Remote safety management system for electric power construction site

Abstract

The invention belongs to the field of construction safety management, and discloses an electric power construction site remote safety management system, which comprises camera equipment, edge processing equipment and remote monitoring equipment, the shooting equipment is used for acquiring a field video of a power construction field; the edge processing device comprises a period calculation module, a code rate calculation module, a compression module and a transmission module. The period calculation module is used for calculating a code rate adjustment period based on the on-site video; the code rate calculation module is used for calculating the code rate of video compression according to the code rate adjustment period; the compression module is used for compressing the on-site video according to the code rate; the transmission module is used for transmitting the compressed on-site video to remote monitoring equipment; and the remote monitoring equipment is used for performing construction safety monitoring on the power construction site based on the compressed site video. According to the invention, the code rate and the adjustment period can be dynamically adjusted according to the change of the on-site video content, and when a picture is greatly changed, the transmission effect is timely responded and optimized, so that the real-time performance and the effectiveness of monitoring are ensured.

Description

Technical Field

[0001] The present invention relates to the field of construction safety management, and particularly to a remote safety management system for electric power construction sites. Background Art

[0002] In a remote monitoring system for construction sites based on cameras, if the on-site video is continuously transmitted at a high bit rate, although the clarity of the on-site video received by the remote monitoring center can be ensured, the continuous transmission of high-bit-rate video streams will significantly increase the occupancy rate of network bandwidth, which poses higher requirements for the receiving devices of the monitoring center. To ensure the real-time reception and smooth playback of video streams, the monitoring center needs to be equipped with a relatively high receiving high bandwidth, otherwise video stuttering or even loss of key frames may occur due to network congestion. Especially in complex scenarios where multiple on-site videos from multiple construction sites are transmitted concurrently, this contradiction is particularly prominent.

[0003] In order to reduce the bit rate of video compression for on-site videos before transmission while ensuring the effectiveness of monitoring, the prior art usually calculates the bit rate regularly based on the change situation of the content of the on-site video frame. When the change degree of the content of the video frame is large (for example, a large number of workers appear in the monitoring frame), a higher bit rate is used for compressing the on-site video, and when the change degree of the content of the video frame is small (for example, the objects in the monitoring frame do not change their positions for a long time), a lower bit rate is used for compressing the on-site video, thereby reducing the requirement for the receiving bandwidth of the monitoring center. This method has the following disadvantages: when the time interval between two adjacent bit rate determinations is too small, the average delay of the on-site video received by the remote monitoring center will be too large, because frequent bit rate calculations will occupy more computing resources, affecting the efficiency of video compression based on the bit rate and resulting in a longer overall compression cycle; when the time interval between two adjacent bit rate determinations is too large, it is easy to occur that when the video frame changes greatly, the on-site video is still transmitted at a lower bit rate, thus affecting the quality of the on-site video received by the monitoring center and further affecting the effectiveness of monitoring. Summary of the Invention

[0004] The purpose of the present invention is to disclose a remote safety management system for electric power construction sites to solve the technical problems raised in the background art.

[0005] To achieve the above purpose, the present invention provides the following technical solutions:

[0006] The present invention provides a remote safety management system for electric power construction sites, including a shooting device, an edge processing device, and a remote monitoring device;

[0007] The shooting device is used to acquire the on-site video of the electric power construction site and transmit the on-site video to the edge processing device;

[0008] The edge processing device is set at the power construction site. The edge processing device includes a cycle calculation module, a bit rate calculation module, a compression module, and a transmission module;

[0009] The cycle calculation module is used to calculate the bit rate adjustment cycle based on the on-site video, including:

[0010] Let n represent the serial number of the current bit rate adjustment cycle, and calculate the (n + 1)-th bit rate adjustment cycle based on the on-site video received within the n-th bit rate adjustment cycle;

[0011] The bit rate calculation module is used to calculate the bit rate of video compression according to the bit rate adjustment cycle;

[0012] The compression module is used to compress the on-site video according to the bit rate;

[0013] The transmission module is used to transmit the compressed on-site video to the remote monitoring device;

[0014] The remote monitoring device is used to perform construction safety monitoring on the power construction site based on the compressed on-site video.

