CN104519322B - A kind of machine vision Target Tracking System - Google Patents

Abstract

A kind of machine vision Target Tracking System proposed by the present invention, including master control system, multiple camera devices and one or more robots；Camera device includes camera and drive mechanism, and camera is used for shooting image, and drive mechanism is used to drive camera to rotate；Master control system includes information receiving unit, image mosaic unit, initial point and end point analysis unit, feature extraction unit, hazard analysis unit, warning device and wireless messages transmitter；Robot is worked accordingly provided with wireless messages receiver with receiving robot control instruction and performing.In the present invention, by feature extraction and image mosaic, the active path of each moving target is can extract out, is continuously monitored for each moving target, improve monitoring efficiency, reduction monitoring record consults the time, it is to avoid numerous and diverse problem of redundant image.

Description

A machine vision target tracking system

technical field

The invention relates to the technical field of monitoring, in particular to a machine vision target tracking system.

Background technique

With the development of society, people pay more and more attention to personal safety and privacy. Therefore, various camera monitoring equipment and regional monitoring systems emerge in endlessly. However, there is a problem in the existing monitoring equipment, that is, intelligent and passive video recording, but cannot identify the target in the video recording. Therefore, most of the existing monitoring equipment only play a deterrent effect. The videos of the monitoring equipment will only be retrieved and reviewed after the accident occurs. However, since the monitoring equipment does not have an automatic identification function, when viewing the videos, they can only be identified and extracted manually. The workload is heavy and mistakes are prone to occur.

Contents of the invention

Based on the technical problems existing in the background technology, the present invention proposes a machine vision target tracking system.

A machine vision target tracking system proposed by the present invention includes a main control system, multiple camera devices and one or more robots;

The camera device includes a camera and a driving mechanism, the camera is used to capture images, and the driving mechanism is used to drive the camera to rotate;

The main control system includes information receiving unit, image splicing unit, start point and end point analysis unit, feature extraction unit, risk analysis unit, alarm device and wireless information transmitter;

The information receiving unit is used to receive the image information sent by the camera and send it to the image stitching unit and the feature extraction unit; the feature extraction unit extracts features from a single moving target in the image and sends it to the image stitching unit, and the image stitching unit combines the received image pairs according to the features Analyze the position status of each active target to obtain the target activity path and send it to the start point and end point analysis unit; the start point and end point analysis unit analyzes its appearance position and disappearance position according to the target activity path and sends it to the risk analysis unit; Risk level analysis, control the alarm device to work according to the risk level and generate robot control instructions; the wireless information transmitter is used to send the robot control instructions in the form of wireless signals;

The robot is equipped with a wireless information receiver to receive robot control instructions and perform corresponding work.

Preferably, the danger analysis unit judges the degree of danger by judging whether the target's appearance position and disappearance position are normal. When the target's appearance position is abnormal, the danger analysis unit generates a robot control instruction to control the robot to intercept the target for identity verification; when the target's appearance position normal and the disappearance position is abnormal, the risk analysis unit generates robot control instructions to control the robot to collect the scene information of the disappearance position or block the scene, and control the alarm device to alarm; when the target position is abnormal and the disappearance position is abnormal, the risk analysis unit generates a robot The control instruction is used to control the robot to block the exit, and to control the alarm device to alarm.

Preferably, the alarm device is provided with a first-level alarm and a second-level alarm. When the target appears at a normal position and disappears at an abnormal position, the first-level alarm is used; when the target appears at an abnormal position and disappears at an abnormal position, a second-level alarm is used.

Preferably, the alarm device is an audible and visible alarm.

In the present invention, through feature extraction and image splicing, the activity path of each moving object can be extracted, and each moving object can be continuously monitored, so as to improve monitoring efficiency, reduce monitoring record review time, and avoid complicated problems of redundant images.

In the present invention, the location of appearance and disappearance of the target is separately extracted according to the activity path to analyze the risk degree, so that the monitoring strategy of avoiding the less important is realized, which is beneficial to improving the monitoring efficiency and ensuring the safety of the monitoring area.

In the present invention, by controlling the robot to check the identity of the target and collect or block the on-site information, the safety of the staff is ensured and the casualty rate of dangerous accidents is reduced. Moreover, the robot is driven by wireless communication, which is conducive to the real-time nature of instruction sending and execution, and improves the efficiency of accident handling.

Description of drawings

FIG. 1 is a schematic structural diagram of a machine vision target tracking system proposed by the present invention.

detailed description

Referring to FIG. 1 , a machine vision target tracking system proposed by the present invention includes a main control system 1 , multiple camera devices 2 and one or more robots.

The camera device 2 includes a camera 21 and a drive mechanism 22 , the camera 21 is used to capture images, and the drive mechanism 22 is used to drive the camera 21 to rotate. In this embodiment, the number and distribution of the camera devices 2 can be determined according to the size and layout of the monitoring area.

The main control system 1 includes an information receiving unit 11 , an image splicing unit 12 , a start point and end point analysis unit 13 , a feature extraction unit 14 , a risk analysis unit 15 , an alarm device 16 and a wireless information transmitter 17 .

The information receiving unit 11 is respectively connected to the camera 21 , the image stitching unit 12 and the feature extraction unit 14 , and is used to receive the image information sent by the camera 21 and send it to the image stitching unit 12 and the feature extraction unit 14 .

