Robot search method and system

Abstract

The present invention discloses a robot search method and system, which includes the following steps: inputting the information of the person being searched and the user information into the robot and synchronizing them to the background management system; the background management system performs historical record retrieval based on the information of the person being searched; the robot autonomously navigates along a set patrol route and automatically plans a path for search based on the historical record location and / or the lost location input by the user; performs person search and matching, and the person search includes human body detection, face recognition, height detection, and color recognition; after a match is consistent, notification and confirmation are performed through the background management system. The present invention can identify and search by automatically planning a path based on the information of the person being searched, is not affected by the environment, has a wide coverage range, has high efficiency and accuracy in searching, and supports target tracking and dynamic location sharing, guiding the searcher to reunite with the person being searched, and preventing the found person from getting lost again.

Description

Technical Field

[0001] The present invention relates to the field of robot manufacturing technology, and more particularly, to a robot human search method and system. Background Art

[0002] Currently, existing robot-based search methods typically rely on fixed surveillance cameras to capture facial images of people, match them against reference facial images, and search for the person being sought. By placing autonomous robots at different locations in the environment, the user inputs an image of the target person into one of the robots. The robot then shares the image with the other robots. All robots then use their on-site cameras to capture facial images of people passing by in the environment and match them against the reference facial images to identify the person. Once a matching person is found, the robot next to the user autonomously navigates to the location of the robot that identified the person being sought, confirming the person's identity.

[0003] However, existing robot-based search methods rely on fixed-point facial image collection, which has a limited coverage area. Furthermore, the robot only collects facial information, and recognition accuracy is easily affected by factors such as camera angle and obstructions. Summary of the Invention

[0004] One purpose of the present invention is to provide a new technical solution for a robot search method, which can at least solve the problems in the existing technology of the robot search method not being affected by the environment and having low recognition accuracy.

[0005] A first aspect of the present invention provides a method for searching for people by a robot, comprising the following steps:

[0006] Input the information of the person being sought and the user information into the robot, and synchronize the information of the person being sought and the user information to the backend management system; wherein the information of the person being sought includes name, gender, height, picture and clothing color category;

[0007] Based on the information of the person being sought and / or the user information, the robot autonomously navigates along a set patrol route and performs historical record retrieval;

[0008] During the walking process, the robot automatically plans a path to search based on the historical location and / or the lost location input by the user;

[0009] The robot collects personnel information during the search process and performs personnel retrieval and matching based on the personnel information. The personnel retrieval includes body detection, face recognition, height detection and color recognition.

[0010] After the personnel information matches the person being sought, notification and confirmation are made through the background management system to complete the robot's person-finding task.

[0011] Optionally, based on the information of the person to be sought and / or the user information, the robot autonomously navigates along a set patrol route, and the step of performing historical record retrieval includes:

[0012] The system uses lidar to scan the environment to build a map, and autonomously navigates along the set patrol route to collect facial image information in the environment.

[0013] Reporting the collected facial image information to the backend management system;

[0014] The time range of the preset historical records is based on the time when the information input by the user is received, and the face graphic information in the environment is retrieved through the background management system;

[0015] The filtered facial image information is matched with the input reference facial image.

[0016] Optionally, the step of automatically planning a path for searching based on the historical record location and / or the lost location input by the user during the walking process of the robot includes:

[0017] Mark the robot's traversable sections in the map scanned by the robot in advance and set the section numbers;

[0018] With the historical record location or the lost location input by the user as the center, circles of different radii are drawn according to the preset search radius step values;

[0019] Sort and combine all the circle segments to generate a complete segment search sequence;

[0020] Autonomously plan the path search, and during the robot walking search process, determine whether the face image received by the background management system matches the reference face image, and continue searching.

[0021] Optionally, the steps of autonomously planning and finding a path include:

[0022] Follow the sequence of road numbers on the map, from small to large or from large to small;

[0023] When the difference between the current road section and the minimum road section number is less than the absolute value of the maximum road section number difference, the robot will walk and search in order from small to large.

