Power plant illegal behavior intelligent alarm method and device

Abstract

The invention belongs to the technical field of electric power safety, and discloses an intelligent warning method and device for violation behaviors of a power plant. The method comprises the following steps: acquiring a plurality of area images of a to-be-monitored power plant through monitoring equipment; a double-flow network structure is utilized to identify the worker and the safety protection equipment worn by the worker based on the regional image, and the monitoring equipment is controlled to track and monitor the worker in real time based on the regional image; acquiring positioning information of a worker through the ultra-bandwidth base station; identifying whether the staff has illegal behaviors or not based on the identification result of the safety protection equipment, the real-time tracking and monitoring result of the staff and the positioning information; and outputting a corresponding alarm prompt when the staff has illegal behaviors. Through the above mode, target identification and tracking are combined with base station positioning, the illegal behaviors of the staff can be monitored in real time, the monitoring efficiency of the illegal behaviors is improved, and the frequency of occurrence of power accidents is effectively reduced.

Description

Technical Field

[0001] The present invention relates to the field of electric power safety technology, and in particular to an intelligent alarm method and device for power plant illegal behavior. Background Art

[0002] In the process of power production and construction, power accidents are inevitable. Power accidents will not only bring economic losses but also cause personal injuries. The main cause of power accidents is the illegal operation of staff.

[0003] At present, power operation sites mainly use manual supervision to correct workers' illegal operations. However, manual supervision is inefficient and highly subjective, and cannot ensure timely monitoring of violations.

[0004] The above contents are only used to assist in understanding the technical solution of the present invention and do not constitute an admission that the above contents are prior art. Summary of the invention

[0005] The main purpose of the present invention is to provide an intelligent alarm method and device for power plant violation behavior, aiming to solve the technical problem that the manual supervision method is inefficient and highly subjective, and cannot ensure timely monitoring of violation behaviors.

[0006] To achieve the above object, the present invention provides a power plant violation behavior intelligent alarm method, the power plant violation behavior intelligent alarm method comprising the following steps:

[0007] Acquire images of several areas of the power plant to be monitored through monitoring equipment;

[0008] Using a dual-stream network structure to identify the worker and the safety protection equipment worn by the worker based on the regional image, and controlling the monitoring device to track and monitor the worker in real time based on the regional image;

[0009] Acquiring the location information of the staff member through an ultra-wideband base station;

[0010] Based on the safety protection equipment identification results, the real-time tracking and monitoring results of the staff and the positioning information, it is identified whether the staff has violated the regulations;

[0011] When the staff member violates the regulations, a corresponding alarm prompt is output.

[0012] In some embodiments, the dual-stream network structure is a neural network model trained using an RGB stream and an HSV stream, and the method further includes:

[0013] After inputting the obtained candidate regions into the ROIpooling layer to adjust the size, the bilinear pooling operation is performed to perform feature fusion, wherein the bilinear pooling operation is to perform feature fusion on two features at the same position of the image;

[0014] The obtained fusion features are sent to the Bounding Box for border regression and Softmax for classification prediction after passing through the fully connected layer. A four-dimensional vector (x, y, w, h) is used to represent a window. The coordinates of the center point of the window are represented by x and y, and the width and height are represented by w and h. The window is adjusted according to the boundary regression result.

[0015] The model accuracy is calculated using the number of correctly classified positive samples, the number of correctly classified negative samples, the number of positive samples, and the number of negative samples. After the model accuracy meets the requirements, it is determined that the dual-stream network structure has completed training.

[0016] In some embodiments, controlling the monitoring device to perform real-time tracking and monitoring of the staff based on the regional image includes:

[0017] determining a blurred region image among a plurality of region images;

[0018] Eliminating the blurred area image from the plurality of area images to obtain a clear area image;

[0019] The clear area image is used to control the monitoring device to track and monitor the staff in real time.