[0015] Further, the remote monitoring device includes a video database, a safety identification module, and a safety warning module;

[0016] The video database is used to store the compressed on-site video transmitted by the transmission module;

[0017] The safety identification module is used to read the compressed on-site video from the video database and perform intelligent identification on the compressed on-site video to determine whether there is an alarm event;

[0018] The safety warning module is used to generate a safety notice based on the alarm event and transmit the safety notice to the electronic terminal used by the safety management personnel.

[0019] Further, obtaining the on-site video of the power construction site and transmitting the on-site video to the edge processing device includes:

[0020] Continuously shoot the construction site at a set frame rate to obtain a frame sequence, and extract the frame sequence with a set time window to form the on-site video.

[0021] Further, calculating the (n + 1)-th bit rate adjustment cycle based on the on-site video received within the n-th bit rate adjustment cycle includes:

[0022] Respectively use t s,n and t d,n to represent the start time and end time of the n-th bit rate adjustment cycle;

[0023] After completely receiving the on-site video transmitted by the shooting device, the edge processing device records the reception time of the on-site video;

[0024] Sort the on-site videos with reception times within the time range [t s,n , t d,n in ascending order of reception time to obtain the on-site video sequence sq n ;

[0025] Based on sq n Calculate the (n + 1)-th bit rate adjustment period.

[0026] Furthermore, calculating the (n + 1)-th bit rate adjustment period based on sq n includes:

[0027] Obtain the on-site video video n that is the last in sq lst ;

[0028] Obtain the sequence frasq composed of the (N - s)-th frame to the N-th frame in video lst , where N represents the total number of frames included in video lst , and s is the frame extraction coefficient;

[0029] Calculate the state coefficient state based on frasq;

[0030] Calculate the (n + 1)-th bit rate adjustment period ratcy n+1 using the following formula:

[0031] ratcy n+1 = ratcy pr × state

[0032] ratcy pr represents the set duration.

[0033] Furthermore, the calculation formula for the state coefficient state is:

[0034]

[0035] fram pr is the preset reference value; fram ave represents the average gray value of the pixel points of the frames in frasq; fram i represents the average gray value of the pixel points in the i-th frame in frasq.

[0036] Furthermore, the calculation formula for fram i is:

[0037]

[0038] Further, calculating the video compression bitrate according to the bitrate adjustment period includes:

[0039]

[0040] comprat n+1 represents the video compression bitrate used in the (n + 1)-th bitrate adjustment period; frac j represents the change value of the j-th frame in frasq; ws j is the influence coefficient of frac j ; comprat pr represents the preset bitrate; fracm is the median of the change values of all frames in frasq, and λ is the weight.

[0041] Further, the security identification module includes a reading unit, a screening unit, a preprocessing unit, and an identification unit;

[0042] The reading unit is used to read the compressed on-site video from the video database;

[0043] The screening unit is used to obtain the frames for identification from the compressed on-site video read by the reading unit;

[0044] The preprocessing unit is used to preprocess the frames obtained by the screening unit to obtain the preprocessed frames;

[0045] The identification unit is used to identify the preprocessed frames to determine whether there is an alarm event.

[0046] Further, obtaining the frames for identification from the compressed on-site video read by the reading unit includes:

[0047] Let L represent the sequence of frames in the compressed on-site video read by the reading unit;

[0048] Extract one frame from L every K frames as the frame for identification, where K is the set reading interval.

[0049] Beneficial effects:

[0050] Compared with the prior art, in the process of remotely managing the safety of a power construction site, the present invention does not use a fixed bit rate to compress and transmit the on-site video. Instead, it first calculates the next bit rate adjustment period based on the on-site video received in the previous bit rate adjustment period, and then calculates the bit rate of video compression based on the next bit rate adjustment period. Thus, while realizing the adaptive change of the bit rate adjustment period, it also realizes the adaptive change of the bit rate of video compression. It can make the bit rate of on-site video compression and transmission lower when the degree of change of the on-site video image is smaller, and vice versa, use a higher bit rate for on-site video compression and transmission. This method of the present invention can make the adjustment period of the bit rate of on-site video compression adaptively change based on the change of the content of the on-site video, and can effectively reduce the delay and ensure the effectiveness of monitoring when the image in the on-site video changes greatly. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0052] Figure 1 It is a schematic diagram of a remote safety management system for a power construction site of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0054] The present invention provides a remote safety management system for a power construction site, including a shooting device, an edge processing device, and a remote monitoring device;