[0024] Optionally, the robot collects personnel information during the search process and performs personnel retrieval and matching on the personnel information, including:

[0025] While the robot is walking, it collects infrared images of the environment;

[0026] performing image binarization on the collected infrared image and performing feature extraction;

[0027] The extracted sample features are input into the classifier and matched with the existing training sample model to obtain the recognition results.

[0028] Optionally, the step of performing personnel search and matching further includes:

[0029] When a person is detected by the infrared sensor, the RGB image of the environment is collected to obtain the bounding box coordinates of each face;

[0030] Perform face alignment to obtain a standard face with uniform angle and posture;

[0031] Perform feature modeling on the standardized facial image to obtain vectorized facial features;

[0032] The facial features are discriminated against the training template of the classifier to obtain a recognition result.

[0033] Optionally, the step of performing personnel search and matching further includes:

[0034] The image coordinate system of the infrared sensor is pre-associated with the coordinate system of the laser radar to obtain the obstacle distance data within the coordinate area provided by the infrared image;

[0035] Based on the obstacle distance data, calculate the human body height measured by the laser line, and take the maximum value as the height measurement value;

[0036] Compare the height value with the height entered by the user to determine potential search targets.

[0037] Optionally, the step of performing personnel search and matching further includes:

[0038] Synchronize the coordinate area position information of the person on the infrared image to the body camera, and perform color recognition on the coordinate area position provided by the infrared image;

[0039] The recognized color category matches the color category input by the user and is determined to be a potential search object.

[0040] Optionally, after the personnel information matches the person to be found, the steps of notifying and confirming through the background management system to complete the robot's person-finding task include:

[0041] After successful face recognition, the user is alerted by voice broadcast based on the input information of the person being sought.

[0042] Number the objects based on their height and color, and keep performing face recognition on the target objects;

[0043] The user confirms the finding on the robot and sends a text message to the user's mobile terminal to complete the search task.

[0044] A second aspect of the present invention provides a robot person-finding system, which is applied to the robot person-finding method described in the above embodiment. The system includes:

[0045] Central control unit, which is used for logic control and data operation of the system;

[0046] A detection module, connected to the central control unit, configured to detect facial image information, perform face recognition, height detection, and color recognition;

[0047] An input module, connected to the central control unit, for inputting information of the person being sought and user information, and synchronizing the information of the person being sought and the user information to a background management system;

[0048] A navigation module is connected to the central control unit and is used to autonomously search and plan a walking path and control the robot to walk autonomously;

[0049] A historical record retrieval module, the historical record retrieval module is connected to the central control unit, and the historical record retrieval module is used for historical record retrieval;

[0050] A personnel search and matching module, connected to the central control unit, for performing personnel search and matching;

[0051] A notification module is connected to the central control unit and is used to notify and confirm through the background management system after the personnel search is consistent.

[0052] The robot-based search method of the present invention can identify and locate a person based on their name, height, image, and clothing color. It is unaffected by the surrounding environment, boasts high recognition accuracy, and is highly efficient. Furthermore, through automatic path planning, it enables mobile search with wide coverage, supports target tracking, and dynamically shares locations, guiding the searcher to reunite with the person, preventing the person from getting lost again.

[0053] Further features and advantages of the present invention will become apparent from the following detailed description of exemplary embodiments of the present invention with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.

[0055] Figure 1 is a flow chart of a robot search method according to an embodiment of the present invention;

[0056] Figure 2 2 is a schematic diagram of a robot person-searching system according to an embodiment of the present invention.

[0057] Reference numerals:

[0058] Detection module 10;

[0059] Input module 20;

[0060] Navigation module 30;

[0061] History record retrieval module 40;

[0062] Personnel search and matching module 50;

[0063] Notification module 60;

[0064] Central control unit 70. DETAILED DESCRIPTION

[0065] Various exemplary embodiments of the present invention will now be described in detail with reference to the accompanying drawings. It should be noted that unless otherwise specifically stated, the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present invention.

