CN116935551A - Perimeter intrusion detection method, system, equipment and storage medium - Google Patents
Perimeter intrusion detection method, system, equipment and storage medium Download PDFInfo
- Publication number
- CN116935551A CN116935551A CN202311197663.2A CN202311197663A CN116935551A CN 116935551 A CN116935551 A CN 116935551A CN 202311197663 A CN202311197663 A CN 202311197663A CN 116935551 A CN116935551 A CN 116935551A
- Authority
- CN
- China
- Prior art keywords
- intrusion
- weather
- sensor
- target
- radar
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 70
- 230000000007 visual effect Effects 0.000 claims abstract description 39
- 238000000034 method Methods 0.000 claims abstract description 19
- 238000012544 monitoring process Methods 0.000 claims description 86
- 239000000428 dust Substances 0.000 claims description 34
- 230000004927 fusion Effects 0.000 claims description 22
- 230000009545 invasion Effects 0.000 claims description 21
- 238000012545 processing Methods 0.000 claims description 19
- 238000004590 computer program Methods 0.000 claims description 11
- 238000012549 training Methods 0.000 claims description 5
- 230000007613 environmental effect Effects 0.000 claims description 4
- 239000013307 optical fiber Substances 0.000 description 46
- 230000008901 benefit Effects 0.000 description 9
- 238000005516 engineering process Methods 0.000 description 8
- 241001465754 Metazoa Species 0.000 description 7
- 230000008569 process Effects 0.000 description 7
- 230000007547 defect Effects 0.000 description 5
- 230000000694 effects Effects 0.000 description 5
- 238000004458 analytical method Methods 0.000 description 4
- 238000009434 installation Methods 0.000 description 4
- 238000013459 approach Methods 0.000 description 3
- 238000013473 artificial intelligence Methods 0.000 description 3
- 230000008859 change Effects 0.000 description 3
- 241000282414 Homo sapiens Species 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 239000000835 fiber Substances 0.000 description 2
- 238000003384 imaging method Methods 0.000 description 2
- 230000002093 peripheral effect Effects 0.000 description 2
- 230000035945 sensitivity Effects 0.000 description 2
- 230000001360 synchronised effect Effects 0.000 description 2
- 238000012876 topography Methods 0.000 description 2
- 230000001960 triggered effect Effects 0.000 description 2
- 241000283690 Bos taurus Species 0.000 description 1
- 241000282832 Camelidae Species 0.000 description 1
- 241000283086 Equidae Species 0.000 description 1
- 206010063385 Intellectualisation Diseases 0.000 description 1
- 241001494479 Pecora Species 0.000 description 1
- 241000282458 Ursus sp. Species 0.000 description 1
- 230000007123 defense Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 239000004576 sand Substances 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000003068 static effect Effects 0.000 description 1
- 238000004148 unit process Methods 0.000 description 1
- 238000012800 visualization Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B13/00—Burglar, theft or intruder alarms
- G08B13/18—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength
- G08B13/181—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using active radiation detection systems
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B13/00—Burglar, theft or intruder alarms
- G08B13/16—Actuation by interference with mechanical vibrations in air or other fluid
- G08B13/1654—Actuation by interference with mechanical vibrations in air or other fluid using passive vibration detection systems
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B13/00—Burglar, theft or intruder alarms
- G08B13/18—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength
- G08B13/189—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using passive radiation detection systems
- G08B13/194—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using passive radiation detection systems using image scanning and comparing systems
- G08B13/196—Actuation by interference with heat, light, or radiation of shorter wavelength; Actuation by intruding sources of heat, light, or radiation of shorter wavelength using passive radiation detection systems using image scanning and comparing systems using television cameras
-
- G—PHYSICS
- G08—SIGNALLING
- G08B—SIGNALLING OR CALLING SYSTEMS; ORDER TELEGRAPHS; ALARM SYSTEMS
- G08B29/00—Checking or monitoring of signalling or alarm systems; Prevention or correction of operating errors, e.g. preventing unauthorised operation
- G08B29/18—Prevention or correction of operating errors
- G08B29/185—Signal analysis techniques for reducing or preventing false alarms or for enhancing the reliability of the system
- G08B29/188—Data fusion; cooperative systems, e.g. voting among different detectors
Abstract
The invention discloses a perimeter intrusion detection method, a system, equipment and a storage medium, wherein the method fuses data of a millimeter wave radar, a vision sensor, a weather sensor and a vibration detector to detect intrusion conditions along a protected perimeter; under non-severe weather, the detection reliability of the radar and the visual sensor is high, and the alarm accuracy is improved by adopting an alarm strategy of the radar and the visual sensor; meanwhile, the visual sensor is possibly shielded or missed detection caused by other reasons, and an alarm strategy of a radar and a vibration detector is adopted, so that manual intervention is not needed, and the intelligent degree of detection is improved.
Description
Technical Field
The invention belongs to the technical field of perimeter environment monitoring, and particularly relates to a perimeter intrusion detection method, a perimeter intrusion detection system, perimeter intrusion detection equipment and a perimeter intrusion detection storage medium based on multi-sensor fusion.
Background
With the continuous development of technology, electronic technology plays an increasingly important role in perimeter safety protection, and various perimeter safety protection technologies such as video monitoring technology, infrared correlation technology, electronic fence, vibration optical fiber, radar and the like are continuously emerging. At present, a monitoring method for linear perimeter intrusion mainly uses a camera and is assisted by equipment such as a pulse electronic fence, a radar, infrared/laser correlation, a vibration optical fiber and the like. The current perimeter early warning system has single anti-intrusion means, and because the linear perimeter protection equipment is applied outdoors, the detection and identification accuracy is greatly affected by the environment, so that the equipment is easy to generate false alarm, has lower reliability in practical application, and is difficult to have the effect of pre-prevention. The main reasons are as follows:
(1) The imaging stability of the camera is relatively poor, the camera is easily influenced by factors such as light, weather and the like, the camera is difficult to work normally in rainy, hazy, haze, snow and sand storm weather, and the topography fluctuation can bring dead zones to visual imaging, so that false alarm and missing report are high;
(2) The radar can cause false alarms due to factors such as tree shaking, small animal activities, multipath reflection and the like, the resolution ratio of the conventional millimeter wave radar is low, the radar is sensitive to metal, the characteristic information of a target cannot be identified, and the visualization is difficult;
(3) The vibration optical fiber has weak adaptability to weather and environment, is easily influenced by strong wind, heavy rain, hail, small animals and the like, generates more false alarm false alarms, is passively detected, and cannot early warn in advance.
Based on the analysis, the requirement of strong adaptability to complex environments, full coverage, all-weather detection and early warning all the day, zero missing report and near zero false report cannot be met by utilizing a single sensor for detection. Therefore, the multi-sensor fusion technology is necessary, the advantages of various sensors are fully combined, the defects of the sensors are made up, the advantages are complemented, and the protection requirement of the linear perimeter can be met.
The prior patent (application publication number CN 115691018A) adopts vibration optical fibers, radars and cameras as intrusion sensing equipment and adopts environment sensing equipment such as rain and snow sensors, so that false alarms caused by environmental influence when the vibration optical fibers are used independently are reduced, the radars and the cameras can perform intrusion alarm automatically and logically, and false alarms are reduced. However, the patent uses the vibration optical fiber and the environment sensing equipment to trigger together, and the triggering mode requires that the alarm can be triggered after the invasion target contacts the fence, so that early warning can not be performed in advance; meanwhile, when the radar and the vibrating optical fiber detect the target and the vision sensor does not detect the target, manual intervention is needed to confirm whether invasion occurs or not, and the intelligent degree is not high; each area uses a camera, so that near targets and far targets in the area cannot be considered at the same time, and the phenomenon of missing report is easy to generate.