[0020] In some embodiments, determining a blurred area image among the plurality of area images comprises:

[0021] Convert several regional images into grayscale images respectively;

[0022] Calculate the grayscale value of each pixel in the grayscale image based on the pixel values ​​corresponding to the RGB channels of each regional image and the preset weights corresponding to the RGB channels;

[0023] Determine the blur of each area image according to the gray value in combination with the Laplace operator and the total number of pixels of each area image;

[0024] The area image whose blurriness is lower than the preset blurriness is used as the blurred area image.

[0025] In some embodiments, the using of the clear area image to control the monitoring device to perform real-time tracking and monitoring of the staff member includes:

[0026] Determining location information of the staff member based on identification of the staff member;

[0027] determining the center coordinates of the region of interest according to the position information and the image size of the clear area image, and determining the coordinates of the center pixel based on the image size of the clear area image;

[0028] Determining a distance offset according to the center coordinates of the region of interest and the coordinates of the center pixel;

[0029] determining an angular offset based on the distance offset;

[0030] The monitoring device is controlled by the angle offset to track and monitor the staff in real time.

[0031] In some embodiments, the obtaining the location information of the staff member through the ultra-wideband base station includes:

[0032] Acquire ultra-wideband signals through ultra-wideband base stations;

[0033] The classification result corresponding to the ultra-wideband signal is obtained by using a positioning algorithm, wherein the positioning algorithm calculates the distance between the input sample ultra-wideband signal and each training data, sorts the matched training data according to the increasing relationship of the distance, selects k points with the smallest distance to the test data, calculates the frequency of occurrence of the category of the k points, and takes the category with the highest frequency of occurrence among the first k points as the predicted classification of the sample ultra-wideband signal;

[0034] The location information of the staff member is obtained based on the classification result.

[0035] In some embodiments, the step of identifying whether the staff has violated the regulations based on the safety protection equipment identification result, the staff's real-time tracking and monitoring result, and the positioning information includes:

[0036] Determining whether the worker is fully wearing safety protection equipment based on the safety protection equipment identification result;

[0037] Determine whether the worker has entered a dangerous area based on the real-time tracking and monitoring results of the worker and the positioning information;

[0038] If the worker does not fully wear safety protection equipment and enters a dangerous area, the worker is deemed to have violated the regulations.

[0039] In addition, to achieve the above-mentioned purpose, the present invention also proposes an intelligent warning device for power plant violation behavior, the intelligent warning device for power plant violation behavior comprising:

[0040] An acquisition module, used to acquire images of several areas of the power plant to be monitored through monitoring equipment;

[0041] A monitoring module, used to identify the staff and the safety protection equipment worn by the staff based on the regional image using a dual-stream network structure, and control the monitoring equipment to track and monitor the staff in real time based on the regional image;

[0042] A positioning module, used to obtain the positioning information of the staff member through an ultra-wideband base station;

[0043] A judgment module, used to identify whether the staff has violated the regulations based on the identification results of the safety protection equipment, the real-time tracking and monitoring results of the staff, and the positioning information;

[0044] The alarm module is used to output corresponding alarm prompts when the staff member violates the regulations.

[0045] In some embodiments, the power plant violation behavior intelligent alarm device further includes a training module;

[0046] The training module is used to input the obtained candidate regions into the ROIpooling layer for resizing, and then perform feature fusion through a bilinear pooling operation, wherein the bilinear pooling operation is to perform feature fusion on two features at the same position of the image;

[0047] The obtained fusion features are sent to the Bounding Box for border regression and Softmax for classification prediction after passing through the fully connected layer. A four-dimensional vector (x, y, w, h) is used to represent a window. The coordinates of the center point of the window are represented by x and y, and the width and height are represented by w and h. The window is adjusted according to the boundary regression result.

[0048] The model accuracy is calculated using the number of correctly classified positive samples, the number of correctly classified negative samples, the number of positive samples, and the number of negative samples. After the model accuracy meets the requirements, it is determined that the dual-stream network structure has completed training.