[0066] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the invention, its application, or uses.

[0067] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered part of the specification.

[0068] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0069] It should be noted that like reference numerals and letters refer to like items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0070] The following describes in detail the robot person-searching method according to an embodiment of the present invention with reference to the accompanying drawings.

[0071] like Figure 1 As shown, the robot search method according to an embodiment of the present invention includes the following steps:

[0072] Input the information of the person being sought and the user information into the robot, and synchronize the information of the person being sought and the user information to the backend management system; the information of the person being sought includes name, gender, height, picture, and clothing color category;

[0073] Based on the information of the person being sought and / or the user, the robot autonomously navigates along the set patrol route and retrieves historical records;

[0074] During the search process, the robot automatically plans a path based on the historical location and / or the lost location input by the user;

[0075] The robot collects personnel information during the search process and performs personnel retrieval and matching based on the information. Personnel retrieval includes body detection, face recognition, height detection, and color recognition.

[0076] After the personnel information matches the person being sought, notification and confirmation will be made through the background management system to complete the robot's search task.

[0077] In other words, in the robot search method of the embodiment of the present invention, see Figure 1 First, the robot can input the information of the person being searched and the user, and then synchronize the information with the backend management system. The information of the person being searched includes name, height, picture, and clothing color category, which facilitates subsequent facial recognition and vital sign detection, improving the efficiency of the search and ensuring that the search process is not affected by the environment and avoiding missed detections.

[0078] Users can use their mobile devices to send the robot information about the person being sought (name, gender, height, picture, clothing color, etc.) and user information (phone number). This can be done by scanning a QR code on the robot or by taking a user-provided picture with the robot's camera. The robot can also be equipped with a touchscreen that displays a map of the environment, allowing users to input the location of the missing person. Once the information is entered, the user is notified to wait at the designated location on the map. After the user confirms the operation, the robot begins searching for the person and synchronizes the input information with the backend management system.

[0079] In the present invention, the robot has the ability to autonomously navigate and move, and supports human body recognition, human height collection, and facial image collection. The surveillance camera is arranged at a fixed position in the environment to support facial image collection.

[0080] In the present invention, human recognition can utilize infrared sensors to detect whether an object is a human. Human height can be acquired through LiDAR ranging and triangulation. Facial image acquisition can utilize cameras on the robot (collectively referred to as body cameras in the present invention) and surveillance cameras fixed to the environment (collectively referred to as surveillance cameras in the present invention). The robot's camera can capture user-provided images. Facial image information captured by the robot's camera includes the image, the robot's real-time location, and the acquisition time. Facial image information captured by the surveillance camera in the environment while in operation can include the image and the acquisition time.

[0081] The backend management system can receive and match image data collected by devices, realize data reception, processing and result distribution, and provide unified control and management of devices such as robots and surveillance cameras. Communication equipment can support 4G / 5G / WiFi / wired networks, etc.

[0082] Based on the person being sought and / or user information, the robot can then autonomously navigate along a pre-set patrol route and retrieve historical records. During its movement, the robot can automatically plan a search path based on historically recorded locations and / or user-entered lost locations. This automated path planning enables on-the-go searching, offering wide coverage, target tracking, and dynamic location sharing, guiding searchers to reunite with the person being sought, preventing the found person from becoming lost again.

[0083] Next, the robot collects personal information during the search process and performs a search and match based on it. This search includes body detection, facial recognition, height detection, and color recognition, improving comprehensiveness and preventing missed detections. Once the information matches the person being searched, notification and confirmation are sent through the backend management system, completing the robot's search.