Disclosure of Invention
Aiming at the problems that the traditional linear perimeter intrusion detection system is easily affected by natural environment, and has high false alarm rate, low intelligent level, inaccurate identification of intrusion targets and difficult early warning realization, the invention provides a perimeter intrusion detection method, a system, equipment and a storage medium, which are used for fusing data of equipment such as millimeter wave radar, a visual sensor, a vibration detector, a meteorological sensor and the like, fully combines the advantages of the sensors, and has the advantages of strong adaptability to complex environment, full coverage, all-weather detection early warning all-day, zero false alarm and near zero false alarm.
The invention solves the technical problems by the following technical scheme: the perimeter intrusion detection method based on multi-sensor fusion, wherein the sensors comprise a radar, a meteorological sensor, a visual sensor and a vibration detector, the radar and the visual sensor are arranged on one side or above the protected perimeter line, and the radar and the visual sensor are synchronously collected, and the detection method comprises the following steps:
step 1: whether an intrusion target exists or not is primarily judged according to the intrusion scanning information of the radar, if the intrusion target exists, whether the intrusion target is bad weather or not is judged according to the weather environment information of the weather sensor, or whether the intrusion target is bad weather or not is judged according to the weather environment information, the intrusion monitoring image of the visual sensor and the vibration information along the protected perimeter of the vibration detector, if the intrusion target exists, the step 2 is carried out; if the weather is not bad, the step 3 is carried out;
Step 2: deeply judging whether an intrusion target exists according to the intrusion scanning information, and if so, sending out a pure radar alarm signal; otherwise, not sending out an alarm signal;
step 3: judging whether an intrusion target exists or not according to the intrusion monitoring image, if so, matching the intrusion target detected by the radar with the intrusion target monitored by the vision sensor, and if the matching is successful, sending out a thunder alarm signal; if the matching is unsuccessful, the step 4 is carried out; if no intrusion target is judged, the step 4 is carried out;
step 4: judging whether an intrusion target exists or not according to the vibration information, if yes, matching the intrusion target detected by the radar with the intrusion target detected by the vibration detector, and if successful, sending out a Lei Zhen alarm signal; if the matching is unsuccessful, an alarm signal is not sent out; if no intrusion target is judged, no alarm signal is sent out.
Further, the vibration detector is arranged on the fence along the protected perimeter in a net hanging manner, or is laid close to the fence in a buried manner.
Further, the protected perimeter is divided into a plurality of monitoring areas along the line, each monitoring area is divided into a plurality of subareas, a set of weather sensors are configured in each monitoring area, a radar and two vision sensors are configured in each subarea, and the two vision sensors are respectively a first vision sensor for monitoring a near target and a second vision sensor for monitoring a far target.
Further, the meteorological sensor comprises a wind speed sensor, a rain and snow sensor, a rainfall sensor, a temperature and humidity sensor, a dust detector, a thick fog detector and a hail sensor; the bad weather is at least one of strong wind, heavy rain, strong snow, flying dust, dense fog and hail weather.
Further, in the step 1, determining whether the weather is severe weather according to weather environment information of the weather sensor specifically includes:
if the acquired value of the wind speed sensor is larger than the wind speed threshold value, the wind speed sensor is in a windy weather;
if the rain and snow sensor detects a rain and snow signal, the acquired value of the rain sensor is larger than the rain threshold value and the acquired value of the temperature and humidity sensor is larger than the temperature and humidity threshold value, the weather is heavy rain; if the rain and snow sensor detects a rain and snow signal, the acquired value of the rain sensor is smaller than the rain threshold value and the acquired value of the temperature and humidity sensor is smaller than the temperature and humidity threshold value, the weather is heavy snow;
if the collection value of the dust detector is larger than the dust threshold value, the dust detector is in dust weather; if the acquisition value of the dense fog detector is larger than the dense fog threshold value, the dense fog weather is the dense fog weather; if the collected value of the hail sensor is larger than the hail threshold value, the hail sensor is in hail weather.
Further, in the step 1, judging whether the weather is bad weather according to weather environment information, intrusion monitoring images and vibration information, specifically including:
obtaining vibration information under normal weather conditions and vibration information under severe weather conditions; taking vibration information under different weather as an input sample, and taking weather environment information corresponding to the vibration information as a sample label to construct a sample data set; constructing a weather prediction model, and training the weather prediction model by utilizing the sample data set to obtain a target weather prediction model;
the method comprises the steps of collecting weather environment information by using a weather sensor and collecting vibration information by using a vibration detector;
if the acquired value of the wind speed sensor is larger than the wind speed threshold value and the predicted result of predicting the vibration information by using the target weather prediction model is strong wind, the predicted result is strong wind weather;
if the rain and snow sensor detects that the rain and snow signal is detected, the acquired value of the rain sensor is larger than the rain threshold value, the acquired value of the temperature and humidity sensor is larger than the temperature and humidity threshold value, and the predicted result of predicting the vibration information by using the target weather prediction model is heavy rain, the predicted result is heavy rain weather; if the rain and snow sensor detects that the rain and snow signal is detected, the acquired value of the rain sensor is smaller than the rain threshold value, the acquired value of the temperature and humidity sensor is smaller than the temperature and humidity threshold value, and the predicted result of predicting the vibration information by using the target weather prediction model is big snow, the predicted result is big snow weather;
If the acquired value of the hail sensor is larger than the hail threshold value and the predicted result of predicting the vibration information by using the target weather prediction model is hail, the hail weather is obtained;
if the collection value of the dust detector is larger than the dust threshold value and the visibility of the intrusion monitoring image is lower than the visibility threshold value, the dust is in dust weather; and if the acquisition value of the dense fog detector is larger than the dense fog threshold value and the visibility of the intrusion monitoring image is lower than the visibility threshold value, the dense fog weather is obtained.
Further, in the step 2, determining whether there is an intrusion target according to the intrusion scan information in depth specifically includes:
determining the confidence coefficient of the intrusion target according to the intrusion scanning information;
judging whether the confidence coefficient is larger than a confidence coefficient threshold value, if so, judging that an intrusion target exists; otherwise, judging that the target is not invaded.
Further, in the step 3, after determining that there is an intrusion target according to the intrusion monitoring image, before matching the intrusion target detected by the radar with the intrusion target monitored by the vision sensor, the detection method further includes:
determining the type of the intrusion target and the position and the area of the intrusion target in the image according to the intrusion monitoring image;
Judging whether the area of the intrusion target in the image and the aspect ratio of the intrusion target in the image are in accordance with the corresponding threshold range, if so, matching the intrusion target detected by the radar with the intrusion target monitored by the vision sensor, otherwise, turning to step 4.
Further, in the step 3, the matching of the intrusion target detected by the radar with the intrusion target monitored by the vision sensor specifically includes:
converting the position of the intrusion target detected by the radar into the same space coordinate system;
calculating the distance deviation between the position of the intrusion target detected by the radar and the position of the intrusion target monitored by the vision sensor under the space coordinate system;
and if the distance deviation is smaller than the deviation threshold value, successful matching is achieved.
Based on the same conception, the invention also provides a perimeter intrusion detection system based on multi-sensor fusion, which comprises:
the radar is used for detecting intrusion scanning information along the protected perimeter, and the vision sensor is used for monitoring intrusion monitoring images along the protected perimeter;
Weather sensors for detecting weather environmental information along the protected perimeter;
a vibration detector for detecting vibration information along the protected perimeter;
the data processing device is connected with the radar, the vision sensor, the weather sensor and the vibration detector and is specifically used for:
whether an intrusion target exists is primarily judged according to the intrusion scanning information, if the intrusion target exists, whether the intrusion target is bad weather is judged according to the weather environment information, or whether the intrusion target is bad weather is judged according to the weather environment information, the intrusion monitoring image and the vibration information;
if the target is bad weather, deeply judging whether the target is invaded according to the invasion scanning information, if the target is invaded, sending out a pure radar alarm signal, otherwise, not sending out an alarm signal;
if the weather is not bad, judging whether an intrusion target exists according to the intrusion monitoring image, if the intrusion target exists, matching the intrusion target detected by the radar with the intrusion target monitored by the vision sensor, and if the intrusion target detected by the radar is successfully matched with the intrusion target monitored by the vision sensor, sending out a thunder alarm signal;
if the intrusion target detected by the radar is not successfully matched with the intrusion target detected by the visual sensor or the intrusion target is judged to be not existed according to the intrusion monitoring image, judging whether the intrusion target exists according to the vibration information, if the intrusion target exists, matching the intrusion target detected by the radar with the intrusion target detected by the vibration detector, and if the intrusion target detected by the radar is successfully matched with the intrusion target detected by the vibration detector, sending out Lei Zhen alarm signals; if the intrusion target detected by the radar is not successfully matched with the intrusion target detected by the vibration detector or the intrusion target is judged to be absent according to the vibration information, an alarm signal is not sent;
And the alarm device is connected with the data processing device and used for sending out an alarm according to the pure radar alarm signal, the thunder alarm signal or the thunder vibration alarm signal.