[0055] The shooting device is used to obtain the on-site video of the power construction site and transmit the on-site video to the edge processing device;

[0056] The edge processing device is arranged at the power construction site. The edge processing device includes a period calculation module, a bit rate calculation module, a compression module, and a transmission module;

[0057] The period calculation module is used to calculate the bit rate adjustment period based on the on-site video, including:

[0058] Use \(n\) to represent the serial number of the current bitrate adjustment period, and calculate the \((n + 1)\)-th bitrate adjustment period based on the on-site video received within the \(n\)-th bitrate adjustment period;

[0059] The bitrate calculation module is used to calculate the bitrate of video compression according to the bitrate adjustment period;

[0060] The compression module is used to compress the on-site video according to the bitrate;

[0061] The transmission module is used to transmit the compressed on-site video to the remote monitoring device;

[0062] The remote monitoring device is used to perform construction safety monitoring on the power construction site based on the compressed on-site video.

[0063] Different from the prior art, when remotely managing the safety of the power construction site, the present invention abandons the traditional fixed bitrate compression and transmission method, but adopts a dynamic adjustment strategy. Specifically, the present invention analyzes the content of the on-site video received within the previous bitrate adjustment period, calculates the duration of the next bitrate adjustment period, and further determines the bitrate of video compression based on this period. This design enables the bitrate adjustment period and the video compression bitrate to adaptively change according to the change of the on-site video content.

[0064] When the change of the on-site video picture is small, the system will compress and transmit at a lower bitrate, thus saving bandwidth resources; while when the picture change is large, the system will automatically increase the bitrate to ensure the clarity and smoothness of the video, and at the same time effectively reduce the latency. In this way, the present invention can dynamically adjust the bitrate and the adjustment period according to the change of the on-site video content, so as to respond in time and optimize the transmission effect when the picture changes significantly, ensuring the real-time and effectiveness of monitoring.

[0065] Furthermore, the remote monitoring device includes a video database, a safety recognition module and a safety warning module;

[0066] The video database is used to store the compressed on-site video transmitted by the transmission module;

[0067] The safety recognition module is used to read the compressed on-site video from the video database and perform intelligent recognition on the compressed on-site video to determine whether there is an alarm event;

[0068] The safety warning module is used to generate a safety notice based on the alarm event and transmit the safety notice to the electronic terminal used by the safety management personnel.

[0069] The safety notice generated by the present invention includes the location where the alarm event occurred and the type of the alarm event (such as not wearing a safety helmet).

[0070] The electronic terminals used by safety management personnel include computers installed in the monitoring center, mobile phones used by safety management personnel, etc.

[0071] Furthermore, compressing the on-site video according to the bit rate includes:

[0072] Using standards such as H.264 or H.265, compress the on-site video according to the bit rate of video compression determined by the bit rate calculation module to obtain the compressed on-site video.

[0073] H.264 is a codec standard based on block-based motion compensation prediction coding and transform coding, combined with intra-frame prediction to reduce redundancy.

[0074] Furthermore, obtaining the on-site video of the power construction site and transmitting the on-site video to the edge processing device includes:

[0075] Continuously shoot the construction site at the set frame rate to obtain a frame sequence, and extract the frame sequence with a set time window to form the on-site video.

[0076] In the present invention, the set frame rate can be 25fps, that is, 25 frames per second.

[0077] The frames in the frame sequence are stored in the order of imaging time from early to late. It can be set to automatically restart the generation of the frame sequence every day, that is, clear the frame sequence at 0:00 every day and store the newly obtained frames into the frame sequence, so as to avoid excessive data occupancy of the frame sequence.

[0078] Furthermore, using T to represent the set time window, then in the frame sequence, extract a frame every T seconds. Using Z to represent the shooting frame rate, the number of consecutive frames included in the on-site video is T×Z.

[0079] In the present invention, the set time window can be 2 seconds, that is, the duration of each on-site video is 2 seconds. By extracting the frame sequence, it is possible to avoid excessive delay of the on-site video due to excessive data sent at one time.

