methods can also be used to obtain a virtual accessory matching the physical accessory as the target accessory. For example, instead of searching the decoration material library using feature information, a virtual accessory matching the physical accessory can be directly generated as the target accessory based on the feature information. It should be noted that the above examples are merely a few possible examples of obtaining target accessories in practical applications. The specific methods for obtaining target accessories are not limited here (for example, the two methods mentioned above, namely, searching for virtual accessories in the decoration material library and directly generating virtual accessories, can be executed simultaneously, and then the virtual accessory that best matches the physical accessory can be selected as the target accessory from the virtual accessories obtained by the two methods respectively, or all the virtual accessories obtained by the two methods can be used as target accessories, so that drivers and passengers can choose from them). No further examples will be given.

[0030] In one implementation scenario, if the detection result indicates that the robot is not wearing any physical accessories, as one possible example, the process can be terminated directly, i.e., the search in the decoration material library and the following prompts for dressing up are no longer performed. Alternatively, as another possible example, the process can return to the steps described above, which involve detecting the robot wearing physical accessories based on the captured image and obtaining the detection result in the image. This process can then be repeated. Of course, the above examples are merely a few possible implementation methods when the detection result indicates that the robot is not wearing any physical accessories in a practical application. Other possible implementation methods are not limited here, nor will they be listed one by one.

[0031] Step S13: Based on the current state of the virtual robot wearing virtual accessories, determine whether to prompt the driver or passenger to dress up the virtual robot with the target accessories.

[0032] In one implementation scenario, virtual robots can be displayed through vehicle-mounted displays, naked-eye 3D imaging, etc. The display methods for virtual robots are not limited here, nor will they be listed one by one.

[0033] In one implementation scenario, the current state of a virtual robot wearing virtual accessories can include two possible scenarios: the virtual robot is not wearing virtual accessories, or the virtual robot is wearing virtual accessories. Furthermore, if the current state indicates the virtual robot is wearing virtual accessories, it can also include the characteristic information of the virtual accessories worn by the virtual robot (such as field values ​​in various preset fields). The above examples are merely a few possible instances of the current state of a virtual robot wearing virtual accessories in practical applications; other possible scenarios are not limited here, nor will they be listed individually.

[0034] In one implementation scenario, as a possible approach, after retrieving the target accessory and determining the current state of the virtual robot wearing the accessory, it can directly detect whether the current state of the virtual robot wearing the accessory matches the target accessory. If the detection indicates a match, no prompt is needed; otherwise, if the detection indicates a mismatch, the driver / passenger can be prompted to dress the virtual robot in the target accessory. It should be noted that if the current state of the virtual robot wearing the accessory is that the virtual robot is wearing the target accessory, it can be determined that the current state of the virtual robot wearing the accessory matches the target accessory. Conversely, if the current state of the virtual robot wearing the accessory is that the virtual robot is not wearing the accessory or is wearing the accessory but not the target accessory, it can be determined that the current state of the virtual robot wearing the accessory does not match the target accessory.

[0035] In another implementation scenario, as a possible alternative, distinct from the aforementioned implementation, before determining whether to prompt the driver / passenger to dress up the virtual robot with the target accessory based on the current state of the virtual robot wearing virtual accessories, it is possible to obtain the consistency result between the current detection result of the physical robot wearing physical accessories and the previous detection result of the physical robot wearing physical accessories. Based on this, it is possible to determine whether to prompt the driver / passenger to dress up the virtual robot with virtual accessories based on the current state and the consistency result. This method, by combining the current state with the consistency result to determine whether to prompt the driver / passenger to dress up the virtual robot with the target accessory, can reduce frequent prompts due to inconsistencies between the virtual robot and the physical robot in terms of accessories.

[0036] In a specific implementation scenario, as a possible example, as mentioned earlier, jewelry detection can be performed in real time. Therefore, the detection result for the robot wearing jewelry could be the result obtained by detecting the image captured at the current moment, while the previous detection result could be the result obtained by detecting the image captured at the moment before the current moment. Alternatively, as another possible example, as mentioned earlier, jewelry detection can be performed only when the vehicle starts. Therefore, the detection result for the robot wearing jewelry could be the result obtained by detecting the image captured when the vehicle started this time, while the previous detection result could be the result obtained by detecting the image captured when the vehicle started last time. Of course, the above examples only illustrate the possible meanings of the current and previous detection results under different detection methods. Other possible scenarios are not limited here, nor will they be listed in detail.

