Station intelligent safety system based on video detection

Abstract

The invention discloses a station intelligent safety system based on video detection. The system comprises an acquisition layer, a response execution layer, a data storage layer, a data processing layer and a command center layer. Wherein the acquisition layer comprises a first-class acquisition module, a second-class acquisition module and a third-class acquisition module; the first-class acquisition module is arranged in a platform area, an outdoor operation area, a cargo area, a hoisting area and an indoor office area of a railway station and is used for acquiring first-class image information of the arranged areas; the second-class acquisition module is arranged in a cargo area and a hoisting area and is used for acquiring second-class image information of the arranged areas; the three types of acquisition modules are arranged at a high point capable of overlooking a railway station and are used for acquiring panoramic coverage image information of the station; and the data processing layer is used for carrying out processing and item point identification on the acquired image information. Therefore, intelligent early warning and monitoring of vehicles, personnel and goods in the station are realized, and safety and stability of production and transportation order in the station are guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of railway station security management, and in particular to a station intelligent security system based on video detection. Background Art

[0002] The construction of intelligent railways is a trend in the development of railway technology in the world today. Railway freight stations, as an important part of railway construction, are an indispensable part of railway transportation.

[0003] Railway freight terminals are vast, handling a wide variety of cargo, undergoing frequent loading and unloading operations, and experiencing the constant flow of personnel and vehicles, creating significant challenges for security management. While most railway terminals in my country have deployed varying numbers of cameras, traditional terminal video surveillance systems are typically used solely for general security purposes. Due to the large size of the terminals and the numerous camera locations, assigning dedicated personnel to review surveillance footage in real time is virtually impossible. Consequently, internal terminal surveillance footage can only be used for post-event inquiries, failing to provide real-time early warning capabilities. Consequently, problems such as cargo loss, casualties, and various operational violations continue to occur within the terminals. Therefore, it is crucial and urgent to introduce mature intelligent video surveillance technology to enable intelligent early warning monitoring of vehicles, personnel, and cargo within the terminals, thereby ensuring the safety and stability of production and transportation within the terminals. Summary of the Invention

[0004] In response to the above problems, the present invention proposes a terminal intelligent safety system based on video detection, so as to realize intelligent early warning monitoring of vehicles, personnel and goods in the station, and ensure the safety and stability of production and transportation order in the station.

[0005] The intelligent security system for the station based on video detection includes: an acquisition layer, a response execution layer, a data storage layer, a data processing layer and a command center layer.

[0006] Among them, the acquisition layer includes Class I acquisition modules, Class II acquisition modules and Class III acquisition modules; Class I acquisition modules are set in the platform area, outdoor operation area, cargo area, hoisting area and indoor office area of ​​the railway station, and are used to collect Class I image information of the set area; Class II acquisition modules are set in the cargo area and hoisting area, and are used to collect Class II image information of the set area; Class III acquisition modules are set at a high point overlooking the railway station, and are used to collect panoramic coverage image information of the station; Class II image information refers to image information used to identify perimeter intrusion points; Class I image information refers to image information other than Class II image information and panoramic coverage image information.

[0007] The response execution layer includes personnel response devices and equipment response devices. The personnel response devices are installed on the personnel in the station, and the equipment response devices are installed on the mobile devices in the station; the mobile devices include forklifts, transport vehicles and cranes; the personnel response devices and equipment response devices both include early warning elements, ID positioning elements, audio transceiver elements and control elements; the ID positioning elements upload the identity and location information of the personnel or mobile devices to the data storage layer in real time.

[0008] The data storage layer is set up in the computer room and is used to store the image information collected by the acquisition layer, the sample data required for image recognition, and the identity and location information of personnel and mobile devices in the station.

[0009] The data processing layer is set up in the computer room and is used to process the collected image information and identify items, and determine the item point identification results; the data processing layer identifies the items of the second type of image information collected by the second type of acquisition module, including: unauthorized perimeter intrusion identification in the cargo area, unauthorized perimeter intrusion identification in key cargo locations within the cargo area, and perimeter intrusion identification in the dangerous area of ​​crane operation.

[0010] The command center layer is used to obtain the collected image information and item point recognition results, and perform early warning and intelligent processing based on the item point recognition results.

[0011] In one possible implementation, the second-category acquisition module includes a dual-purpose camera device, an intelligent projection device, and a voice device; the dual-purpose camera device has a built-in standard camera element and a special camera element; the second-category acquisition module uses the standard camera element of the dual-purpose camera device in the normal working mode to collect two types of image information including the perimeter markings of the cargo area, key cargo locations within the cargo area, and the crane operation danger zone in real time; when the command center layer obtains the second-category image information including the perimeter markings, and determines that the current environment is in a harsh environment based on the second-category image information including the perimeter markings, it issues a first acquisition instruction to the second-category acquisition module to enable the second-category acquisition module to enter the harsh environment working mode; the harsh environment working mode includes the intelligent projection device projecting an optical warning line on the perimeter of the cargo area, key cargo locations within the cargo area, and the crane operation danger zone, the optical warning line being a light band with a characteristic stripe pattern; the width w of the light band is in the range of 5 cm to 30 cm; and the special camera element is used to collect the two types of image information including the optical warning line in real time.

[0012] In one possible implementation, the characteristic stripe pattern includes more than five stripes. When the optical warning line is projected onto a plane, the stripes are straight stripes, and the direction of the stripes is consistent with the direction of the light band. The widths of the multiple stripes are the same or gradually change, and the spacing between adjacent stripes is the same or gradually changes.

[0013] In one possible implementation, the intelligent projection device includes a projector whose light source is a modulated light source. A dedicated imaging element includes optical components specifically adapted to the wavelength and frequency of the modulated light. This allows the dedicated imaging element to capture the optical warning line formed by the light source of the intelligent projection device.

[0014] In one possible implementation, in a harsh environment working mode, the second-category acquisition module also uses a standard camera element to periodically collect two types of image information including perimeter markings of the cargo area, key cargo locations within the cargo area, and crane operation danger zones, and sends the periodically collected two types of image information including perimeter markings of the cargo area, key cargo locations within the cargo area, and crane operation danger zones to the command center layer.

[0015] In a possible implementation, the normal working mode further includes: calibrating the position of the intelligent projection device so that the optical warning line projected by the intelligent projection device under normal conditions coincides with the perimeter marker.

[0016] In one possible implementation, the intelligent projection device also includes a posture adjustment mechanism and a projection controller. The process of calibrating the position of the intelligent projection device includes: simultaneously activating the standard camera element and the dedicated camera element of the dual-purpose camera device; activating the intelligent projection device; the projection controller controls the posture adjustment mechanism to set the initial projection angle of the projector based on the position of the perimeter marker in the second type of image information captured by the standard camera element; the projection controller superimposes the two types of image information captured by the dedicated camera element and the standard camera element to generate a superimposed image. When it is found that the overlap between the optical warning line and the perimeter marker in the superimposed image does not reach the set value and there is a deviation, the projection controller controls the posture adjustment mechanism to continuously adjust the projection angle of the projector based on the size of the deviation until the overlap between the optical warning line and the perimeter marker in the superimposed image reaches the set value. The set value is not less than 99%. The position calibration of the optical warning line of the second type acquisition module should be performed at least once a day.