[0080] Furthermore, calculating the (n + 1)th bit rate adjustment period based on the on-site video received within the nth bit rate adjustment period includes:

[0081] Respectively use t s,n and t d,n to represent the start time and end time of the nth bit rate adjustment period;

[0082] After the edge processing device completely receives the on-site video transmitted by the shooting device, record the reception time of the on-site video;

[0083] The reception time within the time range [t s,n ,td,n Sort the on-site videos of [[ ]] in ascending order of reception time to obtain the on-site video sequence sq. n ;

[0084] Based on sq n Calculate the (n + 1)-th bit rate adjustment period.

[0085] The present invention sorts the on-site videos based on the reception time, which is beneficial for subsequent calculation of the bit rate adjustment period based on the latest obtained on-site video, enabling the bit rate adjustment period to be calculated based on the latest on-site video, thereby ensuring the effectiveness of the calculated bit rate adjustment period.

[0086] Furthermore, calculating the (n + 1)-th bit rate adjustment period based on sq includes: n Calculate the (n + 1)-th bit rate adjustment period, including:

[0087] Obtain the on-site video video that is the last in sq n ; lst ;

[0088] Obtain the sequence frasq composed of the (N - s)-th frame to the N-th frame in video, where N represents the total number of frames in video lst ; s is the frame extraction coefficient; lst Calculate the state coefficient state based on frasq;

[0089] Calculate the (n + 1)-th bit rate adjustment period ratcy using the following formula

[0090] : n+1 :

[0091] ratcy n+1 = ratcy pr × state

[0092] ratcy pr represents the set duration.

[0093] The bit rate adjustment period of the present invention can be calculated based on the state coefficient calculated based on frasq. Therefore, the bit rate adjustment period can adaptively change with the change of the picture content in the on-site video. When the state coefficient is larger, the bit rate adjustment period is smaller, and when the state coefficient is smaller, the bit rate adjustment period can be made larger.

[0094] In the present invention, the set duration can be 30 seconds.

[0095] In the present invention, s can be one-tenth of the number of frames in video lst , that is

[0096] Further, the calculation formula for the state coefficient state is as follows:

[0097]

[0098] fram pr is a preset reference value; fram ave represents the average gray value of the pixel points of the frames in frasq; fram i represents the average gray value of the pixel points in frame i in frasq.

[0099] The state coefficient of the present invention can adaptively change according to the similarity degree of the average gray value of the pixel points of the frames in frasq. Therefore, if the difference in the average gray value of the pixel points of the frames in frasq is larger, the state coefficient is smaller. In this way, the calculated bitrate adjustment period can be made smaller, so that the bitrate can be updated more timely to ensure the effectiveness of the compressed on-site time transmitted to the remote monitoring device. That is, when the change amplitude of the picture content in the on-site video is larger, the bitrate adjustment period is shortened and the bitrate is increased in a timely manner, so that the compressed on-site video transmitted to the remote monitoring device contains more detailed content.

[0100] Further, the calculation formula for fram i is as follows:

[0101]

[0102] Further, calculating the bitrate of video compression according to the bitrate adjustment period includes:

[0103]

[0104] comprat n+1 represents the bitrate of video compression used in the (n + 1)-th bitrate adjustment period; frac j represents the change value of the j-th frame in frasq; ws j is the influence coefficient of frac j ; comprath pr represents the preset bitrate; fracm is the median of the change values of all frames in frasq, and λ is the weight.

[0105] When calculating the bit rate of video compression, the present invention calculates from two aspects: the bit rate adjustment period and the change value of the frames. Therefore, when the bit rate adjustment period is smaller and the overall change value of the frames in frasq is larger, the bit rate of the video compression of the present invention is larger. Thus, when the change range of the picture content of the on-site video is larger, the bit rate of the video compression can be made higher, so that the compressed on-site video has more details. On the contrary, when the bit rate adjustment period is larger and the overall change value of the frames in frasq is smaller, the bit rate of the video compression is lower, reducing the space occupied by the compressed on-site video and reducing the storage pressure on the remote monitoring device.

