TW201941099A - A method and its equipment of locking interaction target for intelligent device - Google Patents

Abstract

A method and its equipment of locking interaction target for intelligent device is proposed in this patent. It includes acquiring surrounding images and recognizing target from the intelligent device in its monitoring range, set it as one of candidates, and acquire state information of each candidate and judge its interaction intention with intelligence device. The interaction target of the intelligent device will be chosen from the candidates with interaction intention. In the implement example, the candidates with interaction intention are chosen based on state information of all candidates, and the final interaction target with intelligent device is obtained from the interaction candidates. This method can find the target with the maximum interaction probability for intelligent device and avoid launching intelligent device mistakenly for interacting with target with no interaction intention.

Description

Method and device for determining interactive target of smart device

The invention relates to the technical field of smart devices, and in particular, to a method and a device for determining an interactive target of a smart device.

With the development of smart device technology, the process of active interaction between smart devices and people has been realized. Specifically, the smart device detects a certain range of objects. When a human face is detected, the person is determined as an interaction target, activated and actively interacts with the person.

However, using the above-mentioned method for determining an interactive target by a smart device, a person detected by the smart device may have no willingness to interact with the smart device. For example, someone passes by in front of the smart device. Although a person is detected, there is no intention of interaction. If a smart device is activated when a person is detected, it may cause the smart device to start incorrectly. It can be seen that the above-mentioned method for determining the interactive target by the smart device has low accuracy in determining the interactive target, and is likely to cause a false start.

The present invention aims to solve at least one of the technical problems in the related technology.

The invention proposes a method for determining an interactive target of a smart device, which realizes the screening of targets without interactive intent from candidate targets, and then screens out interactive targets from targets with interactive intent, thereby avoiding selecting targets without interactive intent as The interaction target improves the accuracy of determining the interaction target and reduces the false start of the smart device.

An embodiment of one aspect of the present invention provides a method for determining an interactive target of a smart device, including: acquiring an environment image within a monitoring range of the smart device, and performing target recognition on the environmental image; The identified target is used as a candidate target to obtain the status information of the candidate target; for each candidate target, according to the corresponding status information, determine whether there is an interaction intention to interact with the smart device; and to select from the candidate targets with interaction intention Interaction target of this smart device.

As a possible implementation manner of the embodiment of one aspect of the present invention, the obtaining the status information of the candidate target includes: obtaining a distance between the candidate target and the smart device; and for each candidate target, according to the corresponding status information, Determining whether there is an interaction intent to interact with the smart device includes: for each candidate target, determining whether the distance between the candidate target and the smart device is less than or equal to a preset threshold value for the distance, and at the distance Whether the dwell time within the threshold range exceeds a preset time threshold; if the distance between the candidate target and the smart device is less than or equal to the distance threshold and the dwell time exceeds the time threshold, determine The candidate target has an interaction intention to interact with the smart device.

As a possible implementation manner of an embodiment of the aspect of the present invention, obtaining the status information of the candidate target includes: obtaining a distance between the candidate target and the smart device, and a face angle of the candidate target; Candidate targets, according to corresponding state information, to determine whether there is an interaction intention to interact with the smart device, including: for each candidate target, determining whether the distance between the candidate target and the smart device is less than or equal to a preset distance threshold Whether the face angle of the candidate target is within a preset angle range; if the distance between the candidate target and the smart device is less than or equal to a preset distance threshold, and the face angle of the candidate target Within the preset angle range, it is determined that the candidate target has an interaction intention to interact with the smart device.

As a possible implementation manner of the embodiment of one aspect of the present invention, selecting an interactive target of the smart device from the candidate targets having the interaction intention includes: when a plurality of candidate targets are detected, and the candidate targets having the interaction intention are plural At this time, from among the plurality of candidate targets having an interaction intention, a candidate target closest to the smart device is determined; and among the candidate targets closest to the smart device, an interactive target of the smart device is selected.

As a possible implementation manner of the embodiment of one aspect of the present invention, selecting an interactive target of the smart device from the candidate targets closest to the smart device includes: when the candidate target closest to the smart device is When plural, query whether there is a face image of a candidate target closest to the smart device in the face image database of the registered user of the smart device; The face image of the closest candidate target of the smart device, then use the candidate target closest to the smart device as an interaction target; if there is no candidate closest to the smart device in the face image library A face image of the target, randomly select a candidate target closest to the smart device as an interactive target; if there are a plurality of face images of the candidate target closest to the smart device in the face image library , The candidate target closest to the smart device that is first queried is used as the interactive target.

As a possible implementation manner of an embodiment of the aspect of the present invention, obtaining the distance between the candidate target and the smart device includes: obtaining a depth map through a depth camera in the smart device, and obtaining the depth map according to the depth map. The distance between the target and the smart device; or, shooting the candidate target through the binocular vision camera in the smart device, calculating the parallax of the image captured by the binocular vision camera, and calculating based on the parallax The distance between the candidate target and the smart device; or, a laser is emitted into the monitoring range through a laser radar in the smart device; according to the laser returned by each obstacle within the monitoring range, generated Binary map of each obstacle; Fusion each binary map with the environment image to identify the binary map corresponding to the candidate target from all binary maps; according to the binary value corresponding to the candidate target The laser return time of the figure determines the distance between the candidate target and the smart device.

As a possible implementation manner of an embodiment of the aspect of the present invention, obtaining the face angle of the candidate target includes: intercepting a face image of the candidate target from the environment image; and inputting the face image into a pre-training In a good machine learning model, the face angle of the face in the face image is obtained. The method further includes: training the machine learning model in the following manner: collecting and carrying a sample face image, wherein the sample face image The image carries annotation data, which is used to represent the face angle of the sample face; the sample face image is input into the initially built machine learning model for training. When the error of the machine learning model after training is within When the preset error range is within, a trained machine learning model is obtained.

As a possible implementation manner of the embodiment of one aspect of the present invention, after selecting the interaction target of the smart device from the candidate targets having an interaction intention, the method further includes: controlling the smart device to interact with the interaction target; during the interaction process Identifying the center point of the face image of the interaction target; detecting whether the center point of the face image is within a preset image area; if it is not within the image area, obtaining the The path between the center point reaching the center point of the image area; according to the path, controlling the smart device so that the center point of the face image is within the image area.

