CN-122024184-A - Object identification and early warning method and system for public transportation places

CN122024184ACN 122024184 ACN122024184 ACN 122024184ACN-122024184-A

Abstract

The invention discloses an object identification and early warning method and system for a public transportation place, which are characterized in that an original image of an object at a preset position of the public transportation place is obtained, the state characteristics of the object are identified, the characteristic label of the object is generated, the characteristic label is synchronously uploaded to a user end and a platform end, the historical motion record of the user end is determined according to an object searching request initiated by the user end, the motion image of the user end at the public transportation place is extracted from the platform end, an abnormal event of the object is determined by combining the characteristic label, the possible transfer path information of the object is identified according to the local environment image during the abnormal event, and an early warning notification message is sent to the relevant edge end under the platform end. By identifying the characteristic tag of the object in the original state, the object is traced and identified by combining the historical motion condition, the possible transfer space condition of the object is positioned, and the reliable, timely and accurate object identification is realized by using a distributed tracing and searching mode.

Inventors

YU DAN
LAN YUQING
Xing Zhihuan
WANG DANXING

Assignees

慧之安信息技术股份有限公司

Dates

Publication Date: 20260512
Application Date: 20251228

Claims (8)

1. The object identification and early warning method for the public transportation places is characterized by comprising the following steps: Acquiring an original image of an object at a preset position of a public transportation place, and identifying the state characteristics of the object from the original image to generate a characteristic label of the object; Determining a historical motion record of the user terminal according to an object searching request initiated by the user terminal, so as to extract a motion image of the user terminal in the public transportation place from the platform terminal; And according to the possible transfer path information, sending an early warning notification message to the relevant edge end subordinate to the platform end.
2. The method for identifying and pre-warning objects in public transportation according to claim 1, wherein: The method comprises the steps of obtaining an original image of an object at a preset position of a public transportation place, identifying the state characteristics of the object from the original image, and generating a characteristic label of the object, and synchronously uploading the characteristic label to a user side and a platform side, wherein the method comprises the following steps: Acquiring an original three-dimensional image of an object at a preset position of a public transportation place, and carrying out multi-modal identification on the original three-dimensional image to obtain multi-modal characteristics of the object, wherein the multi-modal characteristics comprise three-dimensional contour characteristics, color distribution characteristics and motion gesture characteristics of a global range; performing semantic vector conversion on the multi-modal features to generate feature labels of the objects, wherein the feature labels comprise vector elements corresponding to the three-dimensional contour features, the color distribution features and the motion gesture features; And after the feature tag is uploaded to a platform end, copying and transmitting the feature tag to the user end according to login information of the user end on the platform end.
3. The method for identifying and pre-warning objects in public transportation according to claim 1, wherein: Determining a historical motion record of the user terminal according to an object searching request initiated by the user terminal, so as to extract a motion image of the user terminal in the public transportation place from the platform terminal, and determining an abnormal event of the object according to the feature tag and the motion image, wherein the method comprises the following steps: Extracting equipment identity information of the user side from an object searching request initiated by the user side, and determining a historical motion record of the user side before the object searching request is initiated according to the equipment identity information, wherein the historical motion record comprises historical motion positioning data; According to the historical motion record, determining a historical motion trail coverage space of the user side, and extracting a motion image corresponding to the historical motion trail coverage space in a public transportation place from the platform side; Comparing the characteristic tag with the moving image, calibrating an image frame sequence in which the object exists in the moving image, and identifying an abnormal event occurring in the object from the image frame sequence, wherein the abnormal event refers to the time when the distance between the object and a user exceeds a preset distance threshold value for exceeding a preset time length.
4. The method for identifying and pre-warning objects in public transportation according to claim 1, wherein: and according to the possible transfer path information, sending an early warning notification message to the relevant edge end subordinate to the platform end, wherein the early warning notification message comprises the following components: identifying operation action information of other personnel except a user on the object from the local environment image so as to determine possible transfer path information of the object; and comparing the possible transfer path information with the spatial distribution information of all edge terminals subordinate to the platform terminal, and determining related edge terminals so as to send early warning notification messages to the related edge terminals, wherein the early warning notification messages comprise the feature labels.
5. An object recognition and early warning system for public transportation, comprising: the tag generation module is used for acquiring an original image of an object at a preset position of a public transportation place, and identifying the state characteristics of the object from the original image so as to generate a characteristic tag of the object; The tag uploading module is used for synchronously uploading the feature tag to the user side and the platform side; the image extraction module is used for determining a historical motion record of the user terminal according to an object searching request initiated by the user terminal, so that a motion image of the user terminal in the public transportation place is extracted from the platform terminal; the abnormal event determining module is used for determining an abnormal event of the object according to the characteristic tag and the moving image; the path information determining module is used for identifying possible transfer path information of the object according to the local environment image during the occurrence of the abnormal event; And the early warning notification module is used for sending an early warning notification message to the relevant edge end of the platform end subordinate according to the possible transfer path information.
6. The mass transit location object identification and early warning system of claim 5, wherein: the label generating module is used for acquiring an original image of an object at a preset position of a public transportation place, identifying the state characteristics of the object from the original image, so as to generate a characteristic label of the object, and comprises the following steps: Acquiring an original three-dimensional image of an object at a preset position of a public transportation place, and carrying out multi-modal identification on the original three-dimensional image to obtain multi-modal characteristics of the object, wherein the multi-modal characteristics comprise three-dimensional contour characteristics, color distribution characteristics and motion gesture characteristics of a global range; And carrying out semantic vector conversion on the multi-modal feature to generate a feature tag of the object, wherein the feature tag comprises vector elements corresponding to the three-dimensional contour feature, the color distribution feature and the motion gesture feature. The tag uploading module is used for synchronously uploading the feature tag to a user end and a platform end, and comprises the following steps: And after the feature tag is uploaded to a platform end, copying and transmitting the feature tag to the user end according to login information of the user end on the platform end.
7. The mass transit location object identification and early warning system of claim 5, wherein: The image extraction module is used for determining a historical motion record of the user terminal according to an object search request initiated by the user terminal, so as to extract the motion image of the user terminal in the public transportation place from the platform terminal, and comprises the following steps: Extracting equipment identity information of the user side from an object searching request initiated by the user side, and determining a historical motion record of the user side before the object searching request is initiated according to the equipment identity information, wherein the historical motion record comprises historical motion positioning data; According to the historical motion record, determining a historical motion trail coverage space of the user side, and extracting a motion image corresponding to the historical motion trail coverage space in a public transportation place from the platform side; the abnormal event determining module is configured to determine, according to the feature tag and the moving image, an abnormal event that occurs in the object, including: Comparing the characteristic tag with the moving image, calibrating an image frame sequence in which the object exists in the moving image, and identifying an abnormal event occurring in the object from the image frame sequence, wherein the abnormal event refers to the time when the distance between the object and a user exceeds a preset distance threshold value for exceeding a preset time length.
8. The mass transit location object identification and early warning system of claim 5, wherein: the path information determining module is configured to identify possible transfer path information of the object according to a local environment image during the occurrence of the abnormal event, and includes: identifying operation action information of other personnel except a user on the object from the local environment image so as to determine possible transfer path information of the object; the early warning notification module is configured to send an early warning notification message to the platform end subordinate related edge end according to the possible transfer path information, and includes: and comparing the possible transfer path information with the spatial distribution information of all edge terminals subordinate to the platform terminal, and determining related edge terminals so as to send early warning notification messages to the related edge terminals, wherein the early warning notification messages comprise the feature labels.