[0084] Thus, the robot search method according to an embodiment of the present invention can identify and locate a person based on their name, height, picture, and clothing color, unaffected by the surrounding environment, with high recognition accuracy and efficient search. Furthermore, by automatically planning paths, it enables mobile search with wide coverage, supports target tracking, and dynamically shares locations, guiding the searcher to reunite with the person, preventing the found person from getting lost again.

[0085] According to one embodiment of the present invention, based on the information of the person being sought and / or the user information, the robot autonomously navigates along a set patrol route, and the steps of performing historical record retrieval include:

[0086] It uses lidar to scan the environment to build a map and autonomously navigate along the set patrol route to collect facial image information in the environment.

[0087] Report the collected facial image information to the backend management system;

[0088] The time range of the preset historical records is based on the moment when the user input is received, and the face graphic information in the environment is retrieved through the background management system;

[0089] Match the filtered facial image information with the input reference facial image.

[0090] Specifically, during the search process, the robot uses LiDAR to scan the environment to build a map and autonomously navigate along a pre-set patrol route, collecting facial image information (including images, the robot's real-time location, and acquisition time) and reporting it to the backend management system. While operating, surveillance cameras collect facial image information (including images and acquisition time) and report it to the backend management system. The system uses a preset time range for historical records, with the moment of user input being used as the reference. The backend management system retrieves facial image data collected by all robots and surveillance cameras in the environment and matches the selected facial image data with the input reference facial image.

[0091] If the match is successful, the robot uses the last image as a reference to pre-calibrate the camera's position and number in the robot's map. The robot's position / camera number corresponding to the image is then sent to the user-operated robot. The last image can be understood as the moment closest to the input information. If the match fails, the user's input location is used as a reference. If the match fails and the user does not input a location, the robot autonomously plans a path to find the person. While the robot autonomously searches for the person, other robots and surveillance cameras continue to collect facial image information from the environment and report it to the backend management system.

[0092] According to one embodiment of the present invention, the steps of automatically planning a path for searching based on historical location records and / or lost locations input by a user during the robot's walking process include:

[0093] Mark the robot's traversable sections in the map scanned by the robot in advance and set the section numbers;

[0094] With the historical record location or the lost location input by the user as the center, circles of different radii are drawn according to the preset search radius step values;

[0095] Sort and combine all the circle segments to generate a complete segment search sequence;

[0096] Autonomously plan the path search, and during the robot's walking search process, determine whether the face image received by the background management system matches the reference face image, and continue searching.

[0097] In other words, the robot can pre-mark traversable sections of the map it scans and assign them numbers. Based on the historically recorded location / user-input lost location, the robot searches for the missing location. Circles of varying radii are constructed with the historically recorded location / user-input lost location as the center, using preset search radius steps (e.g., 10m, 20m, 30m, 40m, etc., until the robot's map is completely contained within the circle). First, the smallest radius circle is traversed using map coordinates to identify the sections within it. These sections are then placed at the top of the search queue (if multiple sections exist within the circle, they are sorted in descending order by section number, with smaller sections being searched first, followed by larger sections). Next, the sections within each circle are determined, along with the search order, by ascending radius (excluding sections within circles of the previous radius). Finally, the sections within all circles are sorted and combined to create a complete search order (with sections within circles of smaller radius first).

[0098] According to one embodiment of the present invention, the step of autonomously planning a path search includes:

[0099] Follow the sequence of road numbers on the map, from small to large or from large to small;

[0100] When the absolute value of the difference between the current section of the robot and the minimum section number is less than the maximum section number difference, the robot will walk and search in order from small to large.