[0049] In some embodiments, the monitoring module is used to determine a blurred area image among a plurality of area images;

[0050] Eliminating the blurred area image from the plurality of area images to obtain a clear area image;

[0051] The clear area image is used to control the monitoring device to track and monitor the staff in real time.

[0052] The present invention obtains several regional images of the power plant to be monitored through monitoring equipment; uses a dual-stream network structure to identify the staff and the safety protection equipment worn by the staff based on the regional images, and controls the monitoring equipment to track and monitor the staff in real time based on the regional images; obtains the positioning information of the staff through an ultra-wideband base station; identifies whether the staff has violated the regulations based on the identification results of the safety protection equipment, the real-time tracking and monitoring results of the staff, and the positioning information; and outputs corresponding alarm prompts when the staff has violated the regulations. In the above manner, by using target identification and tracking combined with base station positioning, the staff's violations can be monitored in real time, which improves the monitoring efficiency of violations and effectively reduces the frequency of power accidents. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 It is a flow chart of the first embodiment of the intelligent alarm method for power plant violation behavior of the present invention;

[0054] Figure 2 It is a frame corresponding to the window in the intelligent alarm method for power plant violation behavior of the present invention and a corresponding classification schematic diagram;

[0055] Figure 3 This is a structural block diagram of the first embodiment of the intelligent alarm device for power plant violations of the present invention.

[0056] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0057] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0058] The embodiment of the present invention provides a power plant violation behavior intelligent alarm method, referring to Figure 1 , Figure 1 This is a flow chart of a first embodiment of a method for intelligent warning of power plant violations according to the present invention.

[0059] In this embodiment, the power plant violation behavior intelligent alarm method includes the following steps:

[0060] Step S10: Acquire images of several areas of the power plant to be monitored through monitoring equipment.

[0061] In this embodiment, the executor of this embodiment is an intelligent alarm device for power plant violations, wherein the intelligent alarm device for power plant violations has functions such as data processing, data communication and program running. The intelligent alarm device for power plant violations can be a computer terminal device or other network device, and of course it can also be other devices with similar functions, and this embodiment does not limit this.

[0062] It should be noted that, at present, power operation sites mainly use manual supervision to correct workers' illegal operations. However, manual supervision is inefficient and highly subjective, and cannot ensure timely monitoring of violations.

[0063] In order to solve the above technical problems, this embodiment obtains several regional images of the power plant to be monitored through monitoring equipment; uses a dual-stream network structure to identify the staff and the safety protection equipment worn by the staff based on the regional images, and controls the monitoring equipment to track and monitor the staff in real time based on the regional images; obtains the positioning information of the staff through an ultra-wideband base station; identifies whether the staff has violated the regulations based on the identification results of the safety protection equipment, the real-time tracking and monitoring results of the staff, and the positioning information; and outputs corresponding alarm prompts when the staff has violated the regulations. Through the above method, by using target identification and tracking combined with base station positioning, the violations of the staff can be monitored in real time, which improves the monitoring efficiency of violations and effectively reduces the frequency of power accidents. Specifically, it can be achieved in the following way.

[0064] In a specific implementation, in this embodiment, it is necessary to first obtain several area images of the power plant to be monitored through the monitoring equipment, and the several area images are images of different areas of continuous frames and multiple images of the same area.

[0065] Step S20: using a dual-stream network structure to identify the worker and the safety protection equipment worn by the worker based on the regional image, and controlling the monitoring device to track and monitor the worker in real time based on the regional image.