[0017] The projection controller can control the projection focal length of the projector light source. The normal working mode also includes: the projection controller adjusts the focal length of the projector light source so that the characteristic stripe pattern within the optical warning line projected by the intelligent projection device under normal conditions is flat and distortion-free in the initial state.

[0018] In one possible implementation, for the second type of image information including the optical warning line sent in real time by the second type of acquisition module, the data processing layer obtains the fluctuation change information of each part of the optical warning line by calculating and analyzing the degree of distortion of the characteristic stripe pattern of the optical warning line in the second type of image information, and identifies whether there is a perimeter intrusion based on the fluctuation change information of each part of the optical warning line.

[0019] In a possible implementation, for the second type of image information including the perimeter marker sent in real time by the second type of acquisition module, the data processing layer identifies whether there is a perimeter intrusion based on the second type of image information including the perimeter marker.

[0020] In a possible implementation, the image information refers to video information.

[0021] In one possible implementation, the data processing layer identifies whether there is a perimeter intrusion, including identifying the presence of a perimeter intrusion at the location when the fluctuation difference between a certain location and an adjacent location satisfies at least one of two conditions; the two conditions are: the fluctuation difference △h1 between the location and the adjacent location of the optical warning line at time t1 and the fluctuation difference △h2 between the location and the adjacent location at time t2 satisfy │△h2-△h1│ / (t2-t1)≥5 cm / second, where t2≥t1 and w / k1≤(t2-t1)≤w / k2; and the fluctuation difference △h1 between the location and the adjacent location of the optical warning line at time t1, the fluctuation difference △h2 between the location and the adjacent location at time t2, and the fluctuation difference △h3 between the location and the adjacent location at time t3 satisfy (│△h3-△h2│ / (t3-t2)) / (│△h2-△h1│ / (t2-t1))0.5≥k3, where t3≥t2≥t1, and w / k1≤(t2-t1)=(t3-t2)≤w / k2; k1 is the first state setting value, and the k1 value is in the range of 5-12; k2 is the second state setting value, and the k2 value is in the range of 0.1-5; k3 is the trend setting value, and the k3 value is in the range of 1.1-1.5; the frequency s of the dedicated camera element for collecting the second type of image information is 1 / (w / k1) Hz.

[0022] In one possible implementation, for the second type of image information including the optical warning line sent in real time by the second type of acquisition module, the data processing layer identifies whether there is a perimeter intrusion, including: when the fluctuation difference △h between a certain part of the optical warning line and the adjacent part at a certain moment is ≥10 cm, then it is identified that there is a perimeter intrusion at that part.

[0023] In one possible implementation, for the second type of image information including the optical warning line sent in real time by the second type of acquisition module, the data processing layer identifies whether there is a perimeter intrusion, including: when the height value h1 of a certain part of the optical warning line at time t1 and the height value h2 of the same part at time t2 satisfy │h1-h2│≥10 cm, then it is identified that there is a perimeter intrusion at this part.

[0024] In one possible implementation, when the data processing layer identifies that there is a perimeter intrusion in a crane operation danger zone, the point identification result is confirmed as the presence of a perimeter intrusion in the crane operation danger zone; when the data processing layer identifies that there is a perimeter intrusion in the cargo area or a key cargo location within the cargo area, the identity and location information of the personnel or mobile device in the area where the perimeter intrusion is located is immediately retrieved; if the location information of the personnel or mobile device is consistent with the perimeter intrusion, and the identity information has the access authorization for the cargo area or the key cargo location, the point identification result is determined to be that there is no unauthorized perimeter intrusion in the cargo area or the key cargo location; if the location information of the personnel or mobile device is consistent with the perimeter intrusion, but the identity information does not have the access authorization for the cargo area or the key cargo location, or there is no personnel or mobile device in the terminal that is consistent with the perimeter intrusion, then the point identification result is confirmed as the presence of an unauthorized perimeter intrusion in the cargo area or the key cargo location.

[0025] In one implementation, when the result of the point identification is one or more of: unauthorized perimeter intrusion in a cargo area, unauthorized perimeter intrusion in a key cargo location within a cargo area, and perimeter intrusion in a crane operation danger zone, and the location information of the personnel or mobile device is consistent with the perimeter intrusion, the command center layer performs early warning and intelligent processing based on the point identification result, including: the command center layer sends a processing instruction to the second-class acquisition module and the personnel response device of the person involved; the voice device of the second-class acquisition module reads the identity information of the person involved or the mobile device through voice according to the processing instruction and warns and drives away; at the same time, the control element of the personnel response device activates the vibration, lighting, and alarm functions of the early warning element according to the processing instruction, and the audio transceiver element plays the voice command independently generated by the intelligent processing of the command center layer to the person involved, and feeds back the voice of the person involved to the command center layer.

[0026] In one possible implementation, when the result of the point identification is one or more of: unauthorized perimeter intrusion in a cargo area, unauthorized perimeter intrusion in a key cargo location within a cargo area, and perimeter intrusion in a crane operation danger zone, and there are no personnel or mobile equipment in the terminal that match the perimeter intrusion, the command center layer performs early warning and intelligent processing based on the identification point result, including: the command center layer sends a processing instruction to the second-class acquisition module; the voice device of the second-class acquisition module issues a stern warning and drives away through voice according to the processing instruction, and at the same time, staff are dispatched from the terminal center or deployed from nearby to immediately rush to the area for investigation.

[0027] In one possible implementation, when the Class II acquisition module is in a harsh environment working mode, the warning and intelligent processing performed by the command center layer based on the identification results also include: instructing the intelligent projection device to project a red or red and green flashing optical warning line, which serves as a warning and deterrent to the parties involved and as a guide for staff who come to investigate.

[0028] In a possible implementation, the device response device further includes an execution element, which is connected to a power device of the mobile device and can intervene in the power device of the mobile device.

[0029] In one possible implementation, when there is an unauthorized perimeter intrusion in a cargo area or a key cargo location, and the location information of a person or mobile device is consistent with the perimeter intrusion, the warning and intelligent processing performed by the command center layer based on the item identification item also includes: the command center layer sends a processing instruction to the device response device of the mobile device, and the control element of the device response device controls the execution element to shut down the power output of the mobile device in question according to the processing instruction. Then, the command center layer communicates with the person in question through the audio transceiver element of the person in question or the mobile device in question. After confirmation and permission by the command center layer, the control element of the device response device controls the execution element to restore the power output of the mobile device in question according to the processing instruction, allowing the mobile device in question to leave the area.