[0101] That is, in autonomous path-finding, the robot can walk and search in the order of the road segment numbers on the map, from small to large / from large to small. If the difference between the robot's current road segment and the smallest road segment number is less than the absolute value of the difference with the largest road segment number, the robot will follow the order from small to large; otherwise, the robot will follow the order from large to small. During the robot's walking search, if the backend management system receives a facial image captured by a surveillance camera and successfully matches it, the backend management system will send the camera number to the searching robot. The robot will then proceed to the camera's location based on the camera number and continue searching at a new reference position. During the robot's walking search, if the backend management system receives a facial image captured by another robot and successfully matches it, the backend management system will assign the search task to the robot that currently captures the matching face and notify the robot originally responsible for the search to cancel the task.

[0102] In some specific embodiments of the present invention, the steps of collecting personnel information during the search process and matching the personnel information with personnel include:

[0103] While the robot is walking, it collects infrared images of the environment;

[0104] Perform image binarization on the collected infrared image and perform feature extraction;

[0105] The extracted sample features are input into the classifier and matched with the existing training sample model to obtain the recognition results.

[0106] Optionally, the step of performing personnel search and matching further includes:

[0107] When a person is detected by the infrared sensor, the RGB image of the environment is collected to obtain the bounding box coordinates of each face;

[0108] Perform face alignment to obtain a standard face with uniform angle and posture;

[0109] Perform feature modeling on the standardized facial image to obtain vectorized facial features;

[0110] The facial features are discriminated against the training template of the classifier to obtain the recognition result.

[0111] Optionally, the step of performing personnel search and matching further includes:

[0112] The image coordinate system of the infrared sensor is pre-associated with the coordinate system of the laser radar to obtain the obstacle distance data within the coordinate area provided by the infrared image;

[0113] Based on the obstacle distance data, calculate the human height measured by the laser line and take the maximum value as the height measurement value;

[0114] Compare the height value with the height entered by the user to determine potential search targets.

[0115] Optionally, the step of performing personnel search and matching further includes:

[0116] Synchronize the coordinate area position information of the person on the infrared image to the camera, and perform color recognition on the coordinate area position provided by the infrared image;

[0117] The recognized color category matches the color category input by the user and is determined to be a potential search object.

[0118] In other words, in person retrieval and matching, human body detection can be performed first. As the robot walks, the infrared sensor captures infrared images of the environment. After image acquisition, the color image is binarized and converted into a grayscale image. The binarized image is then subjected to noise filtering, image segmentation, and feature extraction. Finally, the extracted sample features are input into a classifier and matched with the existing training sample model to obtain the recognition result.

[0119] If the recognition result is that there is someone, then return to face recognition. If the recognition result is that there is no one, then continue walking and collecting infrared images

[0120] Next, facial recognition can be performed. When the robot detects a person using its infrared sensor, it uses a camera to capture an RGB image of the environment. This image is processed by a face detector to obtain the bounding box coordinates of each face. Face alignment is then performed, using a similarity transformation based on key facial features (such as the corners of the eyes, the tip of the nose, and the corners of the mouth) to obtain a standardized face with uniform angles and postures. Next, feature modeling is performed on the standardized facial image to generate vectorized facial features. Finally, the facial features are compared with the classifier's training template to obtain the recognition result.

[0121] In addition, the image coordinate system of the infrared sensor and the image coordinate system of the camera can be pre-associated and mapped (based on the field of view and installation position settings of the infrared sensor and camera). After the infrared sensor detects a person, the coordinate area position information of the person on the infrared image is synchronized to the camera. After the camera collects the RGB image, it performs face recognition on the coordinate area position provided by the infrared image based on the coordinate mapping relationship.

[0122] If a face is recognized but does not match the face of the person being searched for, the robot continues walking and detecting; if a face is recognized and matches the face of the person being searched for, a notification and confirmation will be made; if a face is not recognized (for example, the person is facing away from the robot / obstructed), the robot will jump to detecting the person's physical characteristics (height, clothing color).