The method for determining an interactive target of a smart device according to an embodiment of the present invention obtains an environmental image within the monitoring range of the smart device, performs target recognition on the environmental image, and uses the target identified from the environmental image as a candidate target. To obtain the status information of the candidate target, and for each candidate target, determine whether there is an interaction intention to interact with the smart device according to the corresponding status information, and select the interaction target of the smart device from the candidate targets having the interaction intention. In this embodiment, according to the status information of the candidate targets, from all candidate targets, candidate targets with interaction intentions are screened out, and among the candidate targets with interaction intentions, interactive targets are selected for smart devices, so that the selected The interaction target is most likely to have the interaction intention with the smart device, avoiding the target without the interaction intention as the interaction target, improving the accuracy of determining the interaction target, and reducing the false startup of the smart device.

An embodiment of another aspect of the present invention provides an interactive target determination device for a smart device, including: a first acquisition module, configured to select an interactive target of the smart device from a candidate target having an interactive intent, and then obtain the smart target The environment image within the monitoring range of the type device performs target recognition on the environment image; the second acquisition module is configured to obtain the target information of the candidate target from the target identified in the environmental image ; A judging module for each candidate target and judging whether there is an interaction intention to interact with the smart device according to the corresponding state information; a selection module for selecting the smart type from the candidate targets having an interaction intention Interaction target of the device.

As a possible implementation manner of another embodiment of the present invention, the second acquisition module is specifically configured to: acquire a distance between the candidate target and the smart device; the judgment module is specifically configured to: for each candidate A target to determine whether the distance between the candidate target and the smart device is less than or equal to a preset threshold value of the distance, and whether the length of stay within the range of the threshold value exceeds the preset time threshold value; if the If the distance between the candidate target and the smart device is less than or equal to the distance critical value and the dwell time exceeds the time critical value, it is determined that the candidate target has an interaction intention to interact with the smart device.

As a possible implementation manner of another embodiment of the present invention, the second obtaining module is specifically configured to: obtain a distance between the candidate target and the smart device, and a face angle of the candidate target; the judgment mode The group is specifically used to: For each candidate target, determine whether the distance between the candidate target and the smart device is less than or equal to a preset distance threshold, and whether the candidate's face angle is within a preset angle range Inside;

If the distance between the candidate target and the smart device is less than or equal to a preset distance threshold, and the face angle of the candidate target is within a preset angle range, it is determined that the candidate target exists with the smart device Interactive intent.

As a possible implementation manner of another embodiment of the present invention, the selection module includes: a determining unit, configured to: when a plurality of candidate targets are detected, and when the candidate targets having the interaction intention are plural, the interaction intention exists from the plural. Among the candidate targets, a candidate target closest to the smart device is determined; and a selecting unit is configured to select an interactive target of the smart device from the candidate target closest to the smart device.

As a possible implementation manner of another embodiment of the present invention, the selection unit is specifically configured to: when the candidate target closest to the smart device is plural, query the registered user face image of the smart device Whether a face image of a candidate target closest to the smart device exists in the database; if a face image of a candidate target closest to the smart device exists in the face image library, the one The candidate candidate closest to the smart device is used as the interaction target; if the face image library does not exist in the face image of the candidate candidate closest to the smart device, a random distance to the smart device is selected randomly. The nearest candidate target is used as the interactive target; if there are multiple face images of the candidate target closest to the smart device in the face image library, the candidate closest to the smart device will be searched first. Goals serve as interactive goals.

As a possible implementation manner of another embodiment of the present invention, the second acquisition module is specifically configured to: acquire a depth map through a depth camera in the smart device, and obtain the target and the smart device according to the depth map. The distance between them; or, shooting the candidate target through the binocular vision camera in the smart device, calculating the parallax of the image captured by the binocular vision camera, and calculating the candidate target and the smart type according to the parallax The distance between the devices; or, a laser is emitted into the monitoring range by a laser radar in the smart device; according to the laser returned by each obstacle within the monitoring range, a binary value of each obstacle is generated Figure; Fusion each binary image with the environment image, identify the binary image corresponding to the candidate target from all binary images; determine the laser return time of the binary image corresponding to the candidate target, and determine Find out the distance between the candidate target and the smart device.

As a possible implementation manner of another embodiment of the present invention, the second acquisition module is specifically configured to: intercept a face image of the candidate target from the environment image; input the face image into a pre-trained In a machine learning model, the face angle of the face in the face image is obtained; the device further includes: an acquisition module for collecting a face image carrying a sample, wherein the sample face image carries annotations Data, the labeled data is used to represent the face angle of the sample face; the training module is used to input the sample face image into the initially constructed machine learning model for training. When the machine learning model is trained, When the error is within the preset error range, the trained machine learning model is obtained.

As a possible implementation manner of another embodiment of the present invention, the device further includes: a first control module for controlling the smart device to interact with the interaction target; an identification module for use in the interaction process, Identify the center point of the face image of the interaction target; a detection module for detecting whether the center point of the face image is within a preset image area; a third acquisition module for detecting the When in the image area, the path between the center point of the face image and the center point of the image area is obtained; a second control module is used to control the smart device according to the path to make the face The center point of the image is within the image area.

The interactive target determining device of the smart device according to the embodiment of the present invention obtains an environmental image within the monitoring range of the smart device, performs target recognition on the environmental image, and uses the target identified from the environmental image as a candidate target. To obtain the status information of the candidate target, and for each candidate target, determine whether there is an interaction intention to interact with the smart device according to the corresponding status information, and select the interaction target of the smart device from the candidate targets having the interaction intention. In this embodiment, according to the status information of the candidate targets, from all candidate targets, candidate targets with interaction intentions are screened out, and among the candidate targets with interaction intentions, interactive targets are selected for smart devices, so that the selected The interaction target is most likely to have the interaction intention with the smart device, avoiding the target without the interaction intention as the interaction target, improving the accuracy of determining the interaction target, and reducing the false startup of the smart device.

Another embodiment of the present invention provides a smart device, which includes a housing, a processor, a memory, a circuit board, and a power circuit. The circuit board is disposed in a space surrounded by the housing, and the processor And the memory is arranged on the circuit board; the power circuit is used for supplying power to each circuit or device of the smart device; the memory is used for storing executable program code; wherein the processor reads the memory by The executable code stored in the body is used to run a program corresponding to the executable code, so as to implement the method for determining an interactive target of the smart device according to the embodiment of the above aspect.