[0106] Further, the change value of the j-th frame is obtained through the following process:

[0107] Determine whether the j-th frame is the last frame in frasq (i.e., determine whether j is equal to s + 1). If so, take the change value of the (j - 1)-th frame as the change value of the j-th frame. If not, obtain the (j + 1)-th frame frac j+1 ;

[0108] Use the following formula to calculate the change value of the j-th frame:

[0109]

[0110] RW and LW are respectively the number of rows and columns of the pixel points included in the frame of the on-site video, pixgra u,v,j and pixgra u,v,j+1 are respectively the gray values of the pixel points at the u-th row and v-th column in frac j and frac j+1 respectively.

[0111] The change value of the present invention is obtained by cumulative summation based on the difference in the gray values of the pixel points at the same position in two adjacent frames. Therefore, when the difference between two adjacent frames is larger, the calculated change value can be made larger, thus reflecting a higher probability of including monitoring targets (such as workers and vehicles at the power construction site) in the monitoring video.

[0112] Further, the weight is 0.6.

[0113] Further, the calculation formula of the influence coefficient is as follows:

[0114]

[0115] frac s+1 is the change value of the (s + 1)-th frame, and η is the weighting parameter.

[0116] The influence coefficient of the present invention is obtained by considering the difference in the change values between the j-th frame and the (s + 1)-th frame in frasq and the magnitude of the value of j. Therefore, when the value of j is larger and the difference in the change values between the j-th frame and the (s + 1)-th frame is smaller, the value of ws j is larger. In this way, it is possible to make the frame with a larger difference in the change value from the change value of the (s + 1)-th frame have a smaller influence on the calculation result of the video compression bitrate as the shooting time is farther from the (s + 1)-th frame. The present invention can preferentially refer to the frames with a later shooting time to calculate the video compression bitrate, so that the calculated bitrate can better adapt to the current construction site situation. Additionally, it can reduce the influence degree of sudden interference on the calculation result. Because if only the two latest obtained frames are considered to calculate the video compression bitrate, when the shooting device is interfered and the difference between the latest obtained frame and the previous frame is too large, it will cause the video compression bitrate to be wrongly increased, increasing the data storage pressure of the remote monitoring device.

[0117] Further, the weighting parameter of the present invention can be 0.5.

[0118] Further, the preset bitrate can be 4 Mbps. The preset bitrate can be set according to the resolution of the on-site video. The higher the resolution, the larger the value.

[0119] Further, the safety identification module includes a reading unit, a screening unit, a preprocessing unit, and an identification unit;

[0120] The reading unit is used to read the compressed on-site video from the video database;

[0121] The screening unit is used to obtain the frames for identification from the compressed on-site video read by the reading unit;

[0122] The preprocessing unit is used to preprocess the frames obtained by the screening unit to obtain the preprocessed frames;

[0123] The identification unit is used to identify the preprocessed frames to determine whether there is an alarm event.

[0124] Further, obtaining the frames for identification from the compressed on-site video read by the reading unit includes:

[0125] Let L represent the sequence of frames in the compressed on-site video read by the reading unit;

[0126] Extract one frame from L every K frames as the frame for identification, where K is the set reading interval.

[0127] By setting the reading interval, it is possible to avoid recognizing all frames. Since the content of frames with close time is generally less changed, obtaining the frames for recognition through the reading interval can not only reduce the pressure of recognition, but also ensure the effectiveness of the recognition process.

[0128] The reading interval of the present invention can be 5.

[0129] Further, preprocess the frames obtained by the screening unit to obtain preprocessed frames, including:

[0130] First, grayscale the frames obtained by the screening unit, and then perform filtering processing to obtain preprocessed frames.

[0131] Through the filtering process, the quality of the frames sent to the recognition unit for recognition can be improved, and the accuracy of the recognition result can be enhanced.

[0132] Further, recognize the preprocessed frames to determine whether there is an alarm event, including:

[0133] It is possible to use trained models such as YOLOv5 and Faster R-CNN to recognize the preprocessed frames to determine whether there is an alarm event.

[0134] The alarm events of the present invention include not wearing a safety helmet, someone appearing in the designated construction restricted area, etc.

[0135] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor limit the invention to the specific embodiments described. Obviously, according to the content of this specification, many modifications and changes can be made. These embodiments are selected and specifically described in this specification to better explain the principles and practical applications of the present invention, so that those skilled in the relevant technical fields can well understand and utilize the present invention. The present invention is only limited by the claims and their full scope and equivalents.