[0123] Next, height detection can be performed. Based on the infrared sensor and lidar's field of view and installation location, the infrared sensor's image coordinate system is pre-mapped to the lidar's coordinate system. Once the infrared sensor detects a person, the person's coordinate area position information on the infrared image is synchronized with the lidar. The lidar then emits multiple laser lines in the vertical (sky-to-ground) direction to scan for obstacles in the environment. Based on the coordinate mapping relationship, it obtains obstacle distance data within the coordinate area provided by the infrared image.

[0124] Based on the obstacle distance measured by the LiDAR, the angle between each laser line speed and the horizontal plane, and the installation height of the LiDAR, the human height measured by each laser line speed is calculated, and the maximum value is taken as the height measurement value. The height value is compared with the height input by the user. If the absolute value of the difference between the two is less than the set value, it is determined to be a potential search object and then enters color recognition. Otherwise, it is determined to be a non-search object, and the robot continues to walk and detect along the route. The slope area is marked on the robot map in advance. If the current coordinates of the robot are in the slope area, height detection is not performed and color recognition is performed directly. If the robot is not in the slope area, based on the distance and angle of the obstacle (pedestrian) detected by the LiDAR, according to the robot's coordinates and the obstacle distance and angle, it is determined whether the obstacle is on the slope. If so, height detection is not performed and color recognition is performed directly.

[0125] Finally, color recognition can be performed. After the infrared sensor detects a person and the lidar detects a potential target object by height, the coordinate area position information of the person on the infrared image is synchronized to the camera. After the camera collects the RGB image, the color recognition of the coordinate area position provided by the infrared image is performed based on the coordinate mapping relationship.

[0126] In the color recognition method, the image corresponding to the coordinate region is converted to the HSV color space. Using a color histogram, the number of pixels for each color value and its proportion of the total image pixels are determined (a pre-set relationship between each color category and color value). This is then combined with the color category input by the user to determine the number and proportion of pixels for the target color category. If the target color proportion is greater than the set value, the color category of the detected object is considered a match. Otherwise, it is considered a mismatch and the object is not being sought, and the robot continues its detection along the route.

[0127] If the recognized color category matches the color category input by the user, it is determined to be a potential search object, and the object is numbered based on the height and color of the recognized object. During the robot's walking process (walking along the planned road section), it continues to perform face recognition on the numbered object, and automatically adjusts the camera angle to keep the target object in the center of the field of view until the recognition result is obtained; if the recognized face does not match the face of the search object, the robot continues to walk and detect; if the recognized face matches the face of the search object.

[0128] According to one embodiment of the present invention, after the person information matches the person being sought, notification and confirmation are performed through the background management system. The steps of completing the robot's person-finding task include:

[0129] After successful face recognition, the system will use voice announcements based on the input information of the person being sought to alert the person.

[0130] Number the objects based on their height and color, and keep performing face recognition on the target objects;

[0131] The user confirms the finding on the robot and sends a text message to the user's mobile terminal to complete the search task.

[0132] That is, after successful facial recognition, the robot uses the entered name of the person to announce the person through voice broadcast. Simultaneously, the robot sends a text message via the GSM module to the user's (the person who initiated the search) mobile terminal, indicating that the person has been found. The user can access the web link in the text message to view the robot's location and proceed to the robot. After locating the person, the robot assigns a number based on the height and color of the identified object and continues to perform facial recognition on the target object. If the target object continues to move, the robot automatically adjusts the camera angle to keep the target object in the center of its field of view. It also adjusts its position based on the distance to the person obtained from the LiDAR scan to ensure that the distance between the target object and the robot is within the set range. At this point, the robot will no longer continue to travel along the planned route.

[0133] The user (the person who initiated the search task) confirms the search has been found on the robot. The robot then sends a text message verification code to the user's mobile device. The user enters the verification code on the robot, and the search task ends if the verification is successful. If the robot sends a "person found" text message to the user but does not confirm the search within the set time, the search task ends automatically.

[0134] If all road sections are searched but no matching person is found, the robot will send a text message to the user's (the party who initiated the search task) mobile terminal through the GSM module to inform the user that the person was not found and the robot ends the search task.