Another embodiment of the present invention provides a computer program product. The computer program product includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a processor, the program instructions are implemented as described above. Aspects The method for determining an interactive target of a smart device.

Another embodiment of the present invention provides a non-transitory computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the interactive target determination of the smart device according to the embodiment of the above aspect is determined. method.

Additional aspects and advantages of the present invention will be given in part in the following description, part of which will become apparent from the following description, or be learned through the practice of the present invention.

An embodiment of the present invention is described in detail below. An example of the embodiment is shown in the drawings, wherein the same or similar reference numerals indicate the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the drawings are exemplary and are intended to explain the present invention, but should not be construed as limiting the present invention.

The following describes a method and an apparatus for determining an interactive target of a smart device according to an embodiment of the present invention with reference to the accompanying drawings.

According to the embodiments of the present invention, when a smart device detects a human face, a method for determining an interactive target using a person as an interactive target may use a target that does not have the willingness to interact with the smart device as an interactive target, thereby causing a smart device The problem of mis-starting, this paper proposes an interactive target determination method for smart devices.

The method for determining an interactive target of a smart device according to an embodiment of the present invention selects candidate targets with interaction intent from all candidate targets according to the status information of the candidate targets, and further selects intelligent targets from candidate targets with interactive intent. The device selects the interaction target, so that the selected interaction target is most likely to have the interaction intention with the smart device, avoiding the target without the interaction intention as the interaction target, improving the determination accuracy of the interaction target, and reducing the smart device. By mistake.

FIG. 1 is a schematic flowchart of a method for determining an interactive target of a smart device according to an embodiment of the present invention.

As shown in FIG. 1, the method for determining an interactive target of a smart device includes: Step 101: Obtain an environmental image within a monitoring range of the smart device, and perform target recognition on the environmental image.

In this embodiment, the smart device may be a robot, a smart home appliance, or the like.

The smart device is equipped with a camera device, such as a camera. The smart device can immediately capture the environment image within the monitoring range through the camera device. After the environment image is acquired, the environment image can be detected to identify targets that enter the monitoring range. Among them, the goal here can be understood as people.

Taking the recognition of people in an environment image as an example, the smart device can recognize the people in the environment image through face detection or human body detection. Specifically, an outline of an object is extracted from an environment image, and the extracted outline of the object is compared with a pre-existing face contour or a contour of a human body. When the similarity between the extracted contour and the preset contour exceeds a preset threshold, it can be considered that a person is recognized from the environmental image. Therefore, all people in the environment image can be identified by this method.

Step 102: Obtain the status information of the candidate target from the target identified in the environmental image.

In this embodiment, when a target is identified from an environment image, the identified target is used as a candidate target. For example, when someone enters the surveillance range of the robot, the robot can identify people who enter the surveillance range from the collected environmental images, and these people are all candidate targets.

After identifying the target in the environment image, obtain the status information of the candidate target, such as the position of the target, the target's staying time within the distance threshold, the number of times the target is recognized within a preset time period, etc. The status information of the candidate target determines whether the candidate target has an interaction intention to interact with the smart device.

Step 103: For each candidate target, determine whether there is an interaction intention to interact with the smart device according to the corresponding state information.

In related technologies, after a smart device recognizes a human face, it directly uses the human as an interaction target to interact with the human. However, the person identified by the smart device may not have the willingness to interact with the smart device, which may cause a false activation.

In this embodiment, for each candidate target, it is determined whether the candidate target has an interaction intention according to the status information of the candidate target.

As a possible implementation manner, the number of times a candidate target is recognized within a preset time period is obtained, and the number of times is compared with the preset number of times. If the number of times the target is recognized within a preset time period is greater than the preset number of times, it can be considered that the target often appears and there is an interaction intention with the smart device.

For example, in the past month, the number of times a robot at the front desk of a company recognized a person was 4 times greater than a preset number of times, indicating that the person is a frequent visitor to the company, and it can be determined that there is an interaction intention between the person and the robot.

In this embodiment, according to the state information of candidate targets, candidate targets without interaction intention are filtered out from the candidate targets, so that it is possible to avoid determining targets without interaction intention as interaction targets.

Step 104: Select an interactive target of the smart device from the candidate targets having an interaction intention.

In this embodiment, in order to further improve the accuracy of determining the interaction target and reduce the probability of the intelligent device being started incorrectly, the interaction target of the intelligent device may be continued to be selected from the candidate targets having the interaction intention, thereby making the interaction target the most likely Candidates for interaction intent.

If there is only one candidate target with interaction intention, the candidate target is taken as the interaction target. If there are plural targets with interaction intention, the interaction target may be determined according to the distance between the candidate target and the smart device. The subsequent embodiments of the specific process will be described in detail, and will not be repeated here.

On the basis of the above embodiment, for step 103, it is determined whether there is an interaction intention to interact with the smart device according to the corresponding state information. As a possible implementation manner, the distance between the candidate target and the smart device may be determined , And the dwell time of the candidate target within a preset distance critical value range to determine whether the candidate target has an interaction intention. FIG. 2 is a schematic flowchart of another method for determining an interactive target of a smart device according to an embodiment of the present invention.

As shown in FIG. 2, the method for determining an interactive target of a smart device includes: Step 201: Obtain an environmental image within a monitoring range of the smart device, and perform target recognition on the environmental image.

In this embodiment, for a smart device to acquire an environment image within a monitoring range and a method for performing target recognition by the environment image, refer to related content recorded in the foregoing embodiment, and details are not described herein again.

In step 202, the target identified from the environment image is used as a candidate target, and the distance between the candidate target and the smart device is obtained.

It can be understood that the closer the distance between the candidate target and the smart device, the greater the possibility of interaction intention between the candidate target and the smart device. Therefore, in this embodiment, the candidate target and the smart device are The distance between them is used as one of the basis for judging whether the candidate target exists and the interaction intention of interacting with the smart device.

In this embodiment, the distance between the candidate target and the smart device may be obtained through a depth camera, a binocular vision camera, or a laser radar.

As a possible implementation manner, a depth camera is configured in the smart device, and a depth map of the candidate target is obtained through the depth camera. In specific implementation, a structured light projector can be used to project a controllable light spot, light bar, or light surface structure on a candidate target surface, and an image is obtained by an image sensor in the depth camera. The geometric relationship is used to use the triangle The principle calculates the three-dimensional coordinates of the candidate target, so that the distance between the candidate target and the smart device can be obtained.