[0037] In a specific implementation scenario, after obtaining a consistent result, in response to the current state being characterized as the virtual robot not wearing the target accessory and the consistency result being characterized as inconsistent, the driver or passenger can be prompted to dress the virtual robot in the target accessory. It should be noted that the current state being characterized as the virtual robot not wearing the target accessory can specifically include any of the following: the virtual robot is not wearing the accessory, or the accessory worn by the virtual robot is not the target accessory.

[0038] In a specific implementation scenario, after obtaining a consistency result, no prompt is needed in response to whether the current state indicates that the virtual robot has worn the target accessory or whether the consistency result indicates consistency. In other words, as long as the current state indicates that the robot has worn the target accessory, no prompt is needed regardless of whether the consistency result indicates consistency; similarly, as long as the consistency result indicates consistency, no prompt is needed regardless of whether the current state indicates that the virtual robot has worn the target accessory.

[0039] In one implementation scenario, if the detection result indicates that the robot is wearing a physical accessory, it can further detect whether the accessory currently worn by the robot exists in any mapping relationship within the scene accessory library based on the feature information of the accessory. It should be noted that the mapping relationship can include driving scenes and physical accessories associated with those scenes. For example, the scene accessories could include a mapping relationship between the physical accessory "pirate hat" and the driving scene "self-driving tour." Based on this, in response to the existence of a mapping relationship for the accessory currently worn by the robot, the mapping relationship can be selected as the target relationship. Furthermore, based on the driving scene and physical accessories in the target relationship, a slogan can be constructed, allowing the robot to output the slogan. It should be noted that the slogan can at least be used to enhance the driving atmosphere of the target scene for the driver and passengers by emphasizing the physical accessories in the target relationship. The above method detects whether the physical ornament exists in any mapping relationship in the scene ornament library based on the feature information of the physical ornament. If a mapping relationship exists, it constructs and outputs slogans to enhance the driving atmosphere by combining the physical ornaments in the existing mapping relationship and the driving scene. This can enhance the driver's and passengers' perception of the driving scene and physical ornaments.

[0040] In a specific implementation scenario, as one possible example, the mapping relationships in the scene decoration library can be set by the driver / passenger or other personnel (such as factory settings). Alternatively, as another possible example, the mapping relationships in the scene decoration library can also be automatically updated and generated during the interaction of the onboard robot when certain conditions are detected. Or, as yet another possible example, the mapping relationships in the scene decoration library can be configured and generated using both of the above methods. For example, they can be configured by relevant personnel initially, and then automatically updated and generated later.

[0041] In a specific implementation scenario, as one possible example, if the physical accessory currently worn by the robot exists in any mapping relationship in the scene accessory library, the existing mapping relationship can be selected as the target relationship. Alternatively, as another possible example, in addition to detecting whether the physical accessory currently worn by the robot exists in any mapping relationship in the scene accessory library, it is also possible to further detect whether the current driving scene exists in the same mapping relationship as the currently worn physical accessory. If so, the existing mapping relationship is selected as the target relationship. It should be noted that the current driving scene can be determined by detecting the origin and destination of this drive. For example, if the origin is "workplace" and the destination is "home," the driving scene could be "getting off work"; conversely, if the origin is "home" and the destination is "workplace," the driving scene could be "going to work"; or if the destination is a "tourist attraction," the driving scene could be "playing," etc. Of course, the above examples are only a few possible scenarios for driving scenes in practical applications; other possible scenarios are not limited here, nor will they be listed one by one.