Further, the data processing device comprises at least one edge intelligent computing unit, wherein the edge intelligent computing unit is used for processing one or more of intrusion scanning information of the radar, intrusion monitoring images of the vision sensor and weather environment information of the weather sensor.
Based on the same concept, the present invention also provides a perimeter intrusion detection device, the device comprising:
a memory for storing a computer program;
and the processor is used for realizing the perimeter intrusion detection method based on multi-sensor fusion when executing the computer program.
Based on the same conception, the invention also provides a computer readable storage medium, wherein the computer readable storage medium is stored with a computer program, and the computer program is executed by a processor to realize the perimeter intrusion detection method based on multi-sensor fusion.
Advantageous effects
Compared with the prior art, the invention has the advantages that:
the invention integrates the data of the millimeter wave radar, the vision sensor, the weather sensor and the vibration detector, detects the invasion condition along the protected perimeter, can automatically judge the weather condition and automatically switch to an alarm strategy with higher reliability, works all the day, and has strong complex environment adaptability. In severe weather, the detection reliability of the visual sensor and the vibration detector is reduced, the detection reliability change of the radar is not obvious, and a pure radar alarm strategy is adopted; under non-severe weather, the detection reliability of the radar and the visual sensor is high, the alarm strategy of the radar and the visual sensor is adopted, the alarm accuracy is improved, the high sensitivity of the radar and the good visual classification performance are combined, the common small animals can be intelligently filtered, and false alarm is reduced; meanwhile, the visual sensor is possibly shielded or missed detection caused by other reasons, and an alarm strategy of a radar and a vibration detector is adopted, so that manual intervention is not needed, and the intelligent degree of detection is improved.
According to the invention, the radar is used for triggering detection, and the target can be found before the intrusion target approaches the protected perimeter, so that early warning of intrusion is realized; all alarms (namely pure radar alarm, radar alarm and Lei Zhen alarm) are triggered by the radar, and the high-sensitivity radar avoids the condition of missing report; meanwhile, one radar is provided with two vision sensors (namely, near targets and far targets are detected respectively), the near targets and the far targets are considered, when a plurality of targets are simultaneously at the near and far positions, the targets can be found at the same time, and the missing report phenomenon is improved.
Drawings
In order to more clearly illustrate the technical solutions of the present invention, the drawings that are needed in the description of the embodiments will be briefly described below, it being obvious that the drawing in the description below is only one embodiment of the present invention, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a perimeter intrusion detection method in an embodiment of the invention;
FIG. 2 is a block diagram of a perimeter intrusion detection system in an embodiment of the invention.
Detailed Description
The following description of the embodiments of the present invention will be made more apparent and fully by reference to the accompanying drawings, in which it is shown, however, only some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The technical scheme of the application is described in detail below by specific examples. The following embodiments may be combined with each other, and some embodiments may not be repeated for the same or similar concepts or processes.
Example 1
The method of the application can be suitable for intrusion detection of border lines, airport perimeters, railway perimeters, energy stations and the like, and the embodiment takes the railway line as an example.
The embodiment of the application provides a perimeter intrusion detection method based on multi-sensor fusion, wherein sensors comprise millimeter wave radar, a meteorological sensor, a visual sensor and a vibration detector. In this embodiment, the vibration detector adopts the vibration optical fiber, and the vibration optical fiber is located on the rail along the railway with the mode of hanging the net, and when invasion target can arouse the vibration of vibration optical fiber in the rail process of crossing or crossing, vibration information is gathered to the vibration optical fiber, judges whether there is invasion target invasion according to this vibration information. Or the vibration optical fiber is laid along the line of the fence in a buried mode, when an invasion target approaches the fence, the vibration of the ground can be caused, so that the vibration optical fiber vibrates, vibration information is collected by the vibration optical fiber, and whether the invasion target invades can be judged according to the vibration information.
The radar and the vision sensor are arranged at the same position, in the embodiment, the radar and the vision sensor are arranged at one side or above the railway line, and the radar and the vision sensor synchronously acquire data. The radar actively radiates electromagnetic waves outwards, after the electromagnetic waves hit a moving object, the frequency of the electromagnetic waves reflected by the moving object changes, the radar receives the reflected electromagnetic waves, and a moving object is detected according to the characteristics of the reflected electromagnetic waves, namely whether an intrusion object exists is judged.
A radar is provided with two vision sensors, wherein the two vision sensors are a first vision sensor for monitoring a near target and a second vision sensor for monitoring a far target, the radar and the two vision sensors are used for realizing non-blind area coverage of a monitoring area, the near target and the far target are considered, when a plurality of targets are simultaneously at the near position and the far position, the targets can be simultaneously found, and the missing report phenomenon is improved. In this embodiment, the vision sensor is a video camera, the first vision sensor is a short-focus camera, and the second vision sensor is a long-focus camera. In order to realize non-blind area coverage of the whole railway line, the railway line is divided into a plurality of monitoring areas, each monitoring area is divided into a plurality of subareas, each monitoring area is provided with a set of weather sensors, and each subarea is provided with a radar and two vision sensors. The distance between adjacent radars is determined by the detectable range of each radar. Each radar has a monitoring blind area at a near position, the distance of the monitoring blind area is determined by the range of a detection beam of each radar in the pitching direction, the installation pitch angle of the radar and the installation height of the radar, so that the monitoring blind area of the previous radar is covered when the next radar is installed, and the whole-line blind area-free coverage is realized.
The division of each sub-area is required to be dependent on the detection capability of the radar and vision sensor, the field perimeter shape and the topography relief, which have an effect on the radar and vision sensor's viewing conditions. For example, when the peripheral fence has a corner or a raised terrain or a pit, the range of the subareas needs to be adjusted according to the vision conditions of the radar and the vision sensor, so that the monitoring blind area is not ensured, the radar and the vision sensor as few as possible can be used, the number of sensors is reduced, and the cost is saved.
The weather sensor is used for collecting weather environment information of a corresponding area, and the types of the sensors contained in the weather sensor correspond to the types of severe weather to be detected. In this embodiment, the bad weather is at least one of strong wind, strong rain, strong snow, flying dust, dense fog, and hail weather, and therefore the weather sensor includes a wind speed sensor, a rain and snow sensor, a rainfall sensor, a temperature and humidity sensor, a flying dust detector, a dense fog detector, and a hail sensor.
As shown in fig. 1, the perimeter intrusion detection method based on multi-sensor fusion provided by the embodiment of the invention includes the following steps:
Step 1: acquiring intrusion scanning information acquired by a radar, intrusion monitoring images (including near-target intrusion monitoring images and far-target monitoring images) acquired by a visual sensor, vibration information acquired by a vibration optical fiber and weather environment information acquired by a weather sensor;
step 2: whether an intrusion target exists or not is primarily judged according to the intrusion scanning information, if the intrusion target exists, whether the intrusion target is bad weather or not is judged according to the weather environment information, the intrusion monitoring image and the vibration information, and if the intrusion target exists, the step 3 is carried out; if the weather is not bad, the step 4 is carried out;
step 3: deeply judging whether an intrusion target exists according to the intrusion scanning information, and if the intrusion target exists, sending out a pure radar alarm signal; otherwise, not sending out an alarm signal;
step 4: judging whether an intrusion target exists or not according to the intrusion monitoring image, if so, matching the intrusion target detected by the radar with the intrusion target monitored by the vision sensor, and if the matching is successful, sending out a thunder alarm signal; if the matching is unsuccessful, the step 5 is carried out; if no intrusion target is judged, the step 5 is carried out;
Step 5: judging whether an intrusion target exists according to the vibration information, if so, matching the intrusion target detected by the radar with the intrusion target detected by the vibration detector, and if so, sending out a Lei Zhen alarm signal; if the matching is unsuccessful, an alarm signal is not sent out; if no intrusion target is judged, no alarm signal is sent out.