As another possible implementation manner, a binocular vision camera is configured in the smart device, and the candidate target is photographed through the binocular vision camera. Then, the parallax of the image captured by the binocular vision camera is calculated, and the distance between the candidate target and the smart device is calculated according to the parallax.

FIG. 3 is a schematic diagram of a binocular vision calculation distance provided by an embodiment of the present invention. In Figure 3, the positions of the two cameras are drawn in the actual space. with , And the optical axes of the left and right cameras, the focal planes of the two cameras, and the distance between the focal planes and the planes of the two cameras are .

As shown in Figure 3, with The positions of the same candidate target P in different captured images are respectively. among them, The distance from the point to the left border of the captured image is , The distance from the point to the left border of the captured image is . with There are two cameras, which are on the same plane, and the distance between the two cameras is .

Based on the principle of triangulation, the distance between P in the third image and the plane where the two cameras are located Has the following relationship: ;

Based on this, it can be deduced ,among them, The visual difference of the images taken by the binocular camera of the same candidate target. due to , Is a fixed value, so Can determine the distance between the candidate target and the plane where the camera is located , The distance between the candidate and the smart device.

As another possible implementation manner, a laser radar is configured in the smart device, and the laser is used to emit laser light into the monitoring range. The emitted laser light is reflected when it encounters an obstacle in the monitoring range. The smart device receives the laser beam returned by each obstacle in the monitoring range, and generates a binary map of each obstacle according to the returned laser beam. Then, each binary image is fused with the environment image, and the binary image corresponding to the candidate target is identified from all the binary images. Specifically, the contour or size of each obstacle can be identified according to the binary map of each obstacle, and then the contour or size of each target in the environment image can be matched to obtain the binary value corresponding to the candidate target. Illustration. Then, the laser return time of the binary map corresponding to the candidate target is multiplied by the speed of light and divided by 2 to obtain the distance between the candidate target and the smart device.

It should be noted that other methods for calculating the distance between the candidate target and the smart device are also included in the protection scope of the embodiment of the present invention.

Step 203: For each candidate target, determine whether the distance between the candidate target and the smart device is less than or equal to a preset distance threshold, and whether the dwell time within the distance threshold exceeds the preset time threshold. .

Because when the distance between the candidate target and the smart device is long, the candidate target may not have an interaction intention to interact with the smart device, or the distance is short, but the short stay time of the candidate target may not exist with the smart device. Interactive intent of device interaction.

Therefore, for each candidate target, the distance between the candidate target and the smart device can be compared with a preset distance threshold to determine whether the distance between the selected target and the smart device is less than or equal to the preset The critical distance. If the distance is within the distance critical value range, it is judged whether the time that the candidate target stays within the distance critical value range exceeds a preset time critical value.

In step 204, if the distance is less than or equal to the distance critical value and the staying time exceeds the time critical value, it is determined that the candidate target has an interaction intention to interact with the smart device.

When the distance between the candidate target and the smart device is less than a preset distance threshold, and the stay time of the candidate target within the range of the distance threshold exceeds the preset time threshold, the candidate target may be considered to interact with the smart device. Interactive intent.

Taking a robot as an example, if the distance between the human and the robot is less than 3 meters, and the human stays within 3 meters for more than 2 seconds, it can be considered that the human has an interaction intention to interact with the robot.

Step 205: Select an interactive target of the smart device from the candidate targets having an interaction intention. In this embodiment, step 205 is similar to step 104 in the foregoing embodiment, so it will not be repeated here.

The method for determining an interactive target of a smart device according to the embodiment of the present invention selects all candidate targets by the distance between the candidate target and the smart device and the dwell time of the candidate target within a preset distance critical range. There are candidate targets for interaction intent to interact with the smart device. Compared with directly detecting people as the target of interaction when a human face is detected, false start of the smart device can be reduced.

For step 103, as another possible implementation manner, whether the candidate target has an interaction intention to interact with the smart device may also be determined according to the distance between the candidate target and the smart device and the face angle of the candidate target. FIG. 4 is a schematic flowchart of another method for determining an interactive target of a smart device according to an embodiment of the present invention.

As shown in FIG. 4, the method for determining an interactive target of a smart device includes: Step 301: Obtain an environmental image within a monitoring range of the smart device, and perform target recognition on the environmental image.

In this embodiment, step 301 is similar to step 101 in the foregoing embodiment, so details are not described herein again.

Step 302: Obtain a target identified from the environment image as a candidate target, and obtain a distance between the candidate target and the smart device, and a face angle of the candidate target.

In practice, when a person passes by the robot, if the person turns his head to look at the robot, or when the human face is facing the robot, it means that the person pays more attention to the robot, and the person has an interaction intention to interact with the robot. Therefore, the face angle of the face in the face image can be used as one of the basis for judging whether the candidate target has an interaction intention to interact with the smart device.

In this embodiment, the distance between the candidate target and the smart device and the face angle of the candidate target can be used to determine whether the candidate target has an interaction willingness to interact with the smart device. When obtaining the distance between the candidate target and the smart device, the distance between the candidate target and the smart device may be obtained by using the method described in the foregoing embodiment.

When obtaining the face angle, the face angle can be obtained through a pre-trained machine learning model. Specifically, the face image of the candidate target may be intercepted from the environment image according to the face contour, and then the face image is input into a machine learning model. The machine learning model outputs the face angle in the face image based on the face image.

The face angle may be an angle where the central axis of the face deviates from the central axis of the face image. The central axis of the face includes the central axis in the horizontal direction and the central axis in the vertical direction. The central axis of the corresponding facial image also includes the horizontal direction And the vertical axis. From the face image, the horizontal axis of the horizontal direction of the face and the vertical axis of the vertical direction can be identified, and the angles deviating from the horizontal axis of the horizontal direction of the face image and the vertical axis of the face image are obtained, The angle reached is the face angle.

In this embodiment, a machine learning model may be trained in the following manner. First, a face image is collected, and the face angle is labeled on the face image, so that the sample face image carries label data indicating the face angle of the sample face image. Then, the sample face images are input into the initially constructed machine learning model for training. When the difference between the face angle output by the machine learning model and the labeled face angle is within a preset error range, it can be considered that the machine learning model has been trained.

In this embodiment, obtaining the face angle by using a trained machine learning model can improve the accuracy of the obtained face angle, thereby improving the accuracy of subsequent judgments.