[0042] In a specific implementation scenario, after selecting the target relation, a slogan can be constructed based on the driving scenario and physical objects within the target relation. As a possible example, a large model instruction can be constructed based on the driving scenario and physical objects in the target relation. This large model instruction instructs the large language model to construct a slogan based on the driving scenario and physical objects, requiring the slogan to emphasize how the physical objects in the target relation enhance the driving atmosphere of the driving scenario for the driver and passengers. The output statement of the large language model in response to the large model instruction can then be obtained as the slogan. It should be noted that the above example is merely one possible way to construct slogans in practical applications; other possible construction methods are not limited here, nor will they be listed in detail. For ease of understanding, taking the aforementioned target accessory "black pirate hat" as an example, when it is detected that the target accessory "black pirate hat" exists in the scene accessory library in the mapping relationship "pirate hat - play", this mapping relationship can be selected as the target relationship. Based on the physical accessory "pirate hat" and the driving scene "play" in this target relationship, a slogan statement can be constructed to enhance the driving atmosphere of the driving scene in the target relationship by emphasizing the physical accessory in the target relationship. Such a slogan statement may include, but is not limited to: "Let's put on the pirate hat and set off to explore the world!" etc. The specific content of the slogan statement is not limited here, nor will it be listed one by one.

[0043] In a specific implementation scenario, as mentioned earlier, the mapping relationships in the scene accessory library can be automatically updated and generated during the interaction of the in-vehicle robot when certain conditions are met. Specifically, in response to the fact that the physical accessory currently worn by the robot does not exist in any mapping relationship, it can be determined whether the physical accessory currently worn by the robot and the current driving scene have appeared consecutively multiple times (i.e., co-occurred). Based on the determination result, it can be determined whether to bind the physical accessory currently worn by the robot and the current driving scene as a new mapping relationship, so as to update the scene accessory library based on the new mapping relationship. For example, if the physical accessory currently worn by the robot and the current driving scene have appeared consecutively multiple times (i.e., co-occurred), the physical accessory currently worn by the robot and the current driving scene can be bound as a new mapping relationship. For example, if the driving scene "Playing" and the physical accessory "Pirate Hat" have appeared consecutively multiple times (i.e., co-occurred), then the driving scene "Playing" and the physical accessory "Pirate Hat" can be bound as a new mapping relationship. It should be noted that the aforementioned "multiple occurrences" standard can be set according to actual application needs. For example, when the binding accuracy requirements between the driving scene and the physical accessory are high, the number of occurrences standard can be set higher; conversely, when the binding accuracy requirements between the driving scene and the physical accessory are relatively relaxed, the number of occurrences standard can be set appropriately lower. The specific value of the number of occurrences standard is not limited here. This method determines whether to bind the physical accessory currently worn by the robot and the current driving scene as a new mapping relationship by judging whether they have appeared consecutively multiple times. This enables automatic updating and generation of the mapping relationship during the interaction process of the in-vehicle robot.

[0044] In one implementation scenario, please refer to the following: Figure 2 , Figure 2 This is a schematic diagram of the framework of an embodiment of the interaction system for the vehicle-mounted robot of this application. Figure 2 As shown, a physical robot is the physical device of a vehicle-mounted robot, while a virtual robot is a virtual device of a vehicle-mounted robot (i.e., composed of...). Figure 2The display screen carries the display (or it can be displayed using other methods such as naked-eye 3D imaging). The physical robot can demonstrate its appearance and movements, and passengers can dress it up. The physical robot can communicate directly with the domain controller in the cockpit via wiring harnesses, or wirelessly via Bluetooth, Wi-Fi, or other methods. Furthermore, the display screen can communicate with the physical robot, drive the virtual robot's display, and download resources from a cloud-based decoration resource library for application on the virtual robot. Cameras in the interactive system can capture image data from the physical robot and identify decorations to provide abstracted feature information. The cloud-based decoration resource library provides unified management and continuous updates to the virtual robot's decoration resources, allowing passengers to constantly access new experiences. It should be noted that the decoration resource library can contain not only virtual materials but also model resources corresponding to the virtual robot's applications. The following is a brief description of the interaction process between the various terminals in the interactive system (refer to the foregoing description in the embodiments of this disclosure): First, an image of the physical robot can be captured by a camera; then, a detection operation can be performed on the captured image regarding the wearing of physical accessories to obtain detection results (e.g., the physical robot is wearing physical accessories and the feature information of the worn physical accessories, or the physical robot is not wearing physical accessories); then, based on the aforementioned detection results, whether the virtual robot is wearing virtual accessories is simultaneously detected and compared. If the accessories of the two are consistent, no prompt is made; otherwise, if the accessories of the two are inconsistent, a prompt can be made (e.g., prompting the driver / passenger to replace the accessories on the virtual robot). Furthermore, to reduce frequent prompts due to inconsistencies between the virtual robot and the physical robot's accessories, a prompt can be made only when the physical robot's current accessories have changed from the previous ones. For example, if the physical robot has always worn a hat, but the driver / passenger changes the virtual accessories on the virtual robot, no prompt can be made even if the accessories of the two are inconsistent. Of course, Figure 2 The above brief explanation is merely one possible example in practical application. Other possible situations are not limited here, nor will they be listed one by one.