[0066] The dual-stream network structure in this embodiment is a neural network model trained by RGB stream and HSV stream. The RGB stream corresponds to the RGB color space. The RGB color mixing method can be represented by red, green and blue lights. When the lights of the red, green and blue lights overlap and overlap, the brightness of the light we can observe is equal to the sum of the brightness of the three lights. In the RGB color space, the component sizes of the red, green and blue color channels are all 0-255. The HSV color space corresponding to the HSV stream, the hue H in the HSV color space is measured by angle, with a value of 0-360 degrees. The saturation S represents the degree to which the color is close to the spectral color. The darker the color, the greater the value of the saturation S. The value range of the saturation S is 0% to 100%. The larger the S value, the darker and more saturated the color.

[0067] In this embodiment, the specific training process of the dual-stream network structure is as follows: after inputting the obtained several candidate regions into the ROIpooling layer to adjust the size, the features are fused through the bilinear pooling operation; after the obtained fused features pass through the fully connected layer, they are respectively sent to the Bounding Box for border regression and the Softmax for classification prediction; the model accuracy is calculated using the number of correctly classified positive samples, the number of correctly classified negative samples, the number of positive samples and the number of negative samples, and after the model accuracy meets the requirements, it is determined that the dual-stream network structure has completed training.

[0068] It should be noted that the amount of candidate area data selected in this embodiment can be 300. The bilinear pooling operation is to fuse the two features at the same position of the image. Specifically, the RGB feature can be fused with the HSV feature. The window corresponding to the border regression is represented by a four-dimensional vector (x, y, w, h). The coordinates of the center point of the window are represented by x and y, and the width and height are represented by w and h. The window is adjusted according to the border regression result. The formed border and the corresponding classification can be referred to Figure 2 Finally, the accuracy of the model is tested to determine whether the training is complete. The accuracy calculation formula is:

[0069]

[0070] Among them, TP is the number of correctly classified positive samples, TN is the number of correctly classified negative samples, and N p is the number of positive samples, N neg is the number of negative samples.

[0071] Furthermore, in this embodiment, the above-mentioned regional images can also be used to control the monitoring equipment to perform real-time tracking and monitoring of the staff. The above-mentioned several regional images are composed of a series of continuous images. When the camera carried by the monitoring device is rotated to realize the real-time target tracking function, blurred images will inevitably be captured. In order to perform more accurate tracking and monitoring, these blurred regional images need to be eliminated in this embodiment, and subsequent real-time tracking is performed through the remaining clear regional images.

[0072] The specific process of identifying blurred area images in this embodiment is to convert several area images into grayscale images respectively; calculate the grayscale value of each pixel in the grayscale image based on the pixel value corresponding to the RGB channel of each area image and the preset weight corresponding to each RGB channel; determine the blurriness of each area image according to the grayscale value and in combination with the Laplace operator and the total number of pixels of each area image; and use the area image with a blurriness lower than the preset blurriness as a blurred area image.

[0073] It should be noted that the calculation formula for the gray value of each pixel is as follows:

[0074] I(x,y)=W R R x,y +W G G x,y +W B B x,y

[0075] Among them, W R , W G and W B They are the preset weights corresponding to the RGB channels respectively, R x,y , G x,y and B x,y are the pixel values ​​corresponding to the RGB channels of each region image. The calculation formula for fuzziness is:

[0076]

[0077] Where N is the total number of pixels in each area image, That is, the Laplace operator, The preset fuzziness can be set according to actual needs, and this is not limited in this embodiment.

[0078] Furthermore, the process of real-time tracking and monitoring of the staff is to determine the position information of the staff based on the identification of the staff; determine the center coordinates of the area of ​​interest based on the position information and the image size of the clear area image, and determine the coordinates of the center pixel based on the image size of the clear area image; determine the distance offset based on the center coordinates of the area of ​​interest and the coordinates of the center pixel; determine the angle offset based on the distance offset; and control the monitoring device through the angle offset to perform real-time tracking and monitoring of the staff.