[0135] The robot search method of the present invention supports autonomous path planning based on basic search information, enabling mobile search. The robot supports body detection, facial recognition, and vital sign detection (height, clothing color). The camera angle can be adjusted for detection and identification. Screening-based detection reduces the number of sample tests, improving efficiency while avoiding missed detections due to factors such as angle and occlusion. This method supports the integration of a backend management system with fixed surveillance cameras to achieve combined search and robot scheduling. It also supports SMS notification of the search initiator, enabling result notification and robot location sharing.

[0136] The robot-based search method of this invention utilizes mobile search capabilities, providing wide coverage. It uses facial recognition and vital sign detection for screening, improving efficiency and preventing missed detections. It also utilizes a combination of robots and fixed surveillance cameras for enhanced search efficiency. The robot supports target tracking and dynamic location sharing, guiding the searcher to reunite with the person being sought, preventing the person from becoming separated again.

[0137] According to a second aspect of the present invention, a robot search system is provided, which is applied to the robot search method in the above embodiment, such as Figure 2As shown, the system includes a central control unit 70, a detection module 10, an input module 20, a navigation module 30, a history retrieval module 40, a personnel retrieval and matching module 50, and a notification module 60. Figure 2 The central control unit 70 is responsible for system logic control and data operations. The detection module 10 is connected to the central control unit 70 and is responsible for detecting facial image information, facial recognition, height detection, and color recognition. The input module 20 is connected to the central control unit 70 and is responsible for inputting the person being sought and user information, and synchronizing the person and user information with the backend management system. The navigation module 30 is connected to the central control unit 70 and is responsible for autonomously searching and planning a travel path and controlling the robot's autonomous movement. The history record retrieval module 40 is connected to the central control unit 70 and is responsible for historical record retrieval. The person search and matching module 50 is connected to the central control unit 70 and is responsible for performing person search and matching. Both the history record retrieval module 40 and the person search and matching module 50 can be part of the backend management system. The notification module 60 is connected to the central control unit 70 and is responsible for providing notification and confirmation via the backend management system when a person search match is found.

[0138] The robot search system of the present invention supports autonomous path planning based on basic search information, enabling mobile searches. The robot supports body detection, facial recognition, and vital sign detection (height, clothing color). The camera angle can be adjusted for detection and identification. Screening-based detection reduces the number of sample tests, improving efficiency while avoiding missed detections due to factors such as angle and occlusion. This method supports combining a backend management system with fixed surveillance cameras to implement combined searches and robot scheduling. It also supports SMS notification of the search initiator, enabling result notification and robot location sharing.

[0139] The robot search system of this invention utilizes mobile search technology with wide coverage. It uses facial recognition and vital sign detection for screening and detection, improving efficiency and preventing missed detections. It also utilizes a combination of robots and fixed surveillance cameras to further enhance search efficiency. The robot supports target tracking and dynamic location sharing, guiding searchers to reunite with the person being sought, preventing the person from becoming separated again.

[0140] Although some specific embodiments of the present invention have been described in detail by way of examples, it should be understood by those skilled in the art that the above examples are for illustration only and are not intended to limit the scope of the present invention. It should be understood by those skilled in the art that modifications may be made to the above embodiments without departing from the scope and spirit of the present invention. The scope of the present invention is defined by the appended claims.