Step 303: For each candidate target, determine whether the distance between the candidate target and the smart device is less than or equal to a preset distance threshold, and whether the face angle of the candidate target is within a preset angle range.

In this embodiment, for each candidate target, the distance between the candidate target and the smart device is compared with a preset distance threshold, and the face angle of the candidate target is compared with a preset upper limit value of the angular range. Compare.

Assuming the critical distance value is 3 meters and the angle range is [0 °, 45 °], determine whether the distance between the candidate target and the smart device is less than 3 meters, and compare the face angle with 45 ° to determine the face angle Whether it is within the preset angle range.

In step 304, if the distance between the candidate target and the smart device is less than or equal to a preset distance threshold, and the face angle of the candidate target is within a preset angle range, it is determined that the candidate target interacts with the smart device. Interaction intent.

In this embodiment, when the distance between the candidate target and the smart device is less than or equal to a preset distance critical value, and the face angle of the candidate target is within a preset angle range, it indicates that the candidate target is within the distance critical value range. By paying attention to the smart device, it can be determined that the candidate target has an interaction intention to interact with the smart device. Compared with directly detecting a person as an interactive target, the accuracy of the interactive target confirmation is improved.

Step 305: Select an interactive target of the smart device from the candidate targets having an interaction intention.

In this embodiment, step 305 is similar to step 104 in the foregoing embodiment, and details are not described herein again.

According to the method for determining an interactive target of a smart device according to an embodiment of the present invention, from all candidate targets, the existence of an interaction intention to interact with the smart device is screened from all candidate targets based on the distance between the candidate target and the smart device, and the face angle of the candidate target. Candidate targets can reduce the false start of smart devices compared with directly using people as interactive targets when detecting faces.

It should be noted that when judging whether the candidate target has an interaction intention, the distance between the candidate target and the smart device can also be within the distance threshold, and the time during which the candidate target stays within the distance threshold exceeds the time. When the threshold value and the face angle of the candidate target are within a preset angle range, it is determined that the candidate target has an interaction intention to interact with the smart device. Otherwise, it can be considered that the candidate target does not have an interaction intention to interact with the smart device.

In the above embodiment, for the interactive target of the smart device selected from the candidate targets with interactive intent, when there is only one candidate target with interactive intent, the candidate target with interactive intent may be used as the interactive target. When there are plural candidate targets with interaction intent, the interactive target may be selected from the candidate targets according to the distance between the candidate target and the smart device. FIG. 5 is a schematic flowchart of another method for determining an interactive target of a smart device according to an embodiment of the present invention.

As shown in FIG. 5, for step 104, the method for determining an interactive target of a smart device may include: Step 401: When a plurality of candidate targets are detected, and when a candidate target having an interaction intent is plural, an interaction intention exists from the plural. Among the candidate targets, a candidate target closest to the smart device is determined.

The closer the distance between the candidate target and the smart device, the stronger the intention of interaction between the candidate target and the smart device.

In this embodiment, when the smart device detects a plurality of candidate targets from the environment image, and determines that the candidate targets with interaction intent are also plural, the number of candidate targets with interaction intention and the smart device may be The distances are compared to find the candidate target closest to the smart device from the plurality of candidate targets with interaction intent, so as to screen out the candidate target with strong interaction intent.

Step 402: Select an interactive target of the smart device from the candidate targets closest to the smart device.

In this embodiment, in order to further determine the interaction target of the smart device, it is necessary to select the interaction target of the smart device from among the targets closest to the smart device.

It can be understood that when there is only one candidate target closest to the smart device, the candidate target can be used as the interactive target of the smart device.

When there are a plurality of candidate targets closest to the smart device, it is necessary to select an interactive target of the smart device from the plurality of candidate targets closest to the smart device.

Taking a robot as an example, a robot is placed at the front desk of a company. When a user needs to enter the company, he can register information in the robot, that is, register in the robot. Alternatively, the facial images of registered users can be downloaded from the company website and stored in the robot. Users who have registered on the company website can be registered in the robot synchronously. Generally, the registered users in the robot have stronger interaction intentions than the unregistered users interacting with the robot. Therefore, the interactive target of the smart device can be determined according to whether the candidate target is registered.

In the daily reception work, the robot can collect the face images of visitors or company employees, and use the collected face images of visitors or company employees to build a registered user face image library, or use the website to register and use The person's face image to build the face image library.

As a possible implementation manner, the smart device may locally query whether the candidate candidate closest to the smart device has already registered the smart device. Specifically, the smart device may store a face image database of registered users in advance, and the face image database stores user face images of registered smart devices. When the candidate target closest to the smart device is plural, the face image of the candidate target closest to the smart device may be compared with the face image in the face image library. If a face image of a candidate target closest to the smart device is stored in the face image library, indicating that the candidate target is registered, then the candidate target is used as the interactive target of the smart device.

If there is no face image of the candidate target closest to the smart device in the face image library, it means that the candidate target closest to the smart device is not registered, and the candidate target closest to the smart device can be selected. , Randomly select a candidate target as the interactive target.

If there is a face image of a plurality of candidate targets closest to the smart device in the face image library, indicating that a plurality of candidate targets closest to the smart device have been registered, then the first query and the smart The candidate target closest to the device is used as the interaction target, and a candidate target may also be randomly selected as the interactive target from the candidate targets registered and closest to the smart device.

As another possible implementation manner, when the candidate object closest to the smart device is plural, the face images of all the candidate objects closest to the smart device can be sent to the server, and the server returns the query The result is to the smart device, and the smart device determines the interaction target according to the comparison result.

Specifically, the server stores a face image database of registered users. When the candidate object closest to the smart device is a plural number, the smart device compares the face number of the plural candidate object to the smart device. Image, send to server. The server receives the face image, and inquires whether there is a face image of a candidate target closest to the smart device in the registered user face image database. The server then sends the query results to the smart device. The smart device determines the interaction target of the smart device according to the query result. For a specific determination method, refer to the foregoing method, and details are not described herein again.

For example, A passed by in front of the robot and had no intention of interaction, while B was a frequent visitor to the company and had previously registered. When the distance between A and B and the robot is less than the critical distance of 3 meters and the same distance as the robot, the robot can select the registered B as the interactive target and say hello to B.