Claims

1. A remote safety management system for electric power construction sites, characterized in that: Including shooting equipment, edge processing equipment and remote monitoring equipment; The shooting equipment is used to obtain on-site videos of the power construction site and transmit the on-site videos to the edge processing equipment; The edge processing equipment is set up at the power construction site, and the edge processing equipment includes a cycle calculation module, a code rate calculation module, a compression module and a transmission module; The cycle calculation module is used to calculate the bit rate adjustment cycle based on the live video, including: n represents the serial number of the current bit rate adjustment period, and the n+1th bit rate adjustment period is calculated based on the live video received in the nth bit rate adjustment period; The bit rate calculation module is used to calculate the bit rate of video compression according to the bit rate adjustment period; The compression module is used to compress the live video according to the bit rate; The transmission module is used to transmit the compressed on-site video to the remote monitoring device; Remote monitoring equipment is used to monitor construction safety at power construction sites based on compressed on-site videos.

2. A remote safety management system for electric power construction sites according to claim 1, characterized in that: The remote monitoring equipment includes a video database, a security identification module and a security alarm module; The video database is used to store the compressed live video transmitted by the transmission module; The security identification module is used to read the compressed on-site video from the video database and perform intelligent identification on the compressed on-site video to determine whether there is an alarm event; The safety alarm module is used to generate safety notifications based on alarm events and transmit the safety notifications to electronic terminals used by safety managers.

3. A remote safety management system for electric power construction sites according to claim 1, characterized in that: Obtain live video of the power construction site and transmit it to edge processing equipment, including: The construction site is continuously photographed at a set frame rate to obtain a frame sequence, which is then extracted in a set time window to form a live video.

4. A remote safety management system for electric power construction sites according to claim 1, characterized in that: Calculating the n+1th bit rate adjustment period based on the live video received in the nth bit rate adjustment period includes: Use t s,n and t d,n Indicates the start time and end time of the nth bitrate adjustment cycle; After the edge processing device completely receives the live video transmitted by the shooting device, it records the time when the live video is received; The receiving time is within the time range [t s,n ,t d,n ] are sorted according to the order of receiving time from early to late, and the live video sequence sq is obtained n ; Based on sq n Calculate the n+1th rate adjustment period.

5. A remote safety management system for electric power construction sites according to claim 4, characterized in that: Based on sq n Calculate the n+1th bitrate adjustment period, including: Get sq n The last live video in the list lst ; Get video lst The sequence from the Nsth frame to the Nth frame in frasq, N represents the video lst The total number of frames contained in , s is the frame extraction coefficient; Calculate the state coefficient state based on frasq; Use the following formula to calculate the n+1th rate adjustment period ratcy n+1 : ratcy n+1 =ratcy pr ×state ratcy pr Indicates the set duration.

6. A remote safety management system for electric power construction sites according to claim 5, characterized in that: The calculation formula of the state coefficient state is: fram pr is the preset control value; ave Represents the average grayscale value of the pixel points in the frame in frasq; i Represents the average grayscale value of the pixels in frame i in frasq.

7. A remote safety management system for electric power construction sites according to claim 6, characterized in that: fram i The calculation formula is:

8. The remote safety management system for electric power construction sites according to claim 5, characterized in that: The bit rate of video compression is calculated according to the bit rate adjustment period, including: comprat n+1 Indicates the video compression bit rate used in the n+1th bit rate adjustment cycle; frac j Indicates the change value of the jth frame in frasq; ws j frac j The influence coefficient of pr represents the preset bit rate; fracm is the median of the change value of all frames in frasq, and λ is the weight.

9. A remote safety management system for electric power construction sites according to claim 5, characterized in that: The security identification module includes a reading unit, a screening unit, a pre-processing unit and an identification unit; The reading unit is used for reading the compressed live video from the video database; The screening unit is used to obtain frames for identification from the compressed live video read by the reading unit; The preprocessing unit is used to preprocess the frame obtained by the screening unit to obtain a preprocessed frame; The recognition unit is used to recognize the pre-processed frames and determine whether there is an alarm event.

10. A remote safety management system for electric power construction sites according to claim 9, characterized in that: Acquire frames for recognition from the compressed live video read by the reading unit, including: L represents a sequence of frames in the compressed live video read by the reading unit; One frame is extracted from L every K frames as the frame for recognition, where K is the set reading interval.

Images