[0030] In one possible implementation, when there is a perimeter intrusion in a dangerous area where a crane operates, the command center layer sends a processing instruction to the equipment response device of the crane in the dangerous area where the crane operates, and the control element of the equipment response device controls the execution element to shut down or slow down the power output of the crane in question according to the processing instruction; after the data processing layer removes the perimeter intrusion identification item based on the second-class image information collected by the second-class acquisition module of the area, and after confirmation and permission from the command center layer, the control element of the equipment response device controls the execution element to restore the power output of the crane in question according to the processing instruction.

[0031] In one possible implementation, the identification items of a type of collection module include: one or more of: smoking identification, fire identification, identification of not wearing work clothes, identification of leaving key positions, napping identification, shirtless identification, intelligent identification of loading and unloading operations, intelligent identification of cargo area and cargo space occupancy, and intelligent identification of red card placement.

[0032] In one possible implementation, the image information collected by the three types of acquisition modules is used to be combined with the first type of image information and / or the second type of image information to verify the accuracy of the item point recognition results corresponding to the identification item points of the first type of acquisition module and / or the identification phase points of the second type of acquisition module.

[0033] Beneficial effects of the present invention: The video-detection-based intelligent terminal security system of the present invention, when the image information collection area is in a harsh environment, uses an intelligent projection device to project an optical warning line along the perimeter of the cargo area, key cargo locations within the cargo area, and the crane operation danger zone. The dual-purpose camera device then switches to a dedicated camera element to accurately capture the image information of the optical warning line. Simultaneously, the data processing layer calculates and analyzes the degree of distortion of this image information to obtain fluctuation information of each part of the optical warning line. Based on this fluctuation information, the system identifies whether there has been a perimeter intrusion. This effectively solves the technical problem of accurately collecting image information in harsh environments and overcomes the industry technical difficulty of identifying perimeter markings when buried under water, snow, and dust. This system increases the accuracy rate of perimeter intrusion identification in key areas to over 99.8%.

[0034] The intelligent safety system for the station based on video detection sets personnel response devices and equipment response devices on the personnel and mobile devices in the station respectively, and combines the linkage between the collection layer, response execution layer, data storage layer, data processing layer and command center layer. The command center layer not only issues an early warning of the perimeter intrusion, but more importantly, it can implement intelligent processing in real time, so that the system's successful handling rate of perimeter intrusion points reaches a high level of 99.5%, and it can "accurately identify" and "control" potential accident hazards, which can prevent accidents before they happen and avoid the further development of safety identification points into safety accidents, causing greater loss of life and property.

[0035] The intelligent security system for the station based on video detection obtains dynamic fluctuation data of various parts at continuous time points through video information, and adopts a special dynamic analysis and recognition method based on the fluctuation change information of each part to identify whether there is perimeter intrusion in each part of the optical warning line, avoiding misjudgment and missed judgment of perimeter intrusion points based on a single static image, overcoming the technical difficulties of perimeter marking technology under the burial of water, snow and dust, ensuring the accuracy of perimeter intrusion recognition in key areas, and thus realizing effective monitoring and early warning of key areas of the station.

[0036] In summary, the video detection-based intelligent security system for the station of the present invention can not only realize intelligent early warning monitoring of vehicles, personnel and goods in the station, but also through the linkage and collaboration between the command center layer and the collection layer, response execution layer, data storage layer, data processing layer and other layers, the identification results of perimeter intrusion can be intelligently processed in real time to ensure the safety and stability of the production and transportation order in the station.

[0037] The above general description and the following description are exemplary and explanatory only and are not intended to limit the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] One or more embodiments are exemplarily described by corresponding drawings. These exemplified descriptions and drawings do not limit the embodiments. Elements with the same reference numerals in the drawings are similar elements, and the drawings do not constitute a scale limitation. Among them: Figure 1 This is a schematic block diagram of the composition of a site intelligent safety system based on video detection provided in an embodiment of the present application.

[0039] Figure 2 This is a schematic block diagram of the composition of a second-type acquisition module provided in an embodiment of the present application.

[0040] Figure 3 This is a schematic diagram of the positional relationship between a perimeter marker and an optical warning line provided in an embodiment of the present application.

[0041] Reference numerals: 1, acquisition layer; 11, first-class acquisition module; 12, second-class acquisition module; 121, dual-purpose camera device; 1211, standard camera element; 1212, dedicated camera element; 122, intelligent projection device; 123, voice device; 13, third-class acquisition module; 2, response execution layer; 21, personnel response device; 22, equipment response device; 3. Data storage layer; 4. Data processing layer; 5. Command center level; 6. Perimeter marking; 7. Optical warning line. DETAILED DESCRIPTION

[0042] The following further describes specific embodiments of the present invention with reference to the accompanying drawings. The accompanying drawings are for illustrative purposes only and are not intended to limit the present invention. In the following technical description, for ease of explanation, numerous details are provided to provide a full understanding of the present invention. However, the present invention can still be practiced without these details. In other cases, well-known structures and devices may be simplified for simplicity of illustration.

[0043] The terms "first," "second," and the like in the specification and claims of the present invention and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a particular order or sequence. It should be understood that the terms used in this manner are interchangeable where appropriate to facilitate the description of the embodiments of the present invention.

[0044] In the present invention, the directions or positional relationships indicated by terms such as "upper", "lower", "inner", "middle", "outer", "front", and "back" are based on the directions or positional relationships shown in the accompanying drawings. These terms are mainly used to better describe the present invention, and are not used to limit the indicated devices, elements, or components to have a specific direction, or to be constructed and operated in a specific direction. Moreover, in addition to being used to indicate directions or positional relationships, some of the above terms may also be used to indicate other meanings. For example, the term "upper" may also be used to indicate a certain dependency or connection relationship in certain circumstances. For those of ordinary skill in the art, the specific meanings of these terms in the present invention can be understood according to specific circumstances.

[0045] Furthermore, the terms "disposed," "connected," and "fixed" should be interpreted broadly. For example, "connected" can mean a fixed connection, a removable connection, or an integral structure; it can be a mechanical connection or an electrical connection; it can be a direct connection, an indirect connection through an intermediary, or an internal communication between two devices, elements, or components. Those skilled in the art will understand the specific meanings of these terms in the present invention based on the specific circumstances.

[0046] In a station scenario (such as a railway station), it usually includes platform areas, outdoor work areas, cargo areas, lifting areas, and indoor office areas. These areas need to be monitored to identify the safety status of each area, thereby improving the safety of station management.

[0047] like Figure 1 As shown, the embodiment of the present application provides a video detection-based intelligent security system for a terminal. The video detection-based intelligent security system for a terminal includes an acquisition layer 1, a response execution layer 2, a data storage layer 3, a data processing layer 4, and a command center layer 5.

[0048] The acquisition layer 1 is used to acquire image information within the station and includes a first-class acquisition module 11 , a second-class acquisition module 12 , and a third-class acquisition module 13 .