In the method for determining an interactive target of a smart device according to the embodiment of the present invention, when there is a plurality of candidate targets having an interaction intention, a candidate target closest to the smart device is selected, and when there are a plurality of targets closest to the smart device, By querying the registered user's face image database and selecting the interactive target of the smart device according to the query result, in the related technology, the smart device may not select the interactive target that is most likely to People with interactive intentions in the device, thereby improving the accuracy of determining the interaction target and avoiding the false start of the intelligent device.

In practice, after the smart device determines the interactive target, the interactive target may be in a moving state during the interaction between the smart device and the interactive target. For example, during the process of the robot greeting the interactive target, the human may be in a moving state. In order to enable the smart device to maintain the front-following interaction feature, the embodiment of the present invention further proposes that during the interaction process, the center point of the face image is located in the image area. FIG. 6 is a schematic flowchart of another method for determining an interactive target of a smart device according to an embodiment of the present invention.

After selecting the interactive target of the smart device from the candidate targets with interactive intent, as shown in FIG. 6, the method for determining an interactive target of the smart device further includes: Step 105, controlling the smart device to interact with the interactive target.

In this embodiment, after the interaction target is determined, the smart device starts and interacts with the interaction target. Take a robot as an example. After the robot determines an object to greet, it starts and greets an interactive target, such as "Welcome."

Step 106: During the interaction, identify the center point of the face image of the interaction target.

The face image of the target may be an image of the smallest area of the target image in the environment image.

In this embodiment, during the interaction process, the smart device instantly recognizes the center point of the face image of the interaction target. The center point of the face image is the intersection of the vertical center line and the horizontal center line of the face image.

In step 107, it is detected whether the center point of the face image is within a preset image area. In this embodiment, the preset image area may be a circle area obtained by drawing a circle at a preset size with a center point of the environment image as a circle center. The preset size may be a half of a horizontal size of a face image when a person is at a distance threshold. Of course, it can also be set as required.

The smart device can detect whether the center point of the face image is within the preset image area every preset time, such as every 0.5 seconds, to determine whether the face image is within the preset image area. .

In step 108, if it is not in the image area, a path between the center point of the face image and the center point of the image area is obtained.

In this embodiment, if the center point of the face image is not in the image area, it indicates that the face image that the smart device can capture is not complete, then the center point of the face image is obtained to the center point of the image area Between paths.

Step 109: Control the smart device according to the path so that the center point of the face image is within the image area.

After the smart device obtains a path between the center point of the face image and the center point of the image area, the smart device is controlled according to the path so that the center point of the face image is within the image area.

As a possible implementation method, the center point of the image area can be used as the center of the circle, a rectangular coordinate system can be established, the coordinates of the center point of the face image can be obtained, and the center point of the face image and the center point of the image area can be calculated. The distance between them, and the angle between the center point of the face image and the horizontal direction. After that, control the smart device to rotate the corresponding angle and distance.

Take the robot as an example. If the robot detects the center point of the face image, to the right of the center point of the image area, it means that the person is gradually moving to the right, then the PTZ and chassis of the control robot are rotated to the right to follow the person. To achieve the purpose of gaze.

The method for determining an interactive target of a smart device according to the embodiment of the present invention realizes that the smart device follows the interactive target by detecting whether the center point of the face image is within a preset image area, and enables the smart device to interact with people. Time is more vivid and flexible.

In order to implement the foregoing embodiments, an embodiment of the present invention further provides an interactive target determination device for a smart device. FIG. 7 is a schematic structural diagram of an interactive target determination device for a smart device according to an embodiment of the present invention.

As shown in FIG. 7, the interactive target determination device of the smart device includes a first acquisition module 510, a second acquisition module 520, a determination module 530, and a selection module 540.

The first acquisition module 510 is configured to acquire an environment image within a monitoring range of the smart device, and perform target recognition on the environment image.

The second acquisition module 520 is configured to acquire a target identified from the environment image as a candidate target, and obtain status information of the candidate target.

The determination module 530 is configured to determine, for each candidate target, whether there is an interaction intention to interact with the smart device according to the corresponding state information.

The selection module 540 is configured to select an interaction target of the smart device from the candidate targets having an interaction intention.

In a possible implementation manner of this embodiment, the second obtaining module 520 is specifically configured to: obtain a distance between the candidate target and the smart device; the determining module 530 is specifically configured to: determine a candidate for each candidate target Whether the distance between the target and the smart device is less than or equal to a preset distance threshold, and whether the length of the stay within the distance threshold exceeds the preset time threshold; if the distance between the candidate target and the smart device is If the distance is less than or equal to the distance critical value and the staying time exceeds the time critical value, it is determined that the candidate target has an interaction intention to interact with the smart device. In a possible implementation manner of this embodiment, the second acquisition module 520 is specifically configured to: acquire the distance between the candidate target and the smart device, and the face angle of the candidate target; the judgment module 530 is specifically configured to: For each candidate target, determine whether the distance between the candidate target and the smart device is less than or equal to a preset distance threshold, and whether the face angle of the candidate target is within a preset angle range. If the distance between them is less than or equal to a preset distance threshold and the face angle of the candidate target is within the preset angle range, it is determined that the candidate target has an interaction intention to interact with the smart device.

In a possible implementation manner of this embodiment, the selection module 540 includes: a determining unit, configured to: when a plurality of candidate targets are detected, and when the candidate target having an interaction intention is a plural number, select from the plurality of candidate targets having the interaction intention. To determine a candidate target closest to the smart device; and a selecting unit for selecting an interactive target of the smart device from the candidate targets closest to the smart device.

In a possible implementation manner of this embodiment, the selecting unit is specifically configured to: when there are plural candidate targets closest to the smart device, query whether there is a smart phone in the registered user face image database of the smart device. Face image of the closest candidate target of the smart device; if a face image of a candidate target closest to the smart device exists in the face image library, a candidate target closest to the smart device is taken as Interactive target; if there is no face image of the candidate target closest to the smart device in the face image library, randomly select a candidate target closest to the smart device as the interactive target; if the face image library In the face image of a plurality of candidate targets closest to the smart device in the face, the candidate target closest to the smart device first queried is used as the interaction target. In a possible implementation manner of this embodiment, the second obtaining module 520 is specifically configured to: obtain a depth map through a depth camera in the smart device, and obtain the distance between the target and the smart device according to the depth map; or, Use the binocular vision camera in the smart device to shoot the candidate target, calculate the parallax of the image captured by the binocular vision camera, and calculate the distance between the candidate target and the smart device according to the parallax; or, Laser radar that emits laser light into the monitoring range; based on the laser light returned by each obstacle in the monitoring range, a binary map of each obstacle is generated; each binary map is fused with the environmental image, from All the binary maps corresponding to the candidate target are identified; according to the laser return time of the binary map corresponding to the candidate target, the distance between the candidate target and the smart device is determined.