Description

Object identification and early warning method and system for public transportation places Technical Field The invention relates to the technical field of object identification, in particular to an object identification and early warning method and system for public transportation places. Background The passenger flow of public transportation places such as train stations and airports is huge, and passengers usually carry luggage items with different sizes, and the passengers exist in complex and changeable environments during entering and waiting, so that the luggage items are easy to lose or steal. When the situation occurs, the passenger can only request help from the station platform, and the passenger can only describe the outline characteristics of the passenger when describing the luggage items of the passenger, and can not provide more detailed and accurate outline information, so that the luggage items can not be accurately traced and identified in the public transportation places with huge space, and the searching and tracking efficiency and timeliness of the items are reduced. In addition, the monitoring images generated by the monitoring equipment are utilized to identify and track during the process of searching the luggage articles, but the mode seriously relies on the identification and searching by naked eyes, the efficiency is low, the situation of misidentification easily occurs, the quick and distributed tracking and searching of the objects cannot be ensured, and the dynamic identification reliability of the object space is reduced. Disclosure of Invention Considering that the existing public transportation places search the final luggage articles by means of manual identification, screening and monitoring images, the luggage articles cannot be positioned and tracked quickly and accurately, the situation of misidentification is easy to occur, and the reliability, timeliness and accuracy of dynamic identification of the object space are reduced. In view of the above problems, the present invention has been made to provide an object recognition and early warning method for public transportation sites that overcomes or at least partially solves the above problems, including: Acquiring an original image of an object at a preset position of a public transportation place, and identifying the state characteristics of the object from the original image to generate a characteristic label of the object; Determining a historical motion record of the user terminal according to an object searching request initiated by the user terminal, so as to extract a motion image of the user terminal in the public transportation place from the platform terminal; And according to the possible transfer path information, sending an early warning notification message to the relevant edge end subordinate to the platform end. Optionally, acquiring an original image of an object at a preset position of a public transportation place, identifying a state characteristic of the object from the original image, and generating a characteristic tag of the object, wherein the step of synchronously uploading the characteristic tag to a user side and a platform side comprises the following steps: Acquiring an original three-dimensional image of an object at a preset position of a public transportation place, and carrying out multi-modal identification on the original three-dimensional image to obtain multi-modal characteristics of the object, wherein the multi-modal characteristics comprise three-dimensional contour characteristics, color distribution characteristics and motion gesture characteristics of a global range; performing semantic vector conversion on the multi-modal features to generate feature labels of the objects, wherein the feature labels comprise vector elements corresponding to the three-dimensional contour features, the color distribution features and the motion gesture features; And after the feature tag is uploaded to a platform end, copying and transmitting the feature tag to the user end according to login information of the user end on the platform end. Optionally, according to the object searching request initiated by the user terminal, determining a historical motion record of the user terminal, so as to extract a motion image of the user terminal in the public transportation place from the platform terminal, and according to the feature tag and the motion image, determining an abnormal event of the object, including: Extracting equipment identity information of the user side from an object searching request initiated by the user side, and determining a historical motion record of the user side before the object searching request is initiated according to the equipment identity information, wherein the historical motion record comprises historical motion positioning data; According to the historical motion record, determining a historical motion trail coverage space of the user side, and extracting a moti