[0049] The second-category acquisition module is installed in the cargo area and the lifting area to collect second-category image information in the installed area. Second-category image information refers to image information used to identify perimeter intrusion points. Perimeter intrusion can be understood as the unauthorized intrusion of foreign objects (such as personnel or mobile equipment) into the truck area, the perimeter of key cargo locations within the cargo area, and the perimeter of the crane operation danger zone. Perimeter intrusion points (i.e., points for identifying the second-category image information collected by the second-category acquisition module, or identification points of the second-category acquisition module) include: identification of unauthorized perimeter intrusions into the cargo area, identification of unauthorized perimeter intrusions into key cargo locations within the cargo area, and identification of perimeter intrusions into the crane operation danger zone.

[0050] Class I acquisition modules are installed in the platform area, outdoor operation area, cargo area, hoisting area, and indoor office area of ​​a railway station to collect Class I image information of the installed area. Class I image information refers to image information other than Class II image information and panoramic coverage image information. The identification items of the Class I acquisition modules can be understood as the items for identifying the Class I image information collected by the Class I acquisition modules. Exemplarily, the identification items of the Class I acquisition modules include: smoking identification, fire identification, identification of not wearing work clothes, identification of leaving key positions, dozing identification, shirtless identification, intelligent identification of loading and unloading operations, intelligent identification of cargo area and cargo space occupancy, and intelligent identification of red card placement.

[0051] The third-type acquisition module is located at a high point overlooking the railway station and is used to collect panoramic image information of the station. Exemplarily, the image information collected by the third-type acquisition module (i.e., panoramic image information) is combined with the image information collected by the first-type acquisition module and / or the second-type acquisition module to verify the accuracy of the item points identified by the first-type acquisition module and / or the phase points identified by the second-type acquisition module.

[0052] The acquisition layer sends the collected image information to the data processing layer and the data storage layer.

[0053] The data processing layer is located in the computer room and is used to process the collected image information and identify items, and determine the item point identification results. For example, the first-category image information collected by the first-category acquisition module is processed and item points are identified to obtain the item point identification results of the corresponding identification items of the first-category acquisition module; the second-category image information collected by the second-category acquisition module is processed and item points are identified to obtain the item point identification results of the corresponding identification items of the second-category acquisition module; the panoramic coverage image information collected by the third-category acquisition module is processed so that the panoramic coverage image information is marked with the identification items of the first-category acquisition module and / or the identification items of the second-category acquisition module, and the image area corresponding to the identification item point can be magnified and viewed to verify the accuracy of the item point identification results of the identification item point from a full field of view. In addition, the data processing layer is also used to send the item point identification results to the command center layer.

[0054] The command center layer is used to obtain the item point identification results sent by the data processing layer, and perform early warning and intelligent processing based on the item point identification results.

[0055] In some embodiments, the command center layer sends an alert message to the response execution layer based on the item point identification result for early warning, and instructs the response execution layer to perform intelligent processing.

[0056] Combine Figure 1As shown, response execution layer 2 includes a human response device 21 and a device response device 22. Human response device 21 is installed on personnel within the station, while device response device 22 is installed on mobile devices within the station. Both human response device 21 and device response device 22 include a warning element, an ID positioning element, an audio transceiver element, and a control element. The ID positioning element uploads the identity and location information of the person or mobile device to the data storage layer 3 in real time.

[0057] Exemplary mobile equipment includes forklifts, transport trucks, and cranes.

[0058] The data storage layer is set up in the computer room and is used to store the image information collected by the acquisition layer, the sample data required for image recognition, and the identity and location information of personnel and mobile devices in the station.

[0059] like Figure 2 As shown, the embodiment of the present application provides a second type of acquisition module 12. The second type of acquisition module 12 includes a dual-purpose camera device 121, an intelligent projection device 122 and a voice device 123.

[0060] In one implementation, combining Figure 2 As shown, the dual-purpose imaging device 121 includes a standard imaging element 1211 and a dedicated imaging element 1212 built therein.

[0061] In normal working mode, the second-category acquisition module uses standard camera elements to collect two types of image information including the perimeter markings of the cargo area, key cargo locations within the cargo area, and crane operation danger zones in real time, and sends the collected second-category image information including perimeter markings to the command center layer and data processing layer in real time.

[0062] Perimeter markings are used to identify the perimeters of the cargo area, key cargo locations within the cargo area, and crane operation danger zones in normal operation mode. Optionally, the perimeter markings are physical, for example, the perimeter markings can be obtained by one or more methods such as drawing lines, spray painting, and laying tape.

[0063] In the harsh environment working mode, the second-category acquisition module uses a dedicated camera element to collect two-category image information including the cargo area, key cargo locations within the cargo area, and optical warning lines in the crane operation danger zone in real time, and sends the real-time collected second-category image information including the optical warning lines to the data processing layer in real time.

[0064] Optical warning lines are used to identify the perimeter of the cargo area, key cargo locations within the cargo area, and the crane operation danger zone in harsh environment operation mode. Optionally, the optical warning lines are optical projections, specifically projected by an intelligent projection device onto the perimeter of the cargo area, key cargo locations within the cargo area, and the crane operation danger zone.

[0065] Exemplarily, the optical warning line is a light stripe with a characteristic stripe pattern; the width w of the light stripe is in the range of 5 cm to 30 cm.

[0066] Exemplarily, the characteristic stripe pattern includes more than 5 stripes. When the optical warning line is projected onto a plane, the stripes are straight stripes, and the direction of the stripes is consistent with the direction of the light band. The widths of the multiple stripes are the same or the widths change gradually, and the spacing between adjacent stripes is the same or changes gradually.

[0067] Optionally, in the harsh environment working mode, before the Class II acquisition module uses a dedicated camera element to collect Class II image information including optical warning lines of the cargo area, key cargo locations within the cargo area, and crane operation danger zones in real time, the Class II acquisition module can also use an intelligent projection device to project optical warning lines on the perimeter of the cargo area, key cargo locations within the cargo area, and crane operation danger zones.

[0068] Exemplarily, the intelligent projection device includes a projector whose light source is a modulated light source. The dedicated imaging element includes optical components specifically adapted to the wavelength and frequency of the modulated light. Thus, the dedicated imaging element is capable of capturing the optical warning line formed by the light source of the intelligent projection device.

[0069] Optionally, in the harsh environment working mode, the second type of acquisition module also uses a standard camera element to periodically collect two types of image information including the perimeter markings of the cargo area, key cargo locations within the cargo area, and the crane operation danger zone, and sends the periodically collected two types of image information including the perimeter markings of the cargo area, key cargo locations within the cargo area, and the crane operation danger zone to the command center layer.

[0070] Optionally, the normal working mode further includes: calibrating the position of the intelligent projection device so that the optical warning line projected by the intelligent projection device in the normal working mode coincides with the perimeter marker.

[0071] Combine Figure 3 As shown, the embodiment of the present application provides a schematic diagram of a perimeter marker 6 and an optical warning line 7. It should be noted that Figure 3 The width difference between the perimeter marker 6 and the optical warning line 7 shown in the figure is only for the convenience of marking and does not represent the actual width difference. In the normal working mode, the optical warning line 6 projected by the intelligent projection device coincides with the perimeter marker 7.