[0079] It should be noted that, assuming that the coordinates corresponding to the staff member's location information are (x a ,y b ), the image size of the clear area image is w,h, then the center coordinates of the region of interest are (C x ,C y ), where C x =x a +w / 2,C y =y b +h / 2, the coordinates of the center pixel (T x ,T y ), where T x =w / 2,T y =h / 2, the calculated distance offsets are Δx=2×(C x / Tx -0.5), Δy=2×(C y / T y -0.5). The angle offset calculation formula is:

[0080]

[0081] Among them, ΔD is the above-mentioned Δx or Δy, Kp is the adjustment proportional coefficient, and err(t) corresponds to the error of the monitoring device itself.

[0082] Step S30: Acquire the location information of the staff member through the ultra-wideband base station.

[0083] When positioning through an ultra-wideband base station, the positioning algorithm can be used to determine the classification result corresponding to the acquired ultra-wideband signal, and the classification result is also the corresponding positioning information of the staff. Among them, the positioning algorithm is to calculate the distance between the input sample ultra-wideband signal and each training data, sort the matched training data according to the increasing relationship of the distance, select the k points with the smallest distance from the test data, calculate the frequency of occurrence of the category of the k points, and use the category with the highest frequency of occurrence among the first k points as the predicted classification of the sample ultra-wideband signal. The category of the above k points is also the corresponding positioning point label number, and positioning can be completed by classifying the positioning point label number.

[0084] Step S40: Based on the safety protection equipment identification result, the real-time tracking and monitoring result of the staff and the positioning information, it is identified whether the staff has violated the regulations.

[0085] It should be noted that based on the safety protection equipment identification results, it can be determined whether the staff is fully wearing safety protection equipment. Fully wearing safety protection equipment means wearing a complete set of safety helmets, safety protection clothing and safety protection gloves as required. Combined with the real-time tracking and monitoring results of the staff and the positioning information, it can be determined whether the staff has entered a dangerous area. If the staff does not fully wear safety protection equipment and enters a dangerous area, it is determined that the staff has violated the regulations.

[0086] Step S50: When the staff member violates the regulations, a corresponding alarm prompt is output.

[0087] It should be understood that if a staff member violates the rules, a corresponding warning prompt will be taken in this embodiment, such as sending an alarm signal to the staff member's mobile terminal, and alerting the staff member through vibration or buzzing.

[0088] This embodiment obtains several regional images of the power plant to be monitored through monitoring equipment; uses a dual-stream network structure to identify the staff and the safety protection equipment worn by the staff based on the regional images, and controls the monitoring equipment to track and monitor the staff in real time based on the regional images; obtains the positioning information of the staff through an ultra-wideband base station; identifies whether the staff has violated the regulations based on the identification results of the safety protection equipment, the real-time tracking and monitoring results of the staff, and the positioning information; and outputs corresponding alarm prompts when the staff has violated the regulations. In the above manner, by using target identification and tracking combined with base station positioning, the staff's violations can be monitored in real time, which improves the efficiency of monitoring violations and effectively reduces the frequency of power accidents.

[0089] Reference Figure 3 , Figure 3 This is a structural block diagram of the first embodiment of the intelligent alarm device for power plant violations of the present invention.

[0090] like Figure 3 As shown, the power plant violation intelligent alarm device proposed in the embodiment of the present invention includes:

[0091] The acquisition module 10 is used to acquire images of several areas of the power plant to be monitored through monitoring equipment;

[0092] A monitoring module 20, for identifying the worker and the safety protection equipment worn by the worker based on the regional image by using a dual-stream network structure, and controlling the monitoring device to perform real-time tracking and monitoring of the worker based on the regional image;

[0093] A positioning module 30, used to obtain the positioning information of the staff member through an ultra-wideband base station;

[0094] A judgment module 40 is used to identify whether the staff has violated the regulations based on the identification result of the safety protection equipment, the real-time tracking and monitoring result of the staff and the positioning information;

[0095] The alarm module 50 is used to output corresponding alarm prompts when the staff member violates the regulations.