[0045] The above scheme detects images of a physical robot to obtain detection results regarding the robot wearing physical accessories. These results include whether the robot is wearing physical accessories and, if so, the characteristic information of the accessories worn by the robot. Responding to the detection results, the scheme identifies the robot as wearing physical accessories. Based on the characteristic information of the physical accessories, a matching virtual accessory is selected as the target accessory. Furthermore, based on the current state of the virtual robot wearing the virtual accessory, the scheme determines whether to prompt the driver / passenger to dress up the virtual robot with the target accessory. Therefore, by detecting accessories on the physical robot and combining this with the current state of the virtual robot wearing accessories, the scheme determines whether to prompt the driver / passenger to dress up the virtual robot with virtual accessories. This eliminates the need for the driver / passenger to manually search for matching virtual accessories to ensure consistency, reducing the complexity of operations required to maintain consistency. On the other hand, by dressing up the virtual robot with the target accessory, the scheme ensures consistency between the physical robot and the virtual robot's appearance as much as possible, even when the physical robot is wearing physical accessories. Therefore, while ensuring the consistency of appearance between physical and virtual robots as much as possible, the complexity of the operations required to maintain this consistency can be reduced.

[0046] Please see Figure 3 , Figure 3 This is a schematic diagram of the framework of an embodiment of the interactive device for an in-vehicle robot according to this application. The interactive device 30 for the in-vehicle robot includes: an image detection module 31, an accessory acquisition module 32, and a dressing prompt module 33. The image detection module 31 is used to detect based on a captured image of the physical robot to obtain a detection result of the physical robot wearing physical accessories in the captured image; wherein, the detection result includes whether the physical robot is wearing physical accessories, and if so, the feature information of the physical accessories worn by the physical robot; the accessory acquisition module 32 is used to, in response to the detection result indicating that the physical robot is wearing physical accessories, obtain a virtual accessory that matches the physical accessories as the target accessory based on the feature information of the physical accessories; the dressing prompt module 33 is used to determine whether to prompt the driver or passenger to dress the virtual robot with the target accessory based on the current state of the virtual robot wearing virtual accessories.

[0047] In the above scheme, the interaction device 30 of the vehicle-mounted robot detects images of the physical robot to obtain detection results regarding the physical robot wearing physical accessories. The detection results include whether the physical robot is wearing physical accessories and, if so, the feature information of the physical accessories worn by the physical robot. In response to the detection results indicating that the physical robot is wearing physical accessories, a virtual accessory matching the physical accessory is obtained as the target accessory based on the feature information of the physical accessory. Based on the current state of the virtual robot wearing the virtual accessory, it is determined whether to prompt the driver or passenger to dress up the virtual robot with the target accessory. Therefore, on the one hand, by detecting accessories on the physical robot and combining it with the current state of the virtual robot wearing accessories, it is determined whether to prompt the driver or passenger to dress up the virtual robot with virtual accessories. This eliminates the need for the driver or passenger to manually search for virtual accessories that match the physical accessories to ensure consistency in appearance, thus reducing the complexity of the operations required to maintain consistency in appearance. On the other hand, by dressing up the virtual robot with the target accessory, it is possible to ensure consistency in appearance between the physical robot and the virtual robot as much as possible when the physical robot is wearing physical accessories. Therefore, while ensuring the consistency of appearance between physical and virtual robots as much as possible, the complexity of the operations required to maintain this consistency can be reduced.