[0072] Exemplarily, the intelligent projection device also includes a posture adjustment mechanism and a projection controller. The process of calibrating the position of the intelligent projection device includes: simultaneously starting the standard camera element and the special camera element of the dual-purpose camera device; starting the intelligent projection device; the projection controller controls the posture adjustment mechanism to set the initial projection angle of the projector based on the position of the perimeter marker in the second type of image information collected by the standard camera element; the projection controller superimposes the two types of image information collected by the special camera element and the standard camera element to generate a superimposed image. When it is found that the overlap between the optical warning line and the perimeter marker in the superimposed image does not reach the set value and there is a deviation, the projection controller will control the posture adjustment mechanism to continuously adjust the projection angle of the projector according to the size of the deviation until the overlap between the optical warning line and the perimeter marker in the superimposed image reaches the set value. The set value is not less than 99%. The position calibration of the optical warning line of the second type acquisition module should be performed at least once a day.

[0073] In one implementation, the Class II acquisition module receives an acquisition instruction from the command center layer. The acquisition instruction includes a first acquisition instruction and a second acquisition instruction. The first acquisition instruction instructs the Class II acquisition module to switch to a harsh environment operating mode, and the second acquisition instruction instructs the Class II acquisition module to switch to a normal operating environment. In response to the acquisition instruction, the Class II acquisition module switches its operating mode, which includes the aforementioned normal operating mode and harsh environment operating mode.

[0074] In one implementation, the command center determines the current environment based on the second-class image information captured by the standard camera element and received from the second-class acquisition module. If the current environment is severe, a first acquisition instruction is sent to the second-class acquisition module; if the current environment is normal, a second acquisition instruction is sent to the second-class acquisition module.

[0075] For example, if the second-category image information captured by the standard camera element is in one or more of the following situations, the command center layer determines that the current environment is a severe environment: the clarity of the second-category image information is less than or equal to a preset clarity (the preset clarity can be set according to the user's actual needs); the acquisition area (cargo area or lifting area) in the second-category image information is in at least one of rain, snow, and windy sand; the perimeter marker in the second-category image information is buried by one or more of accumulated water, snow, and dust. Except for the above situations, the command center layer determines that the current environment is normal.

[0076] In one implementation, the Class II acquisition module defaults to normal operating mode. Specifically, it uses standard imaging elements to capture real-time Class II image information, including perimeter markings of the cargo area, key cargo locations within the cargo area, and crane operation danger zones. This Class II image information, including the perimeter markings, is then transmitted to the command center in real time. The command center receives the Class II image information captured in real time by the standard imaging elements and, based on this information, determines whether the current environment is a harsh environment. If the current environment is a harsh environment, the command center sends a first acquisition instruction to the Class II acquisition module. This causes the Class II acquisition module to switch to a harsh environment mode. In this mode, the Class II acquisition module periodically transmits the Class II image information, including the perimeter markings, to the command center. Based on this information, the command center periodically determines whether the current environment is normal. If the current environment is normal, it sends a second acquisition instruction to the Class II acquisition module, causing it to switch to a normal operating mode.

[0077] In some embodiments, a type of acquisition module includes a high-definition camera, such as a network camera and a spherical camera.

[0078] In some embodiments, the third type of acquisition module includes a high-point camera.

[0079] In some embodiments, for the second type of image information including the optical warning line sent in real time by the second type of acquisition module, the data processing layer obtains the fluctuation change information of each part of the optical warning line by calculating and analyzing the distortion degree of the characteristic stripe pattern of the optical warning line in the second type of image information, and identifies whether there is a perimeter intrusion based on the fluctuation change information of each part of the optical warning line.

[0080] During the development process, the applicant discovered that when the perimeter warning signs in key areas such as the cargo area in the station or key cargo positions in the cargo area, as well as the crane operation danger zone in the hoisting area, are buried in water, snow, or dust, the surface contour height of the area will rise as a whole, or there will be fluctuations. Therefore, when identifying intrusions in the perimeter in this state, if the static state of the height or fluctuation of each part of the optical warning line is obtained based on a single static image calculation and analysis, the recognition accuracy of the item will be reduced, and it is easy to cause misjudgment and missed judgment. The present application obtains dynamic fluctuation data of each part at continuous time points through video information, and uses a special dynamic analysis and recognition method based on the fluctuation change information of each part to identify whether there is perimeter intrusion in each part of the optical warning line, avoiding misjudgment and missed judgment based on a single static image to identify perimeter intrusion points, overcoming the technical difficulties of perimeter marking technology under water, snow, and dust, and ensuring the accuracy of perimeter intrusion recognition in key areas.

[0081] In one implementation, the above-mentioned dynamic analysis and identification method includes: Image information refers to video information; for the second-category image information including the optical warning line sent in real time by the second-category acquisition module, the data processing layer identifies whether there is a perimeter intrusion, including: when the difference in undulation between a certain part and an adjacent part meets at least one of the two conditions, then the part is identified as having a perimeter intrusion.

[0082] The two situations described are: The fluctuation difference △h1 between the optical warning line at time t1 and the adjacent part and the fluctuation difference △h2 between the optical warning line at time t2 and the adjacent part satisfy │△h2-△h1│ / (t2-t1)≥5cm / s, where t2≥t1 and w / k1≤(t2-t1)≤w / k2. And, The fluctuation difference △h1 between the optical warning line at time t1 and the adjacent part, the fluctuation difference △h2 between the optical warning line at time t2 and the fluctuation difference △h3 between the optical warning line at time t3 satisfy (│△h3-△h2│ / (t3-t2)) / (│△h2-△h1│ / (t2-t1)) 0.5 ≥k3, where t3 ≥ t2 ≥ t1, and w / k1 ≤ (t2 - t1) = (t3 - t2) ≤ w / k2; The frequency s of the dedicated camera element for collecting the second type of image information is 1 / (w / k1) Hz; k1 is the first state setting value, and the k1 value is in the range of 5-12; k2 is the second state setting value, and the k2 value is in the range of 0.1-5; k3 is the trend setting value, and the k3 value is in the range of 1.1-1.5.

[0083] This dynamic recognition method is to comprehensively judge the dynamic fluctuation change information and change trend of each part by calculating and analyzing the fluctuation difference change speed information between a certain part and the surrounding adjacent parts at continuous time points, and use this as a basis to identify whether there is perimeter intrusion at each part of the optical warning line, achieving an accuracy rate of more than 99.8% for perimeter intrusion recognition in key areas.

[0084] In another implementation, for the second type of image information including the optical warning line sent in real time by the second type of acquisition module, the data processing layer identifies whether there is a perimeter intrusion, including: when the fluctuation difference △h between a certain part of the optical warning line and the adjacent part at a certain moment is ≥10 cm, then it is identified that there is a perimeter intrusion at that part.

[0085] For example, the portion immediately adjacent to a portion A on the optical warning line may refer to any portion among multiple portions immediately adjacent to the portion A.