The alarm mode of the invention comprises pure radar alarm, lei Zhen alarm and thunder alarm. Under severe weather environment, the credibility of the visual sensor and the vibration optical fiber is very low, the visual sensor is likely to cause rapid performance degradation and even complete failure due to thick fog, heavy rain, heavy snow and ground dust, meanwhile, false alarms of the vibration optical fiber can be obviously increased in heavy rain, heavy wind and hail weather, but the performance degradation of the radar is not obvious under the severe environment, at the moment, the intrusion scanning information collected by the radar is judged to be standard, and a pure radar alarm mode is adopted. Under the non-severe weather environment, the credibility of the radar and the vision sensor is high, and an alarm strategy of the radar and the vision sensor, namely, a radar alarm, can be adopted, so that the accuracy of intrusion target detection is improved; when the vision sensor detects no target due to shielding or other reasons, or the vision sensor detects an intrusion target but does not meet the corresponding threshold range requirement due to a detection frame or is successfully matched with the intrusion target detected by the radar, the intrusion target detected by the vision sensor is filtered, the alarm strategy of using the radar and the vibration optical fiber, namely Lei Zhen, is used for alarming, and the missing report phenomenon is avoided.
In step 2, two methods for judging bad weather are available: one is to judge whether the weather is bad weather according to the weather environment information, and the other is to judge whether the weather is bad weather according to the weather environment information, the intrusion monitoring image and the vibration information. In this embodiment, whether the weather is severe weather is determined according to weather environmental information, and the acquired value of each meteorological sensor is compared with a corresponding threshold value to determine whether the weather is severe weather, which specifically includes:
if the acquired value of the wind speed sensor is larger than the wind speed threshold value, the wind speed sensor is in a windy weather;
if the rain and snow sensor detects a rain and snow signal, the acquired value of the rain sensor is larger than the rain threshold value and the acquired value of the temperature and humidity sensor is larger than the temperature and humidity threshold value, the weather is heavy rain; if the rain and snow sensor detects a rain and snow signal, the acquired value of the rain sensor is smaller than the rain threshold value and the acquired value of the temperature and humidity sensor is smaller than the temperature and humidity threshold value, the weather is heavy snow;
if the collection value of the dust detector is larger than the dust threshold value, the dust detector is in dust weather; if the acquisition value of the dense fog detector is larger than the dense fog threshold value, the dense fog weather is the dense fog weather; if the collected value of the hail sensor is larger than the hail threshold value, the hail sensor is in hail weather.
In order to judge severe weather more accurately, based on weather environment information, intrusion monitoring images and vibration information are combined, for example, shaking of a fence can be caused when weather is heavy wind, heavy rain or hail, so that vibration information is affected; the visibility of the environment can be influenced by the thick fog and dust weather, and then the detection distance of the vision sensor is influenced. In this embodiment, whether the weather is severe weather is judged according to weather environment information, intrusion monitoring images and vibration information, specifically including:
obtaining vibration information under normal weather conditions and vibration information under severe weather conditions; taking vibration information under different weather as an input sample, and taking weather environment information corresponding to the vibration information as a sample label to construct a sample data set; constructing a weather prediction model, and training the weather prediction model by using a sample data set to obtain a target weather prediction model;
the method comprises the steps of collecting weather environment information by using a weather sensor and collecting vibration information by using a vibration detector;
if the acquired value of the wind speed sensor is larger than the wind speed threshold value and the predicted result of predicting the vibration information by using the target weather prediction model is strong wind, the predicted result is strong wind weather;
If the rain and snow sensor detects that the rain and snow signal is detected, the acquired value of the rain sensor is larger than the rain threshold value, the acquired value of the temperature and humidity sensor is larger than the temperature and humidity threshold value, and the predicted result of predicting the vibration information by using the target weather prediction model is heavy rain, the predicted result is heavy rain weather; if the rain and snow sensor detects that the rain and snow signal is detected, the acquired value of the rain sensor is smaller than the rain threshold value, the acquired value of the temperature and humidity sensor is smaller than the temperature and humidity threshold value, and the predicted result of predicting the vibration information by using the target weather prediction model is big snow, the predicted result is big snow weather;
if the acquired value of the hail sensor is larger than the hail threshold value and the predicted result of predicting the vibration information by using the target weather prediction model is hail, the hail weather is obtained;
if the collection value of the dust detector is larger than the dust threshold value and the visibility of the intrusion monitoring image is lower than the visibility threshold value, the dust is in dust weather; and if the acquisition value of the dense fog detector is larger than the dense fog threshold value and the visibility of the intrusion monitoring image is lower than the visibility threshold value, the dense fog weather is obtained.
In this embodiment, the weather prediction model is an artificial intelligence model, such as a neural network model. The target weather prediction model obtained by training the weather prediction model by utilizing the sample data set has weather prediction or identification capability, and the vibration information acquired by the vibration optical fiber in real time is input into the target weather prediction model, so that the current weather condition can be predicted or identified, and the accuracy of weather judgment is improved.
The radar detection range is wide, the monitoring distance is long, the radar is adopted to actively detect the invasion target, and the invasion target can be found and early-warned in advance when the invasion target does not enter the railway perimeter. After the radar detects an intrusion target, judging whether the intrusion target is bad weather, if so, deeply judging whether the intrusion target exists or not according to intrusion scanning information of the radar, wherein the method specifically comprises the following steps of:
determining the confidence coefficient of an intrusion target according to intrusion scanning information;
judging whether the confidence coefficient is larger than a confidence coefficient threshold value, if so, judging that an intrusion target exists, and sending out a pure radar alarm signal; otherwise, judging that the target is not invaded, and not sending out an alarm signal.
The radar judges whether an intrusion target exists according to the characteristics of the reflected electromagnetic waves, and the confidence of the intrusion target can be further determined according to the signal-to-noise ratio of the reflected electromagnetic waves and the times of continuously received electromagnetic waves. The higher the signal-to-noise ratio of the reflected electromagnetic wave, the higher the confidence; the higher the number of consecutively received electromagnetic waves, the higher the confidence.
In step 4, the near target intrusion monitoring image and the far target intrusion monitoring image output by the vision sensor corresponding to the radar are respectively subjected to vision analysis, whether an intrusion target exists in an image picture is judged, if the intrusion target exists, the type of the intrusion target and the position and the area of the intrusion target in the image are output, the position of the intrusion target in the image can be represented by adopting coordinates in an image coordinate system, and the area of the intrusion target is represented by adopting a rectangular vision frame or a detection frame with a certain length and a certain width. In step 4, after judging that there is an intrusion target according to the intrusion monitoring image, before matching the intrusion target detected by the radar with the intrusion target monitored by the vision sensor, the detection method further includes:
Determining the type of an intrusion target and the position and the area of the intrusion target in the image according to the intrusion monitoring image;
judging whether the area of the intrusion target in the image and the aspect ratio of the intrusion target in the image are in accordance with the threshold range of the corresponding type, if so, matching the intrusion target detected by the radar with the intrusion target monitored by the vision sensor, otherwise, turning to step 5. That is, if the area of the visual frame is not within the set area threshold and the aspect ratio of the visual frame is not within the set aspect ratio threshold, it is necessary to perform analysis in combination with the determination result of the vibration information.