[0048] In some disclosed embodiments, the interaction device 30 of the vehicle-mounted robot includes a consistency detection module for obtaining the consistency result between the current detection result of the physical robot wearing physical accessories and the previous detection result of the physical robot wearing accessories; the dressing prompt module 33 is specifically used to determine whether to prompt the driver or passenger to dress up the virtual robot with the target accessories based on the current state and the consistency result.

[0049] In some disclosed embodiments, the dressing-up prompt module 33 includes a first response submodule, used to prompt the driver / passenger to dress the virtual robot with the target accessory in response to the current state indicating that the virtual robot is not wearing the target accessory and the consistency result indicating that it is inconsistent; the dressing-up prompt module 33 includes a second response submodule, used to not prompt in response to the current state indicating that the virtual robot is wearing the target accessory or the consistency result indicating that it is consistent; wherein, the current state indicating that the virtual robot is not wearing the target accessory includes any of the following: the virtual robot is not wearing the accessory, or the accessory worn by the virtual robot is not the target accessory.

[0050] In some disclosed embodiments, the target ornament is retrieved from a decorative material library using feature information. The decorative material library defines virtual ornaments with several preset field values. The feature information includes at least one preset field value. When the feature information is missing some preset field values, the ornament acquisition module 32 includes an ornament filtering submodule for filtering virtual ornaments whose field values ​​match the feature information as candidate ornaments. The ornament acquisition module 32 also includes an inquiry construction submodule for constructing an inquiry statement based on the preset fields in the feature information that are missing field values ​​compared to the candidate ornaments. The inquiry statement prompts the driver / passenger to provide feedback on the preset fields that provide supplementary field values. The ornament acquisition module 32 further includes a supplementation loop submodule for supplementing the feature information based on the driver / passenger's feedback on the inquiry statement and returning the steps of filtering virtual ornaments whose field values ​​match the feature information as candidate ornaments. This process continues until no more feature information is missing, and then the virtual ornament whose field values ​​match all preset fields are selected as the target ornament.

[0051] In some disclosed embodiments, the query construction submodule includes a frequency statistics unit, used to obtain the missing frequency of each preset field by comparing the feature information with the preset fields of each candidate accessory missing field value; the query construction submodule includes a field selection unit, used to select preset fields as target fields according to the order of missing frequency from high to low; the query construction submodule includes a statement construction unit, used to construct a query statement based on the target fields; wherein, the query statement is used to ask the driver and passengers for the field value of the target field.

[0052] In some disclosed embodiments, the image detection module 31 is specifically used to detect captured images based on a sample library to obtain detection results; wherein, the sample library contains several image data of physical robots wearing physical ornaments, and the image data is labeled with the field values ​​of the physical ornaments worn by the physical robots regarding various preset fields; the interactive device 30 of the vehicle robot also includes a sample expansion module, which is used to expand the sample library based on captured images labeled with feature information after the feature information is no longer missing.

[0053] In some disclosed embodiments, when the detection result indicates that the physical robot is wearing physical accessories, the interaction device 30 of the vehicle-mounted robot further includes a mapping detection module, used to detect whether the physical accessories currently worn by the physical robot exist in any mapping relationship in the scene accessory library based on the feature information of the physical accessories currently worn by the physical robot; wherein, the mapping relationship includes driving scenes and physical accessories bound to driving scenes; the interaction device 30 of the vehicle-mounted robot further includes a slogan construction module, used to select the mapping relationship of the physical accessories currently worn by the physical robot as the target relationship in response to the existence of the physical accessories currently worn by the physical robot in the mapping relationship, and construct a slogan statement based on the driving scene and physical accessories in the target relationship; the interaction device 30 of the vehicle-mounted robot further includes a statement output module, used to control the physical robot to output the slogan statement; wherein, the slogan statement is at least used to enhance the driving atmosphere of the driving scene in the target relationship for the driver and passengers by emphasizing the physical accessories in the target relationship.