[0086] Exemplarily, a certain portion A and an adjacent portion each include at least a portion of a stripe.

[0087] In another implementation, for the second type of image information including the optical warning line sent in real time by the second type of acquisition module, the data processing layer identifies whether there is a perimeter intrusion, including: when the height value h1 of a certain part of the optical warning line at time t1 and the height value h2 of the same part at time t2 satisfy │h1-h2│≥10 cm, then it is identified that there is a perimeter intrusion at this part.

[0088] In other embodiments, for the second type of image information including the perimeter marker collected in real time by the second type of acquisition module, the data processing layer identifies whether there is a perimeter intrusion based on the second type of image information including the perimeter marker.

[0089] Optionally, the sample data required for image recognition includes historical second-category image information captured by a standard camera element.

[0090] In one implementation, the data processing layer inputs the second-category image information, including perimeter markings, collected in real time from the second-category acquisition module into a recognition model, generating identification points for the second-category acquisition module output by the recognition model. The recognition model is trained based on the aforementioned historical second-category image information and is used to identify whether the second-category image information, including perimeter markings, represents perimeter intrusions, namely, one or more of the following: perimeter intrusions in the cargo area, perimeter intrusions in key cargo locations within the cargo area, or perimeter intrusions in the crane operation danger zone.

[0091] In some embodiments, when the data processing layer identifies that there is a perimeter intrusion, the data processing layer determines an item point identification result based on the perimeter intrusion situation.

[0092] In one implementation, when the data processing layer identifies that there is a perimeter intrusion in the crane operation danger zone, the item point identification result is confirmed as the existence of a perimeter intrusion in the crane operation danger zone.

[0093] In one implementation, when the data processing layer identifies a perimeter intrusion in a cargo area or a key cargo location within the cargo area, the identity and location information of the person or mobile device in the area where the perimeter intrusion is located is immediately retrieved. If the location information of the person or mobile device is consistent with the perimeter intrusion, and the identity information has access authorization for the cargo area or key cargo location, the point identification result is determined to be that there is no unauthorized perimeter intrusion in the cargo area or key cargo location. If the location information of the person or mobile device is consistent with the perimeter intrusion, but the identity information does not have access authorization for the cargo area or key cargo location, or if there are no people or mobile devices in the terminal that are consistent with the perimeter intrusion, the point identification result is confirmed to be that there is an unauthorized perimeter intrusion in the cargo area or key cargo location.

[0094] The area where the portion with the perimeter intrusion exists may represent a circular area with the portion with the perimeter intrusion existing as the center and a radius of any distance between 3m and 5m.

[0095] Whether the location information of a person or mobile device is consistent with the perimeter intrusion can be understood as whether there is location information of a person or mobile device in the area where the perimeter intrusion is located. If so, it is determined that the location information of the person or mobile device is consistent with the perimeter intrusion. It is understandable that in this case, it can be considered that there are people or mobile devices whose location information can be located in the area where the perimeter intrusion is located, but whether the person or mobile device is authorized requires further judgment. For example, the ID of the person or mobile device at the location can be obtained through the location information, and whether the person or mobile device is authorized can be determined based on the ID. If not, it is determined that the location information of the person or mobile device is inconsistent with the perimeter intrusion. It is understandable that in this case, it can be considered that there are no people or mobile devices whose location information can be located in the area where the perimeter is located. At this time, it can be considered that the intruder is an unauthorized foreign object, such as a person or mobile device that has not registered to enter the station, or an animal.

[0096] In some embodiments, when the result of item identification is one or more of the following: unauthorized perimeter intrusion in a cargo area, unauthorized perimeter intrusion in a key cargo location within a cargo area, and perimeter intrusion in a crane operation danger zone, the data processing layer sends information on whether the location information of the personnel or mobile device is consistent with the perimeter intrusion to the command center layer.

[0097] In some embodiments, when the command center layer receives information on whether the location information of a person or mobile device is consistent with the perimeter intrusion, it performs early warning and intelligent processing on the item identification results (i.e., one or more of the existence of unauthorized perimeter intrusion in a cargo area, unauthorized perimeter intrusion in a key cargo location within a cargo area, and perimeter intrusion in a dangerous area for crane operations) based on whether the location information of the person or mobile device is consistent with the perimeter intrusion.

[0098] In one implementation, when the location information of a person or mobile device matches the perimeter intrusion, the command center layer performs an early warning and intelligent processing based on the identification result. This includes: the command center layer sends a processing instruction to the second-class acquisition module and the personnel response device of the person involved. The voice device of the second-class acquisition module reads the identity information of the person or mobile device through voice according to the processing instruction and warns and drives it away. Simultaneously, the control element of the personnel response device activates the vibration, lighting, and alarm functions of the early warning element according to the processing instruction. The audio transceiver element plays the voice command generated by the intelligent processing of the command center layer to the person involved, and then feeds the person's voice back to the command center layer.

[0099] In another implementation, when the location of a person or mobile device doesn't match the perimeter intrusion, the command center performs an alert and intelligent processing based on the identification results. This includes sending a processing instruction to the second-class acquisition module. The second-class acquisition module's voice device issues a stern warning and urges the person to leave. Simultaneously, personnel are dispatched from the station center or deployed from nearby locations to immediately reconnoiter the area.

[0100] Optionally, when the result of the item identification is one or more of: unauthorized perimeter intrusion in a cargo area, unauthorized perimeter intrusion in a key cargo location within a cargo area, and perimeter intrusion in a crane operation danger zone, when the second-class acquisition module is in the harsh environment working mode, the early warning and intelligent processing performed by the command center layer based on the identification item result also includes: instructing the intelligent projection device to project a red or red and green flashing optical warning line, which serves as a warning and deterrent to the parties involved and as a guide sign for the staff who come to investigate.

[0101] Optionally, when the result of the item recognition is: there is no unauthorized perimeter intrusion in the cargo area, there is no unauthorized perimeter intrusion in the key cargo locations within the cargo area, and there is no perimeter intrusion in the crane operation danger zone, the command center instructs the intelligent projection device to project a green optical warning line.

[0102] In some embodiments, the device response device further includes an execution element, which is connected to the power device of the mobile device and can intervene in the power device of the mobile device.

[0103] In one implementation, when there is an unauthorized perimeter intrusion in a cargo area or a key cargo location, and the location information of a person or mobile device is consistent with the perimeter intrusion, the warning and intelligent processing performed by the command center layer based on the item identification item also includes: the command center layer sends a processing instruction to the device response device of the mobile device, and the control element of the device response device controls the execution element to shut down the power output of the mobile device in question according to the processing instruction. Then, the command center layer communicates with the person in question through the audio transceiver element of the person in question or the mobile device in question. After confirmation and permission by the command center layer, the control element of the device response device controls the execution element to restore the power output of the mobile device in question according to the processing instruction, allowing the mobile device in question to leave the area.