In a possible implementation manner of this embodiment, the second acquisition module 520 is specifically configured to: intercept a face image of a candidate target from an environment image; and input the face image into a pre-trained machine learning model, Obtain a face angle of a face in a face image. In a possible implementation manner of this embodiment, the device further includes: a collection module configured to collect a sample face image, where the sample face image Carrying labeled data, the labeled data is used to represent the face angle of the sample face; the training module is used to input the sample face image into the initially constructed machine learning model for training, and when the trained machine learning model When the error is within a preset error range, a trained machine learning model is obtained.

In a possible implementation manner of this embodiment, the device further includes: a first control module, configured to control the intelligent device and the interaction target to perform the interaction after selecting the interaction target of the intelligent device from the candidate targets having an interaction intention. Interaction; a recognition module for identifying the center point of the face image of the interaction target during the interaction process; a detection module for detecting whether the center point of the face image is within a preset image area; Three acquisition modules are used to acquire the path between the center point of the face image and the center point of the image area when not in the image area; the second control module is used to control the intelligent type according to the path Device so that the center point of the face image is within the image area.

It should be noted that, the foregoing description of the embodiment of the method for determining an interactive target of a smart device is also applicable to the device for determining an interactive target of a smart device in this embodiment, so it will not be repeated here.

The interactive target determining device of the smart device according to the embodiment of the present invention obtains an environmental image within the monitoring range of the smart device, performs target recognition on the environmental image, and uses the target identified from the environmental image as a candidate target. To obtain the status information of the candidate target, and for each candidate target, determine whether there is an interaction intention to interact with the smart device according to the corresponding status information, and select the interaction target of the smart device from the candidate targets having the interaction intention. In this embodiment, according to the status information of candidate targets, from all candidate targets, candidate targets with interaction intent are screened out, and among the candidate targets with interaction intent, interactive targets are selected for smart devices, so that the selected interactions The target is most likely a target with interactive intention with the smart device, avoiding the target without interactive intention as the interactive target, improving the accuracy of determining the interactive target, and reducing the false start of the smart device.

In order to implement the foregoing embodiments, an embodiment of the present invention further provides a smart device.

FIG. 8 is a schematic structural diagram of an embodiment of a smart device according to the present invention. As shown in FIG. 8, the smart device may include a housing 610, a processor 620, a memory 630, a circuit board 640, and a power circuit 650. The circuit board 640 is disposed in the space surrounded by the housing 610, and the processor 620 and the memory 630 are disposed on the circuit board 640. The power circuit 650 is used to supply power to the circuits or components of the smart device. The memory 630 is used to store executable code; the processor 620 reads the executable code stored in the memory 630 to run a program corresponding to the executable code, and is used to execute the interaction of the smart device described in the foregoing embodiment Goal determination methods. In order to implement the above embodiments, an embodiment of the present invention further provides a computer program product. The computer program product includes a computer program stored on a computer-readable storage medium. The computer program includes program instructions. The program instructions are implemented when the processor executes the program instructions. The method for determining an interactive target of a smart device according to the foregoing embodiment.

In order to achieve the above embodiments, an embodiment of the present invention further provides a non-transitory computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the interaction goal of the smart device as described in the above embodiment is achieved. Determine the method.

In the description of this specification, the description with reference to the terms “one embodiment”, “some embodiments”, “examples”, “specific examples”, or “some examples” and the like means specific features described in conjunction with the embodiments or examples , Structure, material, or characteristic is included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms are not necessarily directed to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. In addition, without any contradiction, those skilled in the art may combine and combine different embodiments or examples and features of the different embodiments or examples described in this specification.

In addition, the terms "first" and "second" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of the present invention, the meaning of "plurality" is at least two, for example, two, three, etc., unless it is specifically and specifically defined otherwise.

Any process or method description in a flowchart or otherwise described herein can be understood as representing a module, fragment, or portion of code that includes one or more executable instructions for implementing a custom logic function or step of a process And, the scope of the preferred embodiments of the present invention includes additional implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order according to the functions involved, which should It is understood by those skilled in the art to which the embodiments of the present invention pertain.

The logic and / or steps represented in the flowchart or otherwise described herein, for example, a sequenced list of executable instructions that can be considered to implement a logical function, can be embodied in any computer-readable medium, For use by, or in combination with, an instruction execution system, device, or device (such as a computer-based system, a system that includes a processor, or another system that can fetch and execute instructions from an instruction execution system, device, or device) Or equipment. For the purposes of this specification, a "computer-readable medium" may be any device that can contain, store, communicate, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. More specific examples (non-exhaustive list) of computer-readable media include the following: electrical connections (electronic devices) with one or more wiring, portable computer disk enclosures (magnetic devices), random access memory (RAM ), Read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable media on which the program can be printed, because, for example, by optically scanning the paper or other media, and then editing, interpreting, or other suitable Process to obtain the program electronically and then store it in computer memory.

It should be understood that each part of the present invention may be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, a plurality of steps or methods may be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it may be implemented using any one or a combination of the following techniques known in the art: Discrete logic circuits, dedicated integrated circuits with suitable combinational logic gate circuits, programmable gate arrays (PGA), field programmable gate arrays (FPGA), etc.

Those of ordinary skill in the art can understand that all or part of the steps carried by the methods in the above embodiments can be completed by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and the program is running. , Including one or a combination of the steps of the method embodiments.

In addition, each functional unit in each embodiment of the present invention may be integrated into one processing module, or each unit may exist separately physically, or two or more units may be integrated into one module. The above integrated modules can be implemented in the form of hardware or software functional modules. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

The above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk. Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limitations on the present invention. Those skilled in the art can interpret the above within the scope of the present invention. Embodiments are subject to change, modification, substitution, and modification.