Claims

1. A robot search method, characterized in that: The following steps are involved: Input the information of the person being sought and the user information into the robot, and synchronize the information of the person being sought and the user information to the backend management system; wherein the information of the person being sought includes name, gender, height, picture and clothing color category; Based on the information of the person being sought and / or the user information, the robot autonomously navigates along a set patrol route and retrieves historical records. The steps include: using a laser radar to scan the environment contour to construct a map, and autonomously navigating along the set patrol route to collect facial image information in the environment; reporting the collected facial image information to the backend management system; presetting the time range of the historical records, based on the moment when the user input is received, and retrieving the facial graphic information in the environment through the backend management system; and matching the filtered facial image information with the input reference facial image. During the robot's walking process, the robot automatically plans a path for searching based on the historically recorded location and / or the lost location input by the user. The steps include: pre-marking the robot's traversable sections in a map scanned by the robot and assigning section numbers; drawing circles of different radii based on preset search radius step values ​​with the historically recorded location or the lost location input by the user as the center; sorting and combining the sections of all circles to generate a complete section search sequence; autonomously planning the path for searching, and during the robot's walking search, determining whether the face image received by the backend management system matches the reference face image, and continuing the search; The robot collects personnel information during the search process and performs personnel retrieval and matching based on the personnel information. The personnel retrieval includes body detection, face recognition, height detection and color recognition. After the personnel information matches the person being sought, notification and confirmation are made through the background management system to complete the robot's person-finding task.

2. The robot search method according to claim 1, characterized in that: The steps of autonomous path finding include: Follow the sequence of road numbers on the map, from small to large or from large to small; When the difference between the current road section and the minimum road section number is less than the absolute value of the maximum road section number difference, the robot will walk and search in order from small to large.

3. The robot search method according to claim 1, characterized in that: The robot collects personnel information during the search process and performs personnel retrieval and matching based on the personnel information, including the following steps: While the robot is walking, it collects infrared images of the environment; performing image binarization on the collected infrared image and performing feature extraction; The extracted sample features are input into the classifier and matched with the existing training sample model to obtain the recognition results.

4. The robot search method according to claim 3, characterized in that: The steps for personnel search and matching also include: When a person is detected by the infrared sensor, the RGB image of the environment is collected to obtain the bounding box coordinates of each face; Perform face alignment to obtain a standard face with uniform angle and posture; Perform feature modeling on the standardized facial image to obtain vectorized facial features; The facial features are discriminated against the training template of the classifier to obtain a recognition result.

5. The robot search method according to claim 4, characterized in that: The steps for personnel search and matching also include: The image coordinate system of the infrared sensor is pre-associated with the coordinate system of the laser radar to obtain the obstacle distance data within the coordinate area provided by the infrared image; Based on the obstacle distance data, calculate the human body height measured by the laser line, and take the maximum value as the height measurement value; Compare the height value with the height entered by the user to determine potential search targets.

6. The robot search method according to claim 5, characterized in that: The steps for personnel search and matching also include: Synchronize the coordinate area position information of the person on the infrared image to the body camera, and perform color recognition on the coordinate area position provided by the infrared image; The recognized color category matches the color category input by the user and is determined to be a potential search object.

7. The robot search method according to claim 6, characterized in that: After the personnel information matches the person being sought, notification and confirmation are made through the background management system. The steps of completing the robot's person-finding task include: After successful face recognition, the user is alerted by voice broadcast based on the input information of the person being sought. Number the objects based on their height and color, and keep performing face recognition on the target objects; The user confirms the finding on the robot and sends a text message to the user's mobile terminal to complete the search task.

8. A robot search system, applied to the robot search method according to any one of claims 1 to 7, characterized in that: The system comprises: Central control unit, which is used for logic control and data operation of the system; A detection module, connected to the central control unit, configured to detect facial image information, perform face recognition, height detection, and color recognition; An input module, connected to the central control unit, for inputting information of the person being sought and user information, and synchronizing the information of the person being sought and the user information to a background management system; A navigation module is connected to the central control unit and is used to autonomously search and plan a walking path and control the robot to walk autonomously; A historical record retrieval module, the historical record retrieval module is connected to the central control unit, and the historical record retrieval module is used for historical record retrieval; A personnel search and matching module, connected to the central control unit, for performing personnel search and matching; A notification module is connected to the central control unit and is used to notify and confirm through the background management system after the personnel search is consistent.

Images