[0104] In one implementation, when there is a perimeter intrusion in a crane operation danger zone, the command center layer sends a processing instruction to the equipment response device of the crane in the crane operation danger zone, and the control element of the equipment response device controls the execution element to shut down or slow down the power output of the crane in question (i.e., the crane in the crane operation danger zone) according to the processing instruction; after the data processing layer removes the perimeter intrusion identification item based on the second-class image information collected by the second-class acquisition module of the area, after confirmation and permission from the command center layer, the control element of the equipment response device controls the execution element to restore the power output of the crane in question according to the processing instruction.

[0105] Optionally, the sample data required for image recognition includes a type of historical image information.

[0106] In some embodiments, the data processing layer processes and identifies item points in the first-class image information collected by the first-class acquisition module, and determines the item point recognition results. This includes inputting the first-class image information into a second recognition model and obtaining item point recognition results corresponding to the identification item points of the first-class acquisition module output by the second recognition model, such as a person smoking or a fire in a certain area. The second recognition model is trained using historical first-class image information and is used to identify the identification item points of the first-class acquisition module and obtain corresponding item point recognition results.

[0107] In some embodiments, the command center layer can view data stored by the data storage layer.

[0108] After investigation, the applicant found that weak management of areas such as cargo areas and lifting areas within the terminal is often a major factor leading to problems such as cargo loss, damage, and casualties. For example, after registering to enter the terminal, a person or vehicle may intentionally or unintentionally enter the wrong cargo area or cargo location due to unfamiliarity with the route or ill intentions, which can easily lead to cargo loss; some people or vehicles may mistakenly enter the lifting area, which can easily cause safety accidents such as casualties; and workers and vehicles within the terminal sometimes lack sensitivity to key cargo locations and invade key cargo locations while loading and unloading other cargo locations, affecting the safe storage of high-value, dangerous goods, and other key cargo locations in key cargo locations. Perimeter intrusion monitoring and timely handling of areas such as cargo areas or key cargo locations within cargo areas, as well as crane operation danger zones in the lifting area, can effectively reduce the incidence of the above problems.

[0109] During the development of the intelligent safety system of the present invention, the applicant discovered that, although perimeter warning signs are laid in the aforementioned key areas within the terminal, which can serve as a basis for video recognition detection, when the image information collection area is in a harsh environment such as rain, snow, and sandstorms, on the one hand, the image information is easily blurred, and on the other hand, rain, snow, sandstorms, etc. can easily bury and cover the perimeter warning signs. These factors can adversely affect the effective identification of perimeter intrusions, reduce the recognition accuracy of this point, and result in false positives and missed judgments. In the terminal intelligent safety system of the present invention, when the image information collection area is in a harsh environment, an intelligent projection device projects an optical warning line on the perimeter of the cargo area, key cargo locations within the cargo area, and the crane operation danger zone, and the dual-purpose camera device switches to a dedicated camera element to accurately capture the optical warning line image information. At the same time, the data processing layer calculates and analyzes the degree of distortion of the image information to obtain the fluctuation change information of each part of the optical warning line, and identifies whether there is a perimeter intrusion based on the fluctuation change information of each part of the optical warning line. This method can effectively solve the technical problem of accurately collecting image information in harsh environments, overcome the industry's technical difficulties in identifying perimeter signs buried under accumulated water, snow, and dust, and increase the accuracy of perimeter intrusion identification in key areas to over 99.8%.

[0110] Furthermore, even when existing terminal video detection security systems effectively identify perimeter intrusions in key areas under good visibility, they often only provide real-time warnings to the parties involved and lack real-time response and disposal. By the time disposal personnel arrive, the offending personnel or vehicles have already escaped the cargo area or key cargo locations, compromising the safe storage of goods or causing the goods to be lost. In particular, the inability to respond and handle in real time when personnel or vehicles intrude into the perimeter of a crane operation danger zone can lead to serious casualties and serious consequences. In the intelligent terminal security system of the present invention, by providing personnel response devices and equipment response devices on personnel and mobile devices within the terminal, and integrating linkage and collaboration between the command center layer and the collection layer, response execution layer, data storage layer, and data processing layer, the command center layer not only provides early warning of perimeter intrusions but, more importantly, implements intelligent processing in real time. This system achieves a high success rate of 99.5% in handling perimeter intrusions, achieving both accurate identification and effective control of potential accidents. This prevents identified safety issues from developing into safety accidents, resulting in greater loss of life and property.

[0111] The embodiments of the present invention are not limited to the structures described above and shown in the drawings, and various modifications and changes can be made without departing from the scope thereof. The scope of the present invention is limited only by the appended claims.

Claims

1. A video detection-based intelligent security system for a station, characterized by: include: Collection layer, response execution layer, data storage layer, data processing layer and command center layer; The acquisition layer includes Class I acquisition modules, Class II acquisition modules, and Class III acquisition modules. Class I acquisition modules are set up in the platform area, outdoor operation area, cargo area, hoisting area, and indoor office area of ​​the railway station to collect Class I image information of the set area. Class II acquisition modules are set up in the cargo area and hoisting area to collect Class II image information of the set area. Class III acquisition modules are set up at high points overlooking the railway station to collect panoramic coverage image information of the station. Class II image information refers to image information used to identify perimeter intrusion points. Class I image information refers to image information other than Class II image information and panoramic coverage image information. The response execution layer includes personnel response devices and equipment response devices. Personnel response devices are installed on personnel within the station, while equipment response devices are installed on mobile devices within the station. Mobile devices include forklifts, transport vehicles, and cranes. Both personnel response devices and equipment response devices include warning components, ID positioning components, audio transceiver components, and control components. The ID positioning components upload the identity and location information of the personnel or mobile device to the data storage layer in real time. The data storage layer is set up in the computer room and is used to store the image information collected by the acquisition layer, the sample data required for image recognition, and the identity and location information of personnel and mobile devices in the station; The data processing layer is located in the computer room and is used to process the collected image information, identify items, and determine the item recognition results. The data processing layer identifies the following items for the second type of image information collected by the second type acquisition module: unauthorized perimeter intrusion recognition of the cargo area, unauthorized perimeter intrusion recognition of key cargo locations within the cargo area, and perimeter intrusion recognition of the crane operation danger zone. The command center layer is used to obtain the collected image information and item point recognition results, and perform early warning and intelligent processing based on the item point recognition results.

2. The intelligent security system for a station based on video detection according to claim 1 is characterized in that: The second-type acquisition module includes a dual-purpose camera device, an intelligent projection device and a voice device; the dual-purpose camera device has a built-in standard camera element and a special camera element; the second-type acquisition module uses the standard camera element of the dual-purpose camera device in the normal working mode to collect two types of image information including the perimeter markings of the cargo area, key cargo locations in the cargo area and the crane operation danger zone in real time; when the command center layer obtains the second-type image information including the perimeter markings and determines that the current environment is in a harsh environment based on the second-type image information including the perimeter markings, it sends a first acquisition instruction to the second-type acquisition module to make the second-type acquisition module enter the harsh environment working mode; the harsh environment working mode includes the intelligent projection device projecting an optical warning line on the perimeter of the cargo area, key cargo locations in the cargo area and the crane operation danger zone, and the optical warning line is a light band with a characteristic stripe pattern; the width w of the light band is in the range of 5 cm to 30 cm; the special camera element is used to collect the two types of image information including the optical warning line in real time.