[0096] This embodiment obtains several regional images of the power plant to be monitored through monitoring equipment; uses a dual-stream network structure to identify the staff and the safety protection equipment worn by the staff based on the regional images, and controls the monitoring equipment to track and monitor the staff in real time based on the regional images; obtains the positioning information of the staff through an ultra-wideband base station; identifies whether the staff has violated the regulations based on the identification results of the safety protection equipment, the real-time tracking and monitoring results of the staff, and the positioning information; and outputs corresponding alarm prompts when the staff has violated the regulations. In the above manner, by using target identification and tracking combined with base station positioning, the staff's violations can be monitored in real time, which improves the efficiency of monitoring violations and effectively reduces the frequency of power accidents.

[0097] In some embodiments, the power plant violation behavior intelligent alarm device further includes a training module;

[0098] The training module is used to input the obtained candidate regions into the ROIpooling layer for resizing, and then perform feature fusion through a bilinear pooling operation, wherein the bilinear pooling operation is to perform feature fusion on two features at the same position of the image;

[0099] The obtained fusion features are sent to the Bounding Box for border regression and Softmax for classification prediction after passing through the fully connected layer. A four-dimensional vector (x, y, w, h) is used to represent a window. The coordinates of the center point of the window are represented by x and y, and the width and height are represented by w and h. The window is adjusted according to the boundary regression result.

[0100] The model accuracy is calculated using the number of correctly classified positive samples, the number of correctly classified negative samples, the number of positive samples, and the number of negative samples. After the model accuracy meets the requirements, it is determined that the dual-stream network structure has completed training.

[0101] In some embodiments, the monitoring module 20 is used to determine a blurred area image among a plurality of area images;

[0102] Eliminating the blurred area image from the plurality of area images to obtain a clear area image;

[0103] The clear area image is used to control the monitoring device to track and monitor the staff in real time.

[0104] In some embodiments, the monitoring module 20 is used to convert several regional images into grayscale images respectively;

[0105] Calculate the grayscale value of each pixel in the grayscale image based on the pixel values ​​corresponding to the RGB channels of each regional image and the preset weights corresponding to the RGB channels;

[0106] Determine the blur of each area image according to the gray value in combination with the Laplace operator and the total number of pixels of each area image;

[0107] The area image whose blurriness is lower than the preset blurriness is used as the blurred area image.

[0108] In some embodiments, the monitoring module 20 is used to determine the location information of the staff member based on the identification of the staff member;

[0109] determining the center coordinates of the region of interest according to the position information and the image size of the clear area image, and determining the coordinates of the center pixel based on the image size of the clear area image;

[0110] Determining a distance offset according to the center coordinates of the region of interest and the coordinates of the center pixel;

[0111] determining an angular offset based on the distance offset;

[0112] The monitoring device is controlled by the angle offset to track and monitor the staff in real time.

[0113] In some embodiments, the positioning module 30 is used to obtain an ultra-wideband signal through an ultra-wideband base station;

[0114] The classification result corresponding to the ultra-wideband signal is obtained by using a positioning algorithm, wherein the positioning algorithm calculates the distance between the input sample ultra-wideband signal and each training data, sorts the matched training data according to the increasing relationship of the distance, selects k points with the smallest distance to the test data, calculates the frequency of occurrence of the category of the k points, and takes the category with the highest frequency of occurrence among the first k points as the predicted classification of the sample ultra-wideband signal;

[0115] The location information of the staff member is obtained based on the classification result.

[0116] In some embodiments, the judgment module 40 is used to judge whether the worker is fully wearing the safety protection equipment based on the safety protection equipment identification result;

[0117] Determine whether the worker has entered a dangerous area based on the real-time tracking and monitoring results of the worker and the positioning information;

[0118] If the worker does not fully wear safety protection equipment and enters a dangerous area, the worker is deemed to have violated the regulations.

[0119] An embodiment of the present application also provides an intelligent alarm device for power plant violations, including a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus, and the memory is used to store the intelligent alarm program for power plant violations; the processor is used to implement the above-mentioned intelligent alarm method for power plant violations when executing the program stored in the memory.