[0054] In some disclosed embodiments, the interaction device 30 of the vehicle-mounted robot further includes a co-occurrence determination module, which is used to determine whether the physical ornament currently worn by the physical robot and the current driving scene have appeared consecutively multiple times in response to the physical ornament currently worn by the physical robot not existing in any mapping relationship; the interaction device 30 of the vehicle-mounted robot further includes a binding update module, which is used to determine whether to bind the physical ornament currently worn by the physical robot and the current driving scene as a new mapping relationship based on the determination result, so as to update the scene ornament library based on the new mapping relationship.

[0055] In some disclosed embodiments, if the detection result indicates that the physical robot is not wearing physical accessories, any of the following is performed: ending the process, returning to the step of detecting based on the captured image of the physical robot to obtain the detection result of the physical robot wearing physical accessories in the captured image and looping; and / or, the virtual robot is displayed through at least one of the following methods: vehicle display screen, naked-eye 3D imaging; and / or, the target accessory is directly generated based on feature information.

[0056] Please see Figure 4 , Figure 4 This is a schematic diagram of a framework of an embodiment of the electronic device of this application. The electronic device 40 includes at least a memory 41 and a processor 42 coupled to each other. The memory 41 stores at least program instructions, and the processor 42 is used to execute the program instructions to implement the steps in any of the above-described embodiments of the interaction method for onboard robots. For details, please refer to the foregoing disclosed embodiments, which will not be repeated here.

[0057] Specifically, processor 42 controls itself and memory 41 to implement the steps in any of the above-described embodiments of the interaction method for the onboard robot. Processor 42 can also be referred to as a CPU (Central Processing Unit). Processor 42 may be an integrated circuit chip with signal processing capabilities. Processor 42 can also be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. A general-purpose processor can be a microprocessor or any conventional processor. Furthermore, processor 42 can be implemented using integrated circuit chips.

[0058] In the above scheme, the electronic device 40 detects images of the physical robot to obtain detection results regarding the physical robot wearing physical accessories. The detection results include whether the physical robot is wearing physical accessories and, if so, the feature information of the physical accessories worn by the physical robot. In response to the detection results, which indicate that the physical robot is wearing physical accessories, a virtual accessory matching the physical accessory is obtained as the target accessory based on the feature information of the physical accessory. Based on the current state of the virtual robot wearing the virtual accessory, it is determined whether to prompt the driver or passenger to dress up the virtual robot with the target accessory. Therefore, on the one hand, by detecting accessories on the physical robot and combining it with the current state of the virtual robot wearing accessories, it is determined whether to prompt the driver or passenger to dress up the virtual robot with virtual accessories. This eliminates the need for the driver or passenger to manually search for virtual accessories that match the physical accessories to ensure consistency in appearance, thus reducing the complexity of the operations required to maintain consistency in appearance. On the other hand, by dressing up the virtual robot with the target accessory, it is possible to ensure consistency in appearance between the physical robot and the virtual robot as much as possible when the physical robot is wearing physical accessories. Therefore, while ensuring the consistency of appearance between physical and virtual robots as much as possible, the complexity of the operations required to maintain this consistency can be reduced.

[0059] Please see Figure 5 , Figure 5 This is a schematic diagram of a framework of an embodiment of the computer-readable storage medium of this application. The computer-readable storage medium 50 stores program instructions 51 that can be executed by a processor. The program instructions 51 are used to implement the steps in any of the above-described embodiments of the interaction method of the vehicle-mounted robot.

[0060] In the above scheme, the computer-readable storage medium 50 detects images of the physical robot to obtain detection results regarding the physical robot wearing physical accessories. The detection results include whether the physical robot is wearing physical accessories and, if so, the feature information of the physical accessories worn by the physical robot. In response to the detection results, which indicate that the physical robot is wearing physical accessories, a virtual accessory matching the physical accessory is obtained as the target accessory based on the feature information of the physical accessory. Based on the current state of the virtual robot wearing the virtual accessory, it is determined whether to prompt the driver or passenger to dress up the virtual robot with the target accessory. Therefore, on the one hand, by detecting accessories on the physical robot and combining it with the current state of the virtual robot wearing accessories, it is determined whether to prompt the driver or passenger to dress up the virtual robot with virtual accessories. This eliminates the need for the driver or passenger to manually search for virtual accessories that match the physical accessories to ensure consistency in appearance, thus reducing the complexity of the operations required to maintain consistency in appearance. On the other hand, by dressing up the virtual robot with the target accessory, it is possible to ensure consistency in appearance between the physical robot and the virtual robot as much as possible when the physical robot is wearing physical accessories. Therefore, while ensuring the consistency of appearance between physical and virtual robots as much as possible, the complexity of the operations required to maintain this consistency can be reduced.