3. The intelligent security system for a terminal based on video detection according to claim 2 is characterized in that: For the second-category image information including the optical warning line sent in real time by the second-category acquisition module, the data processing layer calculates and analyzes the degree of distortion of the characteristic stripe pattern of the optical warning line in the second-category image information to obtain the fluctuation change information of each part of the optical warning line, and then identifies whether there is a perimeter intrusion based on the fluctuation change information of each part of the optical warning line; or, With respect to the second type of image information including the perimeter identification sent in real time by the second type of acquisition module, the data processing layer identifies whether there is perimeter intrusion based on the second type of image information including the perimeter identification.

4. The intelligent security system for a terminal based on video detection according to claim 3 is characterized in that: Image information refers to video information. The data processing layer identifies whether there is a perimeter intrusion by identifying the presence of a perimeter intrusion at a location when the difference in fluctuation between a certain location and an adjacent location satisfies at least one of two conditions. The two conditions are: The fluctuation difference △h1 between the optical warning line at time t1 and the adjacent part and the fluctuation difference △h2 between the optical warning line at time t2 and the adjacent part satisfy │△h2-△h1│ / (t2-t1)≥5cm / s, where t2≥t1 and w / k1≤(t2-t1)≤w / k2; and, The fluctuation difference △h1 between the optical warning line at time t1 and the adjacent part, the fluctuation difference △h2 between the optical warning line at time t2 and the fluctuation difference △h3 between the optical warning line at time t3 and the adjacent part satisfy (│△h3-△h2│ / (t3-t2)) / (│△h2-△h1│ / (t2-t1))0.5≥k3, where t3≥t2≥t1, and w / k1≤(t2-t1)=(t3-t2)≤w / k2; k1 is the first state setting value, and the k1 value is in the range of 5-12; k2 is the second state setting value, and the k2 value is in the range of 0.1-5; k3 is the trend setting value, and the k3 value is in the range of 1.1-1.5; the frequency s of the dedicated camera element to collect the second type of image information is 1 / (w / k1) Hz.

5. The intelligent security system for a terminal based on video detection according to claim 3 is characterized in that: When the data processing layer identifies that there is a perimeter intrusion in the crane operation danger zone, the point identification result is confirmed to be that there is a perimeter intrusion in the crane operation danger zone; when the data processing layer identifies that there is a perimeter intrusion in the cargo area or the key cargo location within the cargo area, the identity and location information of the personnel or mobile device in the area where the perimeter intrusion is located is immediately retrieved; if the location information of the personnel or mobile device is consistent with the perimeter intrusion, and the identity information has the access authorization for the cargo area or the key cargo location, the point identification result is determined to be that there is no unauthorized perimeter intrusion in the cargo area or the key cargo location; if the location information of the personnel or mobile device is consistent with the perimeter intrusion, but the identity information does not have the access authorization for the cargo area or the key cargo location, or there is no personnel or mobile device in the station that is consistent with the perimeter intrusion, then the point identification result is confirmed to be that there is an unauthorized perimeter intrusion in the cargo area or the key cargo location.

6. The intelligent security system for a station based on video detection according to claim 5 is characterized in that: The identification result of the item point is one or more of the following: unauthorized perimeter intrusion in a cargo area, unauthorized perimeter intrusion in a key cargo location within a cargo area, and perimeter intrusion in a crane operation danger zone. When the location information of the person or mobile device matches the perimeter intrusion, the command center layer performs early warning and intelligent processing based on the identification result of this point, including: the command center layer sends a processing instruction to the second-class acquisition module and the personnel response device of the person involved; the voice device of the second-class acquisition module reads the identity information of the person involved or the mobile device through voice according to the processing instruction and warns and drives it away; at the same time, the control element of the personnel response device activates the vibration, lighting, and alarm functions of the early warning element according to the processing instruction, and the audio transceiver element plays the voice command independently generated by the intelligent processing of the command center layer to the person involved, and feeds the voice of the person involved back to the command center layer; When there are no personnel or mobile equipment in the station that match the perimeter intrusion, the command center layer will perform early warning and intelligent processing based on the results of the identification items, including: the command center layer sends processing instructions to the second-class acquisition module; the voice device of the second-class acquisition module uses voice to issue a stern warning and drive away according to the processing instructions. At the same time, staff are dispatched from the station center or deployed from nearby to immediately rush to the area for investigation.

7. The intelligent security system for a terminal based on video detection according to claim 6 is characterized in that: When the second-class acquisition module is in the harsh environment working mode, the warning and intelligent processing performed by the command center based on the identification results also include: instructing the intelligent projection device to project a red or red and green flashing optical warning line, which serves as a warning and deterrent to the parties involved and as a guide for the staff who come to investigate.

8. The intelligent security system for a terminal based on video detection according to claim 6 is characterized in that: The device response device further includes an execution element, which is connected to the power device of the mobile device and can intervene in the power device of the mobile device.

9. The intelligent security system for a terminal based on video detection according to claim 8 is characterized in that: When there is an unauthorized perimeter intrusion in a cargo area or a key cargo location, and the location information of a person or mobile device is consistent with the perimeter intrusion, the early warning and intelligent processing performed by the command center layer based on the item identification item also includes: the command center layer sends a processing instruction to the device response device of the mobile device, and the control element of the device response device controls the execution element to shut down the power output of the mobile device in question according to the processing instruction. Then, the command center layer communicates with the person in question through the audio transceiver element of the person in question or the mobile device in question. After the command center layer confirms the permission, the control element of the device response device controls the execution element to restore the power output of the mobile device in question according to the processing instruction, allowing the mobile device in question to leave the area. When there is a perimeter intrusion in a crane operation danger zone, the command center layer sends a processing instruction to the equipment response device of the crane in the crane operation danger zone. The control element of the equipment response device controls the execution element to shut down or slow down the power output of the crane in question according to the processing instruction; after the data processing layer removes the perimeter intrusion identification item based on the second-class image information collected by the second-class acquisition module of the area, the command center layer confirms and permits it, and the control element of the equipment response device controls the execution element to restore the power output of the crane in question according to the processing instruction.

10. The intelligent security system for a terminal based on video detection according to claim 1, characterized in that: The identification items of the first-class collection module include: one or more of the following: smoking identification, fire identification, identification of not wearing work clothes, identification of leaving key positions, identification of dozing, identification of being shirtless, intelligent identification of loading and unloading operations, intelligent identification of cargo area and cargo space occupancy, and intelligent identification of red card placement; The image information collected by the third type of acquisition module is used to be combined with the first type of image information and / or the second type of image information to verify the accuracy of the item point recognition results corresponding to the identification item points of the first type of acquisition module and / or the identification phase points of the second type of acquisition module.

Images