In this embodiment, the visual analysis of the intrusion monitoring image may collect an artificial intelligence algorithm based on a neural network, specifically: the method comprises the steps of obtaining an intrusion monitoring image, manually labeling different images of the intrusion monitoring image, forming a sample data set by the labeled images, then constructing an identification model, training the identification model by using the sample data set to obtain a target identification model, and finally identifying the intrusion monitoring image acquired in real time by using the target identification model to output the type of an intrusion target and a visual frame of the intrusion target.
The types of the invasion targets are different along with different application scenes, and common invasion target types are human beings, vehicles, animals and the like. For different types of intrusion targets, the area threshold range and the aspect ratio threshold range corresponding to the visual frame are also different, and the threshold ranges can be determined in an analog mode, specifically: the human simulation intrusion target, the vision sensor collects images of the intrusion target at different positions, so that the occupied area and the aspect ratio of the intrusion target in the images at different positions are obtained, and the area threshold range and the aspect ratio threshold range are determined. The area threshold range and the aspect ratio threshold range of other types of intrusion targets may be determined in a similar manner. For pedestrian intrusion targets, the visual box is narrow and high, with an aspect ratio typically less than 1; for animal intrusion targets, the aspect ratio is typically around 1 and not less than 1/2.
In non-severe weather, if the radar detects an intrusion target, but the corresponding vision sensor does not find the target all the time, the radar is likely to generate false alarm, so that no alarm is directly generated, and the judgment result of vibration information needs to be further fused for judgment. The purpose of this is primarily to prevent the occurrence of false alarms, as the vision sensor may fail to properly detect an intruding object, such as a type of object that has not been learned by a camouflaged person or artificial intelligence algorithm, including but not limited to large animals such as camels, cattle, sheep, horses, bears, etc., whose invasive activity typically also requires an alarm.
In step 4, in order to avoid false alarm caused by false alarm generated by the radar and the vision sensor at the same time, the vision frame of the intrusion target monitored by the vision sensor is firstly judged, including area and aspect ratio, if the radar and the vision sensor both monitor the intrusion target in non-severe weather. If the area and the aspect ratio of the visual frame are both in accordance with the set threshold range, matching an intrusion target detected by the radar with an intrusion target monitored by the visual sensor; if the set threshold range is not met, the false detection caused by small animals or other objects is likely to be indicated, and the targets are filtered through the measures so as to avoid false alarm. When judging whether the visual frame accords with the set threshold range, a certain error probability exists, so that the judgment result of the vibration information is further combined to prevent the occurrence of false alarm.
In order to avoid false alarm caused by false alarm generated by the radar and the vision sensor at the same time, after the intrusion target monitored by the vision sensor is judged, the intrusion target detected by the radar needs to be further matched with the intrusion target monitored by the vision sensor. In the matching process, the type, the spatial position coordinate, the speed, the scattering cross section area, the signal to noise ratio and the like of the intrusion target detected by the radar can be adopted, and the type, the position and the size in an image coordinate system, the classification confidence and the like of the intrusion target monitored by the vision sensor can be adopted. In this embodiment, the matching process is described by taking a position as an example, and matching between an intrusion target detected by the radar and an intrusion target monitored by the vision sensor needs to be performed on the basis of the same spatial coordinate and time synchronization, so that data acquisition of the radar and the vision sensor needs to be synchronized. In this embodiment, matching an intrusion target detected by a radar with an intrusion target monitored by a vision sensor specifically includes:
Converting the position of the intrusion target detected by the radar into the same space coordinate system;
calculating the distance deviation between the position of the intrusion target detected by the radar and the position of the intrusion target monitored by the vision sensor in a space coordinate system;
if the distance deviation is smaller than the deviation threshold, the matching is successful, which means that the intrusion targets monitored by the radar and the vision sensor are the same target, and a radar alarm signal is generated; otherwise, the matching is unsuccessful, which means that the intrusion targets monitored by the radar and the vision sensor are not the same target, and are likely to be false alarms generated by the radar and the vision sensor, so that the alarm signals are not generated immediately, and judgment is carried out by fusing the judgment results of the vibration information. The purpose of this operation is mainly to prevent false matches from occurring, as the intrusion target detected by the vision sensor is likely to be unsuccessful in matching with the intrusion target detected by the radar due to partial occlusion or the like.
In non-bad weather, if the radar detects an intrusion target, the vision sensor does not detect the intrusion target, or although the vision sensor detects the intrusion target, the judgment result of the vibration information is fused because the area and the aspect ratio of the vision frame of the intrusion target do not accord with the set threshold range, or the intrusion target detected by the vision sensor is not successfully matched with the intrusion target detected by the radar. If the vibrating optical fiber does not detect an intrusion target, not alarming; and if the vibrating optical fiber detects the intrusion target, matching the intrusion target detected by the radar with the intrusion target detected by the vibrating optical fiber. In the matching process, the type, the spatial position coordinate, the speed, the scattering sectional area, the signal to noise ratio and the like of the intrusion target detected by the radar, the type, the position, the classification confidence and the like of the intrusion target detected by the vibration optical fiber can be adopted. The matching of the intrusion target detected by the radar and the intrusion target detected by the vibration optical fiber is performed on the basis of the spatial calibration of the radar and the vibration optical fiber. If the matching is successful, generating a lightning alarming signal; if the matching is not successful, the intrusion target detected by the radar and the intrusion target detected by the vibration optical fiber are not the same target, and are likely to be false alarms generated respectively, so that an alarm signal is not generated.
In this embodiment, the matching between the intrusion target detected by the radar and the intrusion target detected by the vibration optical fiber is the prior art, and specific reference may be made to the patent document with the grant bulletin number CN110126885B, named as a method and a system for monitoring the intrusion target around the railway.
The invention combines the advantages of each sensor, has strong adaptability to complex environment, and has the advantages of full coverage, all weather, detection and early warning in all days, zero missing report and near zero false report.
The detection reliability of the visual sensor and the vibration detector is reduced in severe weather, the detection reliability change of the radar is not obvious, and a pure radar alarm strategy is adopted; under non-severe weather, the detection reliability of the radar and the visual sensor is high, and the alarm accuracy is improved by adopting an alarm strategy of the radar and the visual sensor; meanwhile, the visual sensor is possibly shielded or missed detection caused by other reasons, and an alarm strategy of a radar and a vibration detector is adopted, so that manual intervention is not needed, and the intelligent degree of detection is improved.
The millimeter wave radar is less influenced by illumination and weather factors, has higher stability and higher ranging precision, and can make up the defects of the camera in the aspects; the camera has good visual effect, can accurately classify and identify the invasion target, is convenient for post evidence collection, and can make up the defect of radar in the aspect; the vibration optical fiber can be directly paved on various wire nets or buried under the ground, is suitable for various complex terrains, can detect irregular peripheral defense areas, has high sensitivity, and therefore overcomes the defect that radars and cameras can only be used for viewing scenes, ensures effective detection rate, and can overcome the defect that the vibration optical fiber cannot be early warned in advance.
Therefore, the problem that can be solved by a single sensor is very limited, the millimeter wave radar, the camera and the vibration optical fiber are fused to make advantage complementation, and compared with the traditional perimeter protection means, the millimeter wave radar, the camera and the vibration optical fiber have great advantages in the aspect of target intrusion monitoring and data acquisition in the aspect of linear perimeter protection, can provide accurate, reliable and real-time alarm information, and promote the intellectualization of an alarm system.
Example 2
As shown in fig. 2, the perimeter intrusion detection system based on multi-sensor fusion provided by the embodiment of the invention further comprises a radar, a vision sensor, a weather sensor, a vibration detector, a data processing device and an alarm device; the radar, the vision sensor, the weather sensor and the vibration detector are respectively connected with the data processing device, and the data processing device is connected with the alarm device.
In this embodiment, radar and vision sensor all locate by the one side or the top of protection perimeter along the line, and the radar is used for detecting by the invasion scanning information of protection perimeter along the line, and vision sensor is used for monitoring by the invasion monitoring image of protection perimeter along the line, and radar and vision sensor synchronous acquisition data.