[0061] In some embodiments, the functions or modules of the apparatus provided in this disclosure can be used to perform the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0062] The description of the various embodiments above tends to emphasize the differences between the various embodiments. The similarities or similarities between them can be referred to, and for the sake of brevity, they will not be repeated here.

[0063] In the several embodiments provided in this application, it should be understood that the disclosed methods and apparatus can be implemented in other ways. For example, the apparatus implementations described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0064] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0065] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0066] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0067] If the technical solution of this application involves personal information, the product using this technical solution has clearly informed the user of the personal information processing rules and obtained the user's voluntary consent before processing the personal information. If the technical solution of this application involves sensitive personal information, the product using this technical solution has obtained the user's separate consent before processing the sensitive personal information, and also meets the requirement of "express consent". For example, at personal information collection devices such as cameras, clear and prominent signs are set up to inform users that they have entered the scope of personal information collection and that personal information will be collected. If an individual voluntarily enters the collection scope, it is deemed that they have agreed to the collection of their personal information; or on the personal information processing device, with clear signs / information informing users of the personal information processing rules, authorization is obtained from the individual through pop-up information or by asking the individual to upload their personal information; wherein, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the types of personal information processed.

Claims

1. An interaction method for an onboard robot, characterized in that, include: Based on the detection of images captured of the physical robot, detection results are obtained regarding whether the physical robot is wearing physical jewelry in the captured images; wherein, the detection results include whether the physical robot is wearing the physical jewelry, and if so, the feature information of the physical jewelry worn by the physical robot; In response to the detection result indicating that the physical robot is wearing the physical ornament, a virtual ornament that matches the physical ornament is obtained as the target ornament based on the feature information of the physical ornament; Based on the current state of the virtual robot wearing virtual accessories, determine whether to prompt the driver or passenger to dress up the target accessories for the virtual robot.

2. The method according to claim 1, characterized in that, Before determining whether to prompt the driver or passenger to dress the virtual robot in the target accessory based on the current state of the virtual robot wearing the virtual accessory, the method further includes: Obtain the consistency results between the current test results of the physical robot wearing physical jewelry and the previous test results of the physical robot wearing jewelry; The step of determining whether to prompt the driver or passenger to dress up the target accessory for the virtual robot based on the current state of the virtual robot wearing the virtual accessory includes: Based on the current state and the consistency result, determine whether to prompt the driver or passenger to dress up the target accessory for the virtual robot.

3. The method according to claim 2, characterized in that, The step of determining whether to prompt the driver or passenger to dress up the target accessory for the virtual robot based on the current state and the consistency result includes at least one of the following: In response to the current state indicating that the virtual robot is not wearing the target accessory and the consistency result indicating inconsistency, the driver or passenger is prompted to dress up the virtual robot with the target accessory; No prompt will be made if the current state indicates that the virtual robot has worn the target accessory or the consistency result indicates that it is consistent. The current state being characterized as the virtual robot not wearing the target accessory includes any of the following: the virtual robot is not wearing the accessory, or the accessory worn by the virtual robot is not the target accessory.

4. The method according to claim 1, characterized in that, The target accessory is retrieved from a decorative material library using the feature information. The decorative material library defines virtual accessories with several preset field values. The feature information includes at least one of the preset field values. If the feature information is missing some of the preset field values, obtaining a virtual accessory matching the physical accessory as the target accessory based on the feature information of the physical accessory includes: Virtual accessories whose field values ​​match the feature information are selected as candidate accessories; Based on the feature information and the preset field that is missing the field value of the candidate ornament, a query statement is constructed; wherein, the query statement is used to prompt the driver and passenger to provide feedback on the preset field that is missing the field value; Based on the driver's or passenger's response to the inquiry, the feature information is supplemented, and the step of filtering virtual accessories whose field values ​​match the feature information is returned as candidate accessories, until the feature information is no longer missing, and virtual accessories whose field values ​​of each preset field match the feature information are selected as target accessories.