[0120] The communication bus mentioned in the above-mentioned power plant violation intelligent alarm device can be a peripheral component interconnect standard (English: Peripheral Component Interconnect, referred to as: PCI) bus or an extended industry standard architecture (English: Extended Industry Standard Architecture, referred to as: EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc.

[0121] The communication interface is used for communication between the above-mentioned power plant violation intelligent alarm device and other devices.

[0122] The memory may include a random access memory (RAM) or a non-volatile memory (NVM), such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0123] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, and discrete hardware components.

[0124] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated. The available medium may be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive Solid State Disk (SSD)), etc.

[0125] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0126] Each embodiment in this specification is described in a related manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0127] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

[0128] It should be understood that the above is only an example and does not constitute any limitation on the technical solution of the present invention. In specific applications, technicians in this field can make settings as needed, and the present invention does not limit this.

[0129] It should be noted that the workflow described above is merely illustrative and does not limit the scope of protection of the present invention. In practical applications, technicians in this field can select part or all of them according to actual needs to achieve the purpose of the present embodiment, and no limitation is made here.

[0130] In addition, for technical details that are not described in detail in this embodiment, please refer to the intelligent alarm method for power plant violation behavior provided in any embodiment of the present invention, and will not be repeated here.

[0131] In addition, it should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0132] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0133] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as a read-only memory (ROM) / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0134] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

[0135] It is understandable that the system provided by the embodiment of the present invention corresponds to the method provided by the embodiment of the present invention, and the explanation, examples and beneficial effects of the relevant contents can refer to the corresponding parts in the above method.

Claims

1. An intelligent alarm method for power plant violation behavior, characterized in that: The power plant violation behavior intelligent alarm method comprises: Acquire images of several areas of the power plant to be monitored through monitoring equipment; Using a dual-stream network structure to identify the worker and the safety protection equipment worn by the worker based on the regional image, and controlling the monitoring device to track and monitor the worker in real time based on the regional image; Acquiring location information of the staff member through an ultra-wideband base station; Based on the safety protection equipment identification results, the real-time tracking and monitoring results of the staff and the positioning information, it is identified whether the staff has violated the regulations; When the staff member violates the regulations, a corresponding alarm prompt is output.

2. The power plant violation intelligent alarm method according to claim 1, characterized in that: The dual-stream network structure is a neural network model trained using an RGB stream and an HSV stream, and the method further includes: After inputting the obtained candidate regions into the ROIpooling layer to adjust the size, the bilinear pooling operation is performed to perform feature fusion, wherein the bilinear pooling operation is to perform feature fusion on two features at the same position of the image; The obtained fusion features are sent to the Bounding Box for border regression and Softmax for classification prediction after passing through the fully connected layer. A four-dimensional vector (x, y, w, h) is used to represent a window. The coordinates of the center point of the window are represented by x and y, and the width and height are represented by w and h. The window is adjusted according to the boundary regression result. The model accuracy is calculated using the number of correctly classified positive samples, the number of correctly classified negative samples, the number of positive samples, and the number of negative samples. After the model accuracy meets the requirements, it is determined that the dual-stream network structure has completed training.

3. The power plant violation intelligent alarm method according to claim 1, characterized in that: The controlling the monitoring device to perform real-time tracking and monitoring of the staff based on the regional image includes: determining a blurred region image among a plurality of region images; Eliminating the blurred area image from the plurality of area images to obtain a clear area image; The clear area image is used to control the monitoring device to track and monitor the staff in real time.

4. The power plant violation intelligent alarm method according to claim 3, characterized in that: The step of determining a blurred area image among the plurality of area images comprises: Convert several regional images into grayscale images respectively; Calculate the grayscale value of each pixel in the grayscale image based on the pixel values ​​corresponding to the RGB channels of each regional image and the preset weights corresponding to the RGB channels; Determine the blur of each area image according to the gray value in combination with the Laplace operator and the total number of pixels of each area image; The area image whose blurriness is lower than the preset blurriness is used as the blurred area image.