A radar is provided with two vision sensors, wherein the two vision sensors are a first vision sensor for monitoring a near target and a second vision sensor for monitoring a far target, the radar and the two vision sensors are used for realizing non-blind area coverage of a monitoring area, the near target and the far target are considered, when a plurality of targets are simultaneously at the near position and the far position, the targets can be simultaneously found, and the missing report phenomenon is improved. In this embodiment, the vision sensor is a video camera, the first vision sensor is a short-focus camera, and the second vision sensor is a long-focus camera. In order to realize non-blind area coverage of the whole railway line, the railway line is divided into a plurality of monitoring areas, each monitoring area is divided into a plurality of subareas, each monitoring area is provided with a set of weather sensors, and each subarea is provided with a radar and two vision sensors. The distance between adjacent radars is determined by the detectable range of each radar. Each radar has a monitoring blind area at a near position, the distance of the monitoring blind area is determined by the range of a detection beam of each radar in the pitching direction, the installation pitch angle of the radar and the installation height of the radar, so that the monitoring blind area of the previous radar is covered when the next radar is installed, and the whole-line blind area-free coverage is realized.
The weather sensor is used for collecting weather environment information of a corresponding area, and the types of the sensors contained in the weather sensor correspond to the types of severe weather to be detected. In this embodiment, the bad weather is at least one of strong wind, strong rain, strong snow, flying dust, dense fog, and hail weather, and therefore the weather sensor includes a wind speed sensor, a rain and snow sensor, a rainfall sensor, a temperature and humidity sensor, a flying dust detector, a dense fog detector, and a hail sensor.
In this embodiment, the vibration detector adopts a vibration optical fiber (may also adopt a vibration sensor), the vibration optical fiber is arranged on the rail along the railway in a net hanging manner, when an intrusion target can cause vibration of the vibration optical fiber in the process of crossing or traversing the rail, the vibration optical fiber collects vibration information, and whether the intrusion target intrudes is judged according to the vibration information. Or the vibration optical fiber is laid along the line of the fence in a buried mode, when an invasion target approaches the fence, the vibration of the ground can be caused, so that the vibration optical fiber vibrates, vibration information is collected by the vibration optical fiber, and whether the invasion target invades can be judged according to the vibration information.
The data processing device is specifically used for:
Whether an intrusion target exists is primarily judged according to the intrusion scanning information, if the intrusion target exists, whether the intrusion target is bad weather is judged according to the weather environment information, or whether the intrusion target is bad weather is judged according to the weather environment information, the intrusion monitoring image and the vibration information;
if the target is bad weather, deeply judging whether the target is invaded according to the invasion scanning information, if the target is invaded, sending out a pure radar alarm signal, otherwise, not sending out an alarm signal;
if the weather is not bad, judging whether an intrusion target exists according to the intrusion monitoring image, if the intrusion target exists, matching the intrusion target detected by the radar with the intrusion target monitored by the vision sensor, and if the intrusion target detected by the radar is successfully matched with the intrusion target monitored by the vision sensor, sending out a thunder alarm signal;
if the intrusion target detected by the radar is not successfully matched with the intrusion target detected by the visual sensor or the intrusion target is judged to be not existed according to the intrusion monitoring image, judging whether the intrusion target exists according to the vibration information, if the intrusion target exists, matching the intrusion target detected by the radar with the intrusion target detected by the vibration detector, and if the intrusion target detected by the radar is successfully matched with the intrusion target detected by the vibration detector, sending out Lei Zhen alarm signals; if the intrusion target detected by the radar is not successfully matched with the intrusion target detected by the vibration detector or the intrusion target is judged to be absent according to the vibration information, an alarm signal is not sent.
The alarm device is used for giving an alarm according to the pure radar alarm signal, the thunder alarm signal or the thunder vibration alarm signal.
The data processing device comprises a vibrating optical fiber host, a network switch and a rear-end server, wherein the vibrating optical fiber host, a radar, a vision sensor and an air image sensor are connected with the rear-end server through the network switch, and the vibrating optical fiber host is used for processing vibration information acquired by a vibrating optical fiber and obtaining a judging result; the back-end server is used for processing intrusion scanning information of the radar, intrusion monitoring images of the vision sensor and weather environment information of the weather sensor, and fusing judgment results of the vibration optical fiber host to judge whether an alarm is generated. The back-end server is also used for storing various collected data and corresponding alarm information.
In another embodiment, the data processing device further includes at least one edge intelligent computing unit, the edge intelligent computing unit is connected with one or more of the vibrating fiber host, the radar, the vision sensor and the weather sensor, and is used for processing one or more of intrusion scanning information of the radar, intrusion monitoring images of the vision sensor and weather environment information of the weather sensor, and the vibrating fiber host and the edge intelligent computing unit are connected with the back-end server through the network switch. Meanwhile, the vibration optical fiber host and the edge intelligent computing unit send the processed data to a back-end server through a network switch, and the back-end server fuses all the data to judge whether an alarm is generated or not; or the vibration optical fiber host and the back-end server send the processed data to the edge intelligent computing unit through the network switch, and the edge intelligent computing unit fuses all the data to judge whether an alarm is generated. The edge intelligent computing unit processes the data at the front end, has high instantaneity, and uploads the processed data to the back end server, so that the requirement of data transmission on network bandwidth is reduced, and the computing pressure of the back end server is reduced. Meanwhile, when the network connection between the rear-end server and the front-end radar and other devices is disconnected, data cannot be uploaded to the monitoring center in time, and the problem that data processing cannot be performed in time and whether an alarm is judged or not is avoided by the edge intelligent computing unit.
An edge intelligent computing unit may be provided for each radar, or a plurality of the radar may be provided. When the vibration optical fiber is replaced by the vibration sensor, the vibration sensor is arranged on the fence every several meters, and all the vibration sensors are connected in series through the cable, so that the purpose of long-distance perimeter management and control is achieved.
Example 3
The embodiment of the invention also provides electronic equipment, which comprises: a processor and a memory storing a computer program, the processor being configured to implement the perimeter intrusion detection method as described above when executing the computer program.
Although not shown, the electronic device includes a processor that can perform various appropriate operations and processes according to programs and/or data stored in a Read Only Memory (ROM) or programs and/or data loaded from a storage portion into a Random Access Memory (RAM). The processor may be a multi-core processor or may include a plurality of processors. In some embodiments, the processor may comprise a general-purpose main processor and one or more special coprocessors, such as, for example, a Central Processing Unit (CPU), a Graphics Processor (GPU), a neural Network Processor (NPU), a Digital Signal Processor (DSP), and so forth. In the RAM, various programs and data required for the operation of the electronic device are also stored. The processor, ROM and RAM are connected to each other by a bus. An input/output (I/O) interface is also connected to the bus.
The above-described processor is used in combination with a memory to execute a program stored in the memory, which when executed by a computer is capable of implementing the methods, steps or functions described in the above-described embodiments.
Although not shown, embodiments of the present invention also provide a computer-readable storage medium having a computer program stored thereon, which when executed by a processor, implements the perimeter intrusion detection method as described above.
Storage media in embodiments of the invention include both permanent and non-permanent, removable and non-removable items that may be used to implement information storage by any method or technology. Examples of storage media include, but are not limited to, phase change memory (PRAM), static Random Access Memory (SRAM), dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), read Only Memory (ROM), electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, read only compact disc read only memory (CD-ROM), digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape storage or other magnetic storage devices, or any other non-transmission medium which can be used to store information that can be accessed by a computing device.
The foregoing disclosure is merely illustrative of specific embodiments of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art will readily recognize that changes and modifications are possible within the scope of the present invention.