5. The method according to claim 4, characterized in that, The step of constructing a query statement based on the feature information compared to a preset field where the candidate accessory lacks the field value includes: Based on the feature information, the frequency of missing values ​​of each of the candidate ornaments is obtained by comparing them with the preset fields that are missing values. According to the order of missing frequency from high to low, the preset field is selected as the target field; Based on the target field, a query statement is constructed; wherein the query statement is used to ask the driver or passenger for the field value of the target field.

6. The method according to claim 4, characterized in that, The step of detecting physical accessories worn by the robot in the captured images to obtain detection results in the captured images includes: The captured images are detected based on the sample library to obtain the detection results; wherein, the sample library contains several image data of the physical robot wearing physical ornaments, and the image data is labeled with the field values ​​of the physical ornaments worn by the physical robot with respect to each of the preset fields; After the feature information is no longer missing, the method further includes: The sample library is expanded based on the captured images labeled with the aforementioned feature information.

7. The method according to claim 1, characterized in that, When the detection result indicates that the physical robot is wearing the physical ornament, the method further includes: Based on the feature information of the physical accessories currently worn by the physical robot, it is detected whether the physical accessories currently worn by the physical robot have any mapping relationship in the scene accessory library; wherein, the mapping relationship includes driving scenes and physical accessories bound to the driving scenes; In response to the existence of the physical accessory currently worn by the physical robot in the mapping relationship, the mapping relationship in which the physical accessory currently worn by the physical robot exists is selected as the target relationship, and a slogan statement is constructed based on the driving scenario and the physical accessory in the target relationship; The entity robot is controlled to output the slogan statement; wherein the slogan statement is at least used to enhance the driving atmosphere of the driving scene in the target relationship for the driver and passengers by emphasizing the entity ornament in the target relationship.

8. The method according to claim 7, characterized in that, The method further includes: In response to the fact that the physical accessory currently worn by the physical robot does not exist in any of the mapping relationships, it is determined whether the physical accessory currently worn by the physical robot and the current driving scenario have appeared consecutively multiple times; Based on the judgment result, determine whether to bind the physical accessory currently worn by the physical robot and the current driving scene as a new mapping relationship, so as to update the scene accessory library based on the new mapping relationship.

9. The method according to any one of claims 1 to 8, characterized in that, If the detection result indicates that the physical robot is not wearing the physical ornament, perform any of the following: end the process, return to the step of detecting the physical robot based on the captured image to obtain the detection result of the physical robot wearing the physical ornament in the captured image, and repeat the process. And / or, the virtual robot is displayed through at least one of the following methods: vehicle-mounted display screen, naked-eye 3D imaging; And / or, the target ornament is generated directly based on the feature information.

10. An interactive device for a vehicle-mounted robot, characterized in that, include: An image detection module is used to detect objects in an image of a physical robot and obtain detection results regarding whether the physical robot is wearing physical jewelry in the image. The detection results include whether the physical robot is wearing the physical jewelry and, if so, the feature information of the physical jewelry worn by the physical robot. The accessory acquisition module is used to respond to the detection result indicating that the physical robot is wearing the physical accessory, and obtain a virtual accessory that matches the physical accessory as the target accessory based on the feature information of the physical accessory; The dressing-up prompt module is used to determine whether to prompt the driver or passenger to dress up the virtual robot with the target accessory based on the current state of the virtual robot wearing virtual accessories.

11. An electronic device, characterized in that, It includes at least a memory and a processor coupled to each other, wherein the memory stores at least program instructions, and the processor is used to execute the program instructions to implement the interaction method of the vehicle-mounted robot according to any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The system stores program instructions that can be executed by a processor, the program instructions being used to implement the interaction method of the vehicle-mounted robot according to any one of claims 1 to 9.

Images