5. The intelligent warning method for power plant violation behavior according to claim 3 is characterized in that: The using of the clear area image to control the monitoring device to perform real-time tracking and monitoring of the staff member includes: Determining location information of the staff member based on identification of the staff member; determining the center coordinates of the region of interest according to the position information and the image size of the clear area image, and determining the coordinates of the center pixel based on the image size of the clear area image; Determining a distance offset according to the center coordinates of the region of interest and the coordinates of the center pixel; determining an angular offset based on the distance offset; The monitoring device is controlled by the angle offset to track and monitor the staff in real time.

6. The power plant violation behavior intelligent alarm method according to claim 1, characterized in that: The obtaining the positioning information of the staff member through the ultra-wideband base station includes: Acquire ultra-wideband signals through ultra-wideband base stations; The classification result corresponding to the ultra-wideband signal is obtained by using a positioning algorithm, wherein the positioning algorithm calculates the distance between the input sample ultra-wideband signal and each training data, sorts the matched training data according to the increasing relationship of the distance, selects k points with the smallest distance to the test data, calculates the frequency of occurrence of the category of the k points, and takes the category with the highest frequency of occurrence among the first k points as the predicted classification of the sample ultra-wideband signal; The location information of the staff member is obtained based on the classification result.

7. The power plant violation intelligent alarm method according to any one of claims 1 to 6, characterized in that: The identification of whether the staff has violated the regulations based on the safety protection equipment identification results, the real-time tracking and monitoring results of the staff, and the positioning information includes: Determining whether the worker is fully wearing safety protection equipment based on the safety protection equipment identification result; Determine whether the worker has entered a dangerous area based on the real-time tracking and monitoring results of the worker and the positioning information; If the worker does not fully wear safety protection equipment and enters a dangerous area, the worker is deemed to have violated the regulations.

8. An intelligent warning device for power plant violations, characterized in that: The power plant violation intelligent alarm device comprises: An acquisition module, used to acquire images of several areas of the power plant to be monitored through monitoring equipment; A monitoring module, used to identify the staff and the safety protection equipment worn by the staff based on the regional image using a dual-stream network structure, and control the monitoring equipment to track and monitor the staff in real time based on the regional image; A positioning module, used to obtain the positioning information of the staff member through an ultra-wideband base station; A judgment module, used to identify whether the staff has violated the regulations based on the identification results of the safety protection equipment, the real-time tracking and monitoring results of the staff, and the positioning information; The alarm module is used to output corresponding alarm prompts when the staff member violates the regulations.

9. The intelligent warning device for power plant violation behavior according to claim 8, characterized in that: The power plant violation behavior intelligent alarm device also includes a training module; The training module is used to input the obtained candidate regions into the ROIpooling layer for resizing, and then perform feature fusion through a bilinear pooling operation, wherein the bilinear pooling operation is to perform feature fusion on two features at the same position of the image; The obtained fusion features are sent to the Bounding Box for border regression and Softmax for classification prediction after passing through the fully connected layer. A four-dimensional vector (x, y, w, h) is used to represent a window. The coordinates of the center point of the window are represented by x and y, and the width and height are represented by w and h. The window is adjusted according to the boundary regression result. The model accuracy is calculated using the number of correctly classified positive samples, the number of correctly classified negative samples, the number of positive samples, and the number of negative samples. After the model accuracy meets the requirements, it is determined that the dual-stream network structure has completed training.

10. The intelligent warning device for power plant violation behavior according to claim 8, characterized in that: The monitoring module is used to determine the blurred area image among the plurality of area images; Eliminating the blurred area image from the plurality of area images to obtain a clear area image; The clear area image is used to control the monitoring device to track and monitor the staff in real time.

Images