Claims (13)
1. The perimeter intrusion detection method based on multi-sensor fusion is characterized in that the sensors comprise a radar, a meteorological sensor, a visual sensor and a vibration detector, wherein the radar and the visual sensor are arranged on one side or above the protected perimeter line, and the radar and the visual sensor are synchronously collected, and the detection method comprises the following steps:
step 1: whether an intrusion target exists or not is primarily judged according to the intrusion scanning information of the radar, if the intrusion target exists, whether the intrusion target is bad weather or not is judged according to the weather environment information of the weather sensor, or whether the intrusion target is bad weather or not is judged according to the weather environment information, the intrusion monitoring image of the visual sensor and the vibration information along the protected perimeter of the vibration detector, if the intrusion target exists, the step 2 is carried out; if the weather is not bad, the step 3 is carried out;
Step 2: deeply judging whether an intrusion target exists according to the intrusion scanning information, and if so, sending out a pure radar alarm signal; otherwise, not sending out an alarm signal;
step 3: judging whether an intrusion target exists or not according to the intrusion monitoring image, if so, matching the intrusion target detected by the radar with the intrusion target monitored by the vision sensor, and if the matching is successful, sending out a thunder alarm signal; if the matching is unsuccessful, the step 4 is carried out; if no intrusion target is judged, the step 4 is carried out;
step 4: judging whether an intrusion target exists or not according to the vibration information, if yes, matching the intrusion target detected by the radar with the intrusion target detected by the vibration detector, and if successful, sending out a Lei Zhen alarm signal; if the matching is unsuccessful, an alarm signal is not sent out; if no intrusion target is judged, no alarm signal is sent out.
2. The perimeter intrusion detection method based on multi-sensor fusion of claim 1, wherein: the vibration detector is arranged on the fence along the protected perimeter in a net hanging manner, or is laid close to the fence in a buried manner.
3. The perimeter intrusion detection method based on multi-sensor fusion of claim 1, wherein: dividing the protected perimeter into a plurality of monitoring areas along the line, dividing each monitoring area into a plurality of subareas, configuring a set of weather sensors in each monitoring area, configuring a radar and two vision sensors in each subarea, wherein the two vision sensors are respectively a first vision sensor for monitoring a near target and a second vision sensor for monitoring a far target.
4. A perimeter intrusion detection method based on multisensor fusion according to claim 1 or 3, wherein: the weather sensor comprises a wind speed sensor, a rain and snow sensor, a rainfall sensor, a temperature and humidity sensor, a dust detector, a thick fog detector and a hail sensor; the bad weather is at least one of strong wind, heavy rain, strong snow, flying dust, dense fog and hail weather.
5. The perimeter intrusion detection method based on multi-sensor fusion of claim 4, wherein: in the step 1, determining whether the weather is severe weather according to weather environment information of the weather sensor specifically includes:
If the acquired value of the wind speed sensor is larger than the wind speed threshold value, the wind speed sensor is in a windy weather;
if the rain and snow sensor detects a rain and snow signal, the acquired value of the rain sensor is larger than the rain threshold value and the acquired value of the temperature and humidity sensor is larger than the temperature and humidity threshold value, the weather is heavy rain; if the rain and snow sensor detects a rain and snow signal, the acquired value of the rain sensor is smaller than the rain threshold value and the acquired value of the temperature and humidity sensor is smaller than the temperature and humidity threshold value, the weather is heavy snow;
if the collection value of the dust detector is larger than the dust threshold value, the dust detector is in dust weather; if the acquisition value of the dense fog detector is larger than the dense fog threshold value, the dense fog weather is the dense fog weather; if the collected value of the hail sensor is larger than the hail threshold value, the hail sensor is in hail weather.
6. The perimeter intrusion detection method based on multi-sensor fusion of claim 4, wherein: in the step 1, judging whether the weather is severe weather or not according to weather environment information, intrusion monitoring images and vibration information, specifically comprising:
obtaining vibration information under normal weather conditions and vibration information under severe weather conditions; taking vibration information under different weather as an input sample, and taking weather environment information corresponding to the vibration information as a sample label to construct a sample data set; constructing a weather prediction model, and training the weather prediction model by utilizing the sample data set to obtain a target weather prediction model;
The method comprises the steps of collecting weather environment information by using a weather sensor and collecting vibration information by using a vibration detector;
if the acquired value of the wind speed sensor is larger than the wind speed threshold value and the predicted result of predicting the vibration information by using the target weather prediction model is strong wind, the predicted result is strong wind weather;
if the rain and snow sensor detects that the rain and snow signal is detected, the acquired value of the rain sensor is larger than the rain threshold value, the acquired value of the temperature and humidity sensor is larger than the temperature and humidity threshold value, and the predicted result of predicting the vibration information by using the target weather prediction model is heavy rain, the predicted result is heavy rain weather; if the rain and snow sensor detects that the rain and snow signal is detected, the acquired value of the rain sensor is smaller than the rain threshold value, the acquired value of the temperature and humidity sensor is smaller than the temperature and humidity threshold value, and the predicted result of predicting the vibration information by using the target weather prediction model is big snow, the predicted result is big snow weather;
if the acquired value of the hail sensor is larger than the hail threshold value and the predicted result of predicting the vibration information by using the target weather prediction model is hail, the hail weather is obtained;
if the collection value of the dust detector is larger than the dust threshold value and the visibility of the intrusion monitoring image is lower than the visibility threshold value, the dust is in dust weather; and if the acquisition value of the dense fog detector is larger than the dense fog threshold value and the visibility of the intrusion monitoring image is lower than the visibility threshold value, the dense fog weather is obtained.
7. The perimeter intrusion detection method based on multi-sensor fusion of claim 1, wherein: in the step 2, whether an intrusion target exists or not is deeply judged according to the intrusion scanning information, which specifically comprises the following steps:
determining the confidence coefficient of the intrusion target according to the intrusion scanning information;
judging whether the confidence coefficient is larger than a confidence coefficient threshold value, if so, judging that an intrusion target exists; otherwise, judging that the target is not invaded.
8. The perimeter intrusion detection method based on multi-sensor fusion of claim 1, wherein: in the step 3, after judging that there is an intrusion target according to the intrusion monitoring image, before matching the intrusion target detected by the radar with the intrusion target monitored by the vision sensor, the detection method further includes:
determining the type of the intrusion target and the position and the area of the intrusion target in the image according to the intrusion monitoring image;
judging whether the area of the intrusion target in the image and the aspect ratio of the intrusion target in the image are in accordance with the corresponding threshold range, if so, matching the intrusion target detected by the radar with the intrusion target monitored by the vision sensor, otherwise, turning to step 4.
9. The perimeter intrusion detection method based on multi-sensor fusion of claim 1, wherein: in the step 3, matching the intrusion target detected by the radar with the intrusion target monitored by the vision sensor specifically includes:
converting the position of the intrusion target detected by the radar into the same space coordinate system;
calculating the distance deviation between the position of the intrusion target detected by the radar and the position of the intrusion target monitored by the vision sensor under the space coordinate system;
and if the distance deviation is smaller than the deviation threshold value, successful matching is achieved.
10. A perimeter intrusion detection system based on multi-sensor fusion, the system comprising:
the radar is used for detecting intrusion scanning information along the protected perimeter, and the vision sensor is used for monitoring intrusion monitoring images along the protected perimeter;
weather sensors for detecting weather environmental information along the protected perimeter;
a vibration detector for detecting vibration information along the protected perimeter;
The data processing device is connected with the radar, the vision sensor, the weather sensor and the vibration detector and is specifically used for:
whether an intrusion target exists is primarily judged according to the intrusion scanning information, if the intrusion target exists, whether the intrusion target is bad weather is judged according to the weather environment information, or whether the intrusion target is bad weather is judged according to the weather environment information, the intrusion monitoring image and the vibration information;
if the target is bad weather, deeply judging whether the target is invaded according to the invasion scanning information, if the target is invaded, sending out a pure radar alarm signal, otherwise, not sending out an alarm signal;
if the weather is not bad, judging whether an intrusion target exists according to the intrusion monitoring image, if the intrusion target exists, matching the intrusion target detected by the radar with the intrusion target monitored by the vision sensor, and if the intrusion target detected by the radar is successfully matched with the intrusion target monitored by the vision sensor, sending out a thunder alarm signal;
if the intrusion target detected by the radar is not successfully matched with the intrusion target detected by the visual sensor or the intrusion target is judged to be not existed according to the intrusion monitoring image, judging whether the intrusion target exists according to the vibration information, if the intrusion target exists, matching the intrusion target detected by the radar with the intrusion target detected by the vibration detector, and if the intrusion target detected by the radar is successfully matched with the intrusion target detected by the vibration detector, sending out Lei Zhen alarm signals; if the intrusion target detected by the radar is not successfully matched with the intrusion target detected by the vibration detector or the intrusion target is judged to be absent according to the vibration information, an alarm signal is not sent;
And the alarm device is connected with the data processing device and used for sending out an alarm according to the pure radar alarm signal, the thunder alarm signal or the thunder vibration alarm signal.