610‧‧‧shell

620‧‧‧Processor

630‧‧‧Memory

640‧‧‧circuit board

650‧‧‧Power circuit

The above and / or additional aspects and advantages of the present invention will become apparent and easy to understand from the following description of the embodiments with reference to the accompanying drawings, wherein: FIG. 1 is an interaction target of a smart device provided by an embodiment of the present invention A schematic flowchart of a determination method; FIG. 2 is a schematic flowchart of another method for determining an interactive target of a smart device according to an embodiment of the present invention; FIG. 3 is a schematic diagram of a binocular vision calculation distance provided by an embodiment of the present invention; FIG. 4 is a schematic flowchart of another method for determining an interactive target of a smart device according to an embodiment of the present invention; FIG. 5 is a schematic flowchart of another method for determining an interactive target of a smart device according to an embodiment of the present invention; FIG. 7 is a schematic flowchart of another method for determining an interactive target of a smart device according to an embodiment of the present invention; FIG. 7 is a schematic diagram of a structure of an interactive target determining device for a smart device according to an embodiment of the present invention; A schematic structural diagram of an embodiment of a smart device.

Claims (12)

A method for determining an interactive target of a smart device, comprising the following steps: acquiring an environment image within the monitoring range of a smart device, and performing target recognition on the environment image; The identified target is used as a candidate target to obtain the status information of the candidate target. For each candidate target, according to the corresponding status information, it is determined whether there is an interaction intention to interact with the smart device. From a candidate target that has an interaction intention, Select the interactive target of the smart device. The method according to item 1 of the scope of patent application, wherein the obtaining the status information of the candidate target comprises: obtaining a distance between the candidate target and the smart device; for each candidate target, according to the corresponding status information , Determining whether there is an interaction intention for interacting with the smart device, including: for each candidate target, determining whether a distance between the candidate target and the smart device is less than or equal to a preset threshold for the distance, and Whether the dwell time within the distance threshold exceeds a preset time threshold; if the distance between the candidate target and the smart device is less than or equal to the distance threshold and the dwell time exceeds the time threshold, It is determined that the candidate target has an interaction intention to interact with the smart device. The method according to item 1 of the scope of patent application, wherein the obtaining the status information of the candidate target comprises: obtaining a distance between the candidate target and the smart device, and a face angle of the candidate target; For each candidate target, judging whether there is an interaction intention to interact with the smart device according to the corresponding state information, including: for each candidate target, determining whether the distance between the candidate target and the smart device is less than or equal to a preset Whether the distance between the candidate target and the smart device is less than or equal to the preset distance threshold, and the person of the candidate target is within the preset angle range; If the face angle is within a preset angle range, it is determined that the candidate target has an interaction intention to interact with the smart device. The method according to item 1 of the scope of patent application, wherein the selecting an interactive target of the smart device from the candidate targets having an interaction intention includes: when a plurality of candidate targets are detected, and an interaction intention exists; When the number of candidate targets is plural, a candidate target closest to the smart device is determined from the plurality of candidate targets with interaction intent; and the smart is selected from the candidate targets closest to the smart device. An interactive target for the device. The method according to item 4 of the scope of patent application, wherein the selecting an interactive target of the smart device from the candidate targets closest to the smart device includes: when the closest distance to the smart device is When the number of candidate targets is plural, query whether there is a face image of the candidate target closest to the smart device in the registered face image database of the smart device; if one exists in the face image database, A face image of a candidate candidate closest to the smart device is used as an interactive target; if there is no distance to the smart device in the face image library For the face image of the closest candidate, randomly select a candidate that is closest to the smart device as an interaction target; if there are a plurality of people in the face image database with a candidate that is closest to the smart device For the face image, the candidate target closest to the smart device that is first queried is used as the interactive target. The method according to item 2 of the scope of patent application, wherein the obtaining the distance between the candidate target and the intelligent device includes: obtaining a depth map through a depth camera in the intelligent device, and according to the depth Map to obtain the distance between the target and the smart device; or, use a binocular vision camera in the smart device to shoot the candidate target and calculate a parallax of the image captured by the binocular vision camera , Calculate the distance between the candidate target and the smart device according to the parallax; or, use a laser radar in the smart device to emit laser light into the monitoring range; according to each obstacle in the monitoring range The returned laser generates a binary map of each obstacle; fuses each binary map with the environment image, and identifies the binary map corresponding to the candidate target from all the binary maps; according to the candidate The laser return time of the binary map corresponding to the target determines the distance between the candidate target and the smart device. The method according to item 3 of the scope of patent application, wherein the obtaining the face angle of the candidate target comprises: intercepting a face image of the candidate target from the environment image; and inputting the face image In a pre-trained machine learning model, the face angle of the face in the face image is obtained; the method further includes: training the machine learning model in the following manner: collecting the same face image, wherein the sample person The face image carries annotated data, which is used to represent the face angle of the sample face; the sample face image is input into the machine learning model initially constructed for training, and the machine learning after training When the error of the model is within the preset error range, the trained machine learning model is obtained. The method according to any one of claims 1 to 7, wherein after selecting the interaction target of the smart device from the candidate targets having an interaction intention, the method further includes: controlling the smart device and The interaction target interacts; during the interaction, identifies the center point of the face image of the interaction target; detects whether the center point of the face image is within a preset image area; if it is not in the image area Within, obtain a path between the center point of the face image and the center point of the image area; and according to the path, control the smart device so that the center point of the face image is within the image area. An interactive target determination device for a smart device, comprising: a first acquisition module for acquiring an environmental image within the monitoring range of the smart device, and performing target recognition on the environmental image; a second acquisition A module for obtaining target information from a target identified in the environmental image as a candidate target; a judgment module for each candidate target and judging whether or not there is based on the corresponding status information Interaction intention for interacting with the smart device; a selection module for selecting the interaction target of the smart device from a candidate target having an interaction intention. A smart device, comprising: a housing, a processor, a memory, a circuit board, and a power circuit, wherein the circuit board is disposed in a space surrounded by the housing, and the processor And the memory is arranged on the circuit board; the power circuit is used for supplying power to each circuit or device of the smart device; the memory is used for storing executable program code; wherein the processor reads the memory by The executable code stored in the body is used to run a program corresponding to the executable code, so as to implement the method for determining an interactive target of a smart device as described in any one of the first to eighth patent applications. A non-transitory computer-readable storage medium having stored thereon a computer program, characterized in that, when the program is executed by a processor, the smart device according to any one of claims 1 to 8 of the scope of patent application is implemented Method of interactive target determination. A computer program product. The computer program product includes a computer program stored on a computer-readable storage medium. The computer program includes a program instruction. When the program instruction is executed by a processor, the implementation of items 1 to 8 of the scope of patent application is achieved. The method of any one of clauses.