The feature extraction unit 14 is connected to the image stitching unit 12 and is used to extract features from a single moving object in the image and send it to the image stitching unit 12 . The image stitching unit 12 is connected to the start point and end point analysis unit 13, and analyzes the position and state of each moving target according to the characteristics combined with the received images taken by the camera 21, obtains the target activity path and sends it to the start point and end point analysis unit 13.

In this embodiment, through feature extraction and image splicing, the activity path of each moving object can be extracted, and each moving object can be continuously monitored to improve monitoring efficiency, reduce monitoring record review time, and avoid complicated problems of redundant images.

The start point and end point analysis unit 13 is connected to the risk analysis unit 15 , and analyzes its appearance position and disappearance position according to the target activity path and sends it to the risk analysis unit 15 .

The danger analysis unit 15 is connected to the alarm device 16, and analyzes the danger level of the target according to the position of appearance and disappearance, controls the work of the alarm device 16 according to the degree of danger, and generates robot control instructions.

In this embodiment, the location of appearance and disappearance of the target is extracted separately according to the activity path to analyze the risk level, which realizes the monitoring strategy of avoiding the less important and helping to improve the monitoring efficiency and ensure the safety of the monitoring area.

The wireless information transmitter 17 is connected to the risk analysis unit 15 and is used to send robot control instructions in the form of wireless signals to control the robot to work.

The robot is equipped with a wireless information receiver 3 to receive robot control instructions and perform corresponding work. The number and distribution of robots can be determined according to the size of the monitoring area.

In this embodiment, the specific working method of the risk analysis unit is as follows: the risk analysis unit 15 judges the degree of danger by judging whether the target appearance position and the disappearance position are normal. When the target position is abnormal, the risk analysis unit 15 generates robot control instructions to Control the robot to intercept the target for identity verification; when the target appears at a normal position and disappears at an abnormal position, the risk analysis unit 15 generates a robot control command to control the robot to collect on-site information at the disappearing position or block the scene, and controls the alarm device 16 to alarm; when the target appears If the position is abnormal and the disappearance position is abnormal, the risk analysis unit 15 generates a robot control instruction to control the robot to block the exit, and controls the alarm device 16 to alarm.

In this embodiment, by controlling the robot to check the identity of the target and collect or block the on-site information, the safety of the staff is ensured, which is conducive to reducing the casualty rate of dangerous accidents. Moreover, the robot is driven by wireless communication, which is conducive to the real-time nature of instruction sending and execution, and improves the efficiency of accident handling.

In this embodiment, the alarm device 16 is provided with a first-level alarm and a second-level alarm. When the target appears at a normal position and disappears at an abnormal position, the first-level alarm is used; when the target appears at an abnormal position and disappears at an abnormal position, a second-level alarm is used. The setting of the first-level alarm and the second-level alarm and the second-level alarm is conducive to the staff to take corresponding measures in time,

In this embodiment, the alarm device 16 is an audible and visible alarm, so as to expand the warning range.

The above is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto, any person familiar with the technical field within the technical scope disclosed in the present invention, according to the technical solution of the present invention Any equivalent replacement or change of the inventive concepts thereof shall fall within the protection scope of the present invention.

Claims (3)

1. a kind of machine vision Target Tracking System, it is characterised in that including master control system (1), multiple camera devices (2) and One or more robots；

Camera device (2) includes camera (21) and drive mechanism (22), and camera (21) is used for shooting image, drive mechanism (22) it is used to drive camera (21) to rotate；

Master control system (1) includes information receiving unit (11), image mosaic unit (12), initial point and end point analysis unit (13), feature extraction unit (14), hazard analysis unit (15), warning device (16) and wireless messages transmitter (17)；

Information receiving unit (11) is used to receive the image information of camera (21) transmission and is sent to image mosaic unit (12) With feature extraction unit (14)；Feature extraction unit (14) is to the single-unit activity Objective extraction feature in image and is sent to image Concatenation unit (12), image mosaic unit (12) is combined according to feature receives image to each moving target progress location status Analysis obtains goal activities path and is sent to initial point and end point analysis unit (13)；Initial point and end point analysis unit (13) basis Goal activities path analysis its there is position and disappearance position and be sent to hazard analysis unit (15)；Hazard analysis unit (15) Degree of danger analysis is carried out to target, warning device (16) work is controlled according to degree of danger and robot control instruction is generated； Wireless messages transmitter (17) is used to send robot control instruction in the way of wireless signal；

Robot is worked accordingly provided with wireless messages receiver (3) with receiving robot control instruction and performing；

Hazard analysis unit (15) occurs whether position and disappearance position normally judge its degree of danger by judging target, when When target occurs out of alignment normal, hazard analysis unit (15) generation robot control instruction is entered with control machine people's interception target Row identification checkup；When target occur position it is normal and disappear it is out of alignment often, the control of hazard analysis unit (15) generation robot Instruct and the field data of disappearance position is gathered with control machine people or sealed off the area, and control warning device (16) to alarm；When Target appearance is out of alignment normal and disappearance is out of alignment often, and hazard analysis unit (15) generation robot control instruction is with control machine Device people block outlet, and control warning device (16) to alarm.

2. machine vision Target Tracking System as claimed in claim 1, it is characterised in that warning device (16) is provided with one-level Alarm and second-level alarm, when target occur position it is normal and disappear it is out of alignment often, using primary alarm；When position occurs in target Abnormal and disappearance is out of alignment often, using second-level alarm.

3. machine vision Target Tracking System as claimed in claim 1 or 2, it is characterised in that warning device (16) is aobvious for sound Alarm.