11. The multi-sensor fusion based perimeter intrusion detection system according to claim 10, wherein the data processing device comprises at least one edge intelligent computing unit for processing one or more of intrusion scan information of a radar, intrusion monitoring images of a vision sensor, and weather environment information of a weather sensor.
12. A perimeter intrusion detection device, the device comprising:
a memory for storing a computer program;
a processor for implementing the perimeter intrusion detection method based on multi-sensor fusion according to any one of claims 1 to 9 when executing the computer program.
13. A computer readable storage medium, wherein a computer program is stored on the computer readable storage medium, and when executed by a processor, the computer program implements the perimeter intrusion detection method based on multi-sensor fusion according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311197663.2A CN116935551A (en) | 2023-09-18 | 2023-09-18 | Perimeter intrusion detection method, system, equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311197663.2A CN116935551A (en) | 2023-09-18 | 2023-09-18 | Perimeter intrusion detection method, system, equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116935551A true CN116935551A (en) | 2023-10-24 |
Family
ID=88382934
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311197663.2A Pending CN116935551A (en) | 2023-09-18 | 2023-09-18 | Perimeter intrusion detection method, system, equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116935551A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117310691A (en) * | 2023-11-30 | 2023-12-29 | 浙江宇视科技有限公司 | Multi-mode radar target positioning method, device, electronic equipment and storage medium |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101048806A (en) * | 2004-08-27 | 2007-10-03 | 新加坡科技劲力私人有限公司 | Multi-sensor intrusion detection system |
| CN105389920A (en) * | 2015-11-30 | 2016-03-09 | 北京正安维视科技股份有限公司 | Perimeter alarm method based on two-dimensional laser radar and video analysis |
| CN206136125U (en) * | 2016-08-03 | 2017-04-26 | 湖南华诺星空电子技术有限公司 | Radar video fuses alarm device , system and equipment |
| CN110126885A (en) * | 2018-02-02 | 2019-08-16 | 保定市天河电子技术有限公司 | A kind of railway circumference intrusion target monitoring method and system |
| CN113484858A (en) * | 2021-07-07 | 2021-10-08 | 深圳市商汤科技有限公司 | Intrusion detection method and system |
| KR102377226B1 (en) * | 2021-09-02 | 2022-03-22 | 주식회사 세렉스 | Machine learning-based perimeter intrusion detection system using radar |
| CN115601902A (en) * | 2021-07-08 | 2023-01-13 | 华为技术有限公司（Cn） | Perimeter detection method, device, equipment and system |
| CN115691018A (en) * | 2022-10-26 | 2023-02-03 | 哈尔滨市科佳通用机电股份有限公司 | Railway perimeter intrusion monitoring and early warning method and system based on multi-sensor fusion |
| CN116740874A (en) * | 2022-03-01 | 2023-09-12 | 华为技术有限公司 | Intrusion detection method and related device |
-
2023
- 2023-09-18 CN CN202311197663.2A patent/CN116935551A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101048806A (en) * | 2004-08-27 | 2007-10-03 | 新加坡科技劲力私人有限公司 | Multi-sensor intrusion detection system |
| CN105389920A (en) * | 2015-11-30 | 2016-03-09 | 北京正安维视科技股份有限公司 | Perimeter alarm method based on two-dimensional laser radar and video analysis |
| CN206136125U (en) * | 2016-08-03 | 2017-04-26 | 湖南华诺星空电子技术有限公司 | Radar video fuses alarm device , system and equipment |
| CN110126885A (en) * | 2018-02-02 | 2019-08-16 | 保定市天河电子技术有限公司 | A kind of railway circumference intrusion target monitoring method and system |
| CN113484858A (en) * | 2021-07-07 | 2021-10-08 | 深圳市商汤科技有限公司 | Intrusion detection method and system |
| CN115601902A (en) * | 2021-07-08 | 2023-01-13 | 华为技术有限公司（Cn） | Perimeter detection method, device, equipment and system |
| KR102377226B1 (en) * | 2021-09-02 | 2022-03-22 | 주식회사 세렉스 | Machine learning-based perimeter intrusion detection system using radar |
| CN116740874A (en) * | 2022-03-01 | 2023-09-12 | 华为技术有限公司 | Intrusion detection method and related device |
| CN115691018A (en) * | 2022-10-26 | 2023-02-03 | 哈尔滨市科佳通用机电股份有限公司 | Railway perimeter intrusion monitoring and early warning method and system based on multi-sensor fusion |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117310691A (en) * | 2023-11-30 | 2023-12-29 | 浙江宇视科技有限公司 | Multi-mode radar target positioning method, device, electronic equipment and storage medium |
| CN117310691B (en) * | 2023-11-30 | 2024-02-13 | 浙江宇视科技有限公司 | Multi-mode radar target positioning method, device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112418069B (en) | High-altitude parabolic detection method and device, computer equipment and storage medium | |
| CN103176185B (en) | Method and system for detecting road barrier | |
| CN109920185A (en) | One kind merging the mobile mesh calibration method of detection with video data based on millimetre-wave radar | |
| KR101927364B1 (en) | Outside Intruding and Monitering Radar Syatem Based on Deep -Learning and Method thereof | |
| CN110126885B (en) | Railway perimeter intrusion target monitoring method and system | |
| CN113671480A (en) | Radar and video fusion traffic target tracking method, system, equipment and terminal | |
| AU2017359142B2 (en) | Systems and methods for detecting flying animals | |
| CN114419825A (en) | High-speed rail perimeter intrusion monitoring device and method based on millimeter wave radar and camera | |
| CN116935551A (en) | Perimeter intrusion detection method, system, equipment and storage medium | |
| CN112776856A (en) | Track foreign matter intrusion monitoring method, device and system and monitoring host equipment | |
| CN114445803A (en) | Driving data processing method and device and electronic equipment | |
| CN115184917B (en) | Regional target tracking method integrating millimeter wave radar and camera | |
| CN113484858A (en) | Intrusion detection method and system | |
| CN113219454A (en) | System and method for preventing vehicles and personnel from invading runway based on millimeter wave radar | |
| CN115272810A (en) | Target detection method, system, equipment and medium based on radar vision fusion | |
| CN115034324B (en) | Multi-sensor fusion perception efficiency enhancement method | |
| CN112949566A (en) | Monitoring method, device and system and computer storage medium | |
| CN115083088A (en) | Railway perimeter intrusion early warning method | |
| CN112349055A (en) | Target monitoring device and method based on radar and video linkage | |
| CN115691018B (en) | Railway perimeter intrusion monitoring and early warning method and system based on multi-sensor fusion | |
| KR102440169B1 (en) | Smart guard system for improving the accuracy of effective detection through multi-sensor signal fusion and AI image analysis | |
| CN115856908A (en) | Vehicle speed measurement method and system based on three-dimensional laser radar | |
| CN114882661A (en) | Outdoor early warning method, device, system and computer readable storage medium | |
| CN114219825A (en) | Target contact force tracking method and system under multiple photoelectric distributed deployments | |
| CN117238143B (en) | Traffic data fusion method, system and device based on radar double-spectrum camera |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |