Method and system for detecting number of operator equipment on tower based on iron tower camera

Abstract

The invention discloses an on-tower operator equipment number detection method and system based on an iron tower camera, and the method comprises the following steps: S1, collecting diversified static images, carrying out the preprocessing of the images, and obtaining a to-be-processed image data set; s2, storing the YOLOv8 model with the capability of identifying the signal box and the antenna box; s3, recognizing and counting the signal box and the antenna box in the real-time video stream in combination with a YOLOv8 model and a dynamic counting strategy, and storing a recognizing and counting result in a database; s4, generating corresponding page numbers in different equipment counting dictionaries according to identification counting results in the database, and generating detailed equipment reports according to data recorded by adjacent page numbers in the same equipment counting dictionary; the invention has the beneficial effects that the cost is low and the installation is convenient; the data has accuracy and real-time performance; and history data is accumulated, so that subsequent in-depth analysis of operators is facilitated.

Description

Technical Field

[0001] The present invention relates to target detection technology, and in particular to a method and system for detecting the number of operator equipment on a tower based on a tower camera. Background Art

[0002] With the rapid development of the communications industry, towers, as an important part of base stations, carry a large number of communication equipment, such as antenna boxes and signal boxes. Real-time monitoring and management of these equipment are crucial to ensuring network stability and improving operational efficiency.

[0003] The traditional methods of counting are manual counting and fixed sensor network counting. Manual counting involves technicians conducting on-site inspections and manually recording the number of each tower and its equipment. Although this method can ensure accuracy in small-scale equipment management, it is inefficient when deployed on a large scale. Inspection activities are not only time-consuming, but also not real-time, making it inconvenient for real-time statistics. Fixed sensor networks install various sensors around the towers to monitor the presence and status of equipment in real time. These sensors can provide real-time data feedback to help operators detect equipment failures in a timely manner. However, the initial investment cost of this method is high, especially when installed on multiple towers. Each sensor also requires additional power and network connections, which makes the deployment of fixed sensors complicated and expensive. In addition, the accuracy of the sensors is affected by environmental factors such as weather changes and dust, which may lead to false alarms or missed alarms.

[0004] To this end, the present invention proposes a method and system for detecting the number of operator equipment on a tower based on a tower camera. The method observes and records the changes in the equipment by automatically identifying and classifying objects in the image. Through the accumulation of historical data, operators can conduct in-depth analysis, such as equipment usage trends, failure rates, and maintenance requirements, thereby optimizing resource allocation and maintenance strategies and improving overall operational efficiency. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a method and system for detecting the number of operator equipment on a tower based on a tower camera, which is used to solve the technical problems in the prior art that manual statistics are poor in real time and the deployment of fixed sensors is complex.

[0006] To achieve the above object, the present invention provides a method for detecting the number of operator equipment on a tower based on a tower camera, comprising the following steps: S1. Collect and preprocess a variety of static images to obtain a dataset of images to be processed; S2. Use the YOLOv8 model to train the image data set to be processed, ensure that the YOLOv8 model has the ability to identify devices, specifically, to identify the signal box and the antenna box, and save the YOLOv8 model that has the ability to identify the signal box and the antenna box; S3, combining the YOLOv8 model and the dynamic counting strategy, identifying and counting the signal boxes and antenna boxes in the real-time video stream, and storing the identification and counting results in the database; S4. Generate corresponding page numbers in different device counting dictionaries based on the identification and counting results in the database, and produce a detailed device report based on the data recorded in adjacent page numbers of the same type of device counting dictionary.

[0007] Further: The specific method of S1 is as follows: Obtain static images of the signal box and antenna box taken at different angles, extract images containing only the equipment area from the static images and label them, and classify them into signal box static image dataset and antenna box static image dataset by category; The signal box static image dataset and the antenna box static image dataset are divided into corresponding training sets, validation sets and test sets respectively; The division rule is: randomly select data from the divided data set and classify them, of which 70% of the data is the training set, 15% of the data is the validation set, and 15% of the data is the test set; A variety of static images are collected by cameras installed on the tower and preprocessed to improve data quality. The training set is used for subsequent training models and is the basis for model learning. The model adjusts its parameters (such as weights in the neural network) through training data to find the best fitting curve; the validation set is used to evaluate the performance of the model on unseen data to prevent overfitting; the test set is used to finally evaluate the generalization ability of the model to ensure the performance of the model on real data.

[0008] Further: the specific steps of S2 are as follows: S2-1, the image dataset to be processed obtained in S1 after the YOLOv8 model is imported; S2-2. Configure the hyperparameters of the YOLOv8 model, specifically: the initial learning rate is set to 0.001, the batch size is set to 32, and the training rounds are set to 80; S2-3. The architecture of the YOLOv8 model selects the Nano lightweight minimum version, and combines data enhancement technology, loss function, optimization algorithm, and hyperparameters of the YOLOv8 model to perform multiple rounds of training on the YOLOv8 model to minimize the error between the output of the YOLOv8 model and the real image; S2-4. After each training round, the YOLOv8 model uses the validation set obtained in step S1 to calculate the precision and recall of the model, judge the detection effect of the model, and adjust the initial learning rate and batch size in the hyperparameters of the YOLOv8 model according to the detection effect; repeat the training until the set upper limit of the training round is reached to complete the training; S2-5. After the training is completed, the test set obtained in step S1 is used to bring into the YOLOv8 model to calculate whether the error between the precision and recall rate and the precision and recall rate obtained in S2-4 does not exceed ±0.05%; if yes, the current model parameters are retained; if not, the number of training rounds is increased, and steps S2-3 to S2-5 are repeated until the error between the current precision and recall rate and the precision and recall rate obtained in S2-4 does not exceed ±0.05%, and the current model parameters are retained; In this step, the learning rate is set to avoid oscillations during training and to converge quickly; the batch size is set to speed up the training process while ensuring memory usage, and it is suitable for the computing power of most GPUs; the training epochs is set to 80, and the performance of the validation set is monitored to ensure that the model is fully learned without overfitting.

[0009] Further: The specific method of S3 is as follows: Start the camera installed on the tower to capture the video stream in real time, and use the YOLOv8 model trained in step S2 to classify and count the devices in the video stream; In the process of device classification and counting, to ensure that the same device is counted only once in consecutive frames, the video stream is cut into several frames, and the uniqueness of the frames is determined based on the similarity and timestamp between each frame. The duplicate frames in the several frames cut from the video stream are removed to ensure that the device is identified in different frames; Then, a confidence mechanism is introduced to assign a confidence value to the device identified in each frame, and then the confidence value is evaluated; When the confidence exceeds the preset threshold, it is considered that a new device is detected and the count increases by 1; When the confidence does not exceed the preset threshold, the device is considered to be from a frame that has been detected, and the count remains unchanged; In order to accurately identify and avoid statistical errors in continuous video frames, only the devices in the video stream under a unique frame are counted. When an individual video stream is too long, that is, the picture of its video stream is not the image observed by the camera in one circle, but is larger than the image observed in one rotation, it is possible that the detected device will reappear in the task that needs to be detected. In this regard, a confidence level is given to the device, and a confidence assessment is performed on each detected device to analyze whether the device is a new target. Through this mechanism, the system can effectively distinguish between new targets and repeated targets, thereby improving the accuracy and overall performance of device counting.

[0010] Furthermore, the image containing only the device area in the static image is extracted by removing the background environment image of the non-device area image in the static image by a background subtraction algorithm, and then the acquired device area image is framed with a rectangular frame.

[0011] Furthermore: the data enhancement technology includes flipping, cropping, rotating, and scaling the image, and standardizing the sizes of all images in the image data set to be processed.

[0012] Further: in said S4, the device count dictionary includes a signal box count dictionary and an antenna box count dictionary; The page number is generated by a complete real-time video stream, and is specifically stored in segments according to the time it takes for a camera installed on the tower to rotate one circle, that is, the counting result of the device within the time interval of one rotation of the camera is recorded as a page number; The adjacent page numbers are page numbers generated by the counting results of the device when the two cameras rotate one circle and are stored in sequence; The equipment reports include: equipment increase and decrease reports, equipment failure rate reports and service resource configuration reports.

[0013] Further: the equipment failure rate report is specifically obtained by calculating the frequency of equipment replacement within a set fixed time range; The service resource configuration report is specifically obtained by combining the total number of signal boxes and antenna boxes currently installed on the tower with the user density in the service area of ​​the tower; Since the operator allocates network resources to each region based on the number of Internet users in the region and the number of base stations installed in the region, the present invention can also configure network resources for each base station (signal box and antenna box) in real time based on the number of devices on the tower and the number of monitored Internet users.

[0014] The present invention also provides a tower operator equipment quantity detection system based on a tower camera, comprising: a monitoring camera; The surveillance camera is equipped with a communication module; The communication module communicates wirelessly with the computing device to feed back the video stream to the computing device; The computing device is used to implement the steps and methods described in any one of the above items.

[0015] The beneficial effects of the present invention are: a single tower only needs to correspond to a single camera, which is low-cost and easy to install; the detected equipment and its counting results will be updated in real time to the equipment counting dictionary, and the data is accurate and real-time; historical data is accumulated to facilitate subsequent in-depth analysis by operators. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention and the technical solutions in the prior art, the drawings required for use in the embodiments and the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some implementation methods recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 It is a schematic diagram of the system flow of the present invention; Figure 2 The detection result of the number of operator equipment on the tower of the present invention; Figure 3 This is a diagram of the alarm results generated by the change in the number of operator equipment in the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical scheme and advantages of the present invention more clearly understood, the present invention is described and illustrated below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. Based on the embodiments provided by the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention.

[0019] Obviously, the drawings described below are only some examples or embodiments of the present invention. For ordinary technicians in this field, the present invention can be applied to other similar scenarios based on these drawings without any creative work.

[0020] In addition, it can be understood that although the efforts made in this development process may be complex and lengthy, for ordinary technicians in the field related to the contents disclosed in the present invention, some design, manufacturing or production changes based on the technical contents disclosed in the present invention are just conventional technical means and should not be understood as the contents disclosed in the present invention are insufficient.

[0021] If not otherwise specified, all embodiments and optional embodiments of the present invention can be combined with each other to form a new technical solution.

[0022] See also Figure 1 , the system flow diagram of the present invention comprises the following steps: S1. Collect and preprocess a variety of static images to obtain a dataset of images to be processed; S2. Use the YOLOv8 model to train the image data set to be processed, ensure that the YOLOv8 model has the ability to identify devices, specifically, to identify the signal box and the antenna box, and save the YOLOv8 model that has the ability to identify the signal box and the antenna box; S3, combining the YOLOv8 model and the dynamic counting strategy, identifying and counting the signal boxes and antenna boxes in the real-time video stream, and storing the identification and counting results in the database; S4. Generate corresponding page numbers in different device counting dictionaries based on the identification and counting results in the database, and produce a detailed device report based on the data recorded in adjacent page numbers of the same type of device counting dictionary.

[0023] Specifically: To ensure that the model can accurately identify the signal box and antenna box, a variety of static images are collected by the camera installed on the tower and fed back to the computing device (PC) through wireless communication. The computing device (PC) preprocesses them to improve the data quality. The static images are static images of the signal box and antenna box taken at different angles. The images containing only the equipment area in the static images are extracted and labeled, and classified into signal box static image datasets and antenna box static image datasets by category; and the signal box static image dataset and antenna box static image dataset are divided respectively, specifically divided into corresponding training sets, validation sets and test sets; the division rule is: randomly take data from the divided dataset and classify them, of which 70% of the data is the training set, 15% of the data is the validation set, and 15% of the data is the test set.

[0024] The divided image data set to be processed is imported into the YOLOv8 model for training, and the hyperparameters of the YOLOv8 model are configured. Specifically, in order to avoid oscillation and rapid convergence during training, the initial learning rate is set to 0.001; in order to ensure memory usage and accelerate the training process, and to be suitable for the computing power of most GPUs, the batch size is set to 32; to ensure that the model does not overfit while fully learning, the training rounds are set to 8; the architecture of the YOLOv8 model selects the Nano lightweight minimum version, and combines data enhancement technology, loss function, optimization algorithm, and hyperparameters of the YOLOv8 model to train the YOLOv8 model for multiple rounds, so as to minimize the error between the output of the YOLOv8 model and the real image. Small; at the end of each training round, the YOLOv8 model uses the obtained validation set to calculate the model's precision and recall, judge the model's detection effect, and adjust the initial learning rate and batch size in the YOLOv8 model's hyperparameters according to the detection effect; repeat the training until the set upper limit of the training round is reached to complete the training; after the training is completed, the obtained test set is used to bring the YOLOv8 model to calculate whether the error between the precision and recall and the obtained precision and recall is no more than ±0.05%; if yes, the current model parameters are retained; if not, the number of training rounds is added and the training is repeated until the error between the current precision and recall and the obtained precision and recall is no more than ±0.05%, and the current model parameters are retained.

[0025] Start the camera installed on the tower to capture the video stream in real time. The system controls the camera on the tower to rotate at a fixed angle and time to ensure that the camera can rotate a full circle, thereby recording a video covering the entire tower. The specific process of obtaining the video stream is as follows: First, the system obtains the URL (address) of the video stream corresponding to each camera by calling getVideoRealtimeUrlAPI (video basic information call); the request contains several parameters: streamType (video stream address) is set to 1, indicating that a real-time video stream is requested; protocolType (protocol type) is also set to 1; netType (network type) is set to 1; after sending the request through the callServer (call service) function, the system parses the return result to extract the streamUrl (the read address data). If the return code is 0000, it indicates that the request is successful and the system can proceed to the next step; after successfully obtaining the video stream URL, the system needs to control the camera's pan / tilt (PTZ); this process first locks the pan / tilt and sends a doLockOrReleasePtz (lock and release camera) request with a duration set to 300 seconds; this operation ensures that the camera's position remains stable during the next adjustment; after locking, the system uses the doPtzLocationCmd instruction to adjust the pan / tilt position. The specific parameter settings are as follows: horizontal rotation angle (pan): 9280.0, vertical rotation angle (tilt): 1536.0, zoom level (zoom): 16.0, ensure that the camera can accurately aim at the area to be monitored; after completing the pan / tilt adjustment, the system starts video capture; create an instance and use the RTSP video stream address (streamUrl) obtained previously; when creating the grabber, the system sets a series of parameters, including: rtsp_transport (real-time streaming protocol) is set to TCP, analyzeduration (audio processing duration) is set to 10,000,000 microseconds and probesize (read byte length) is set to 5,000,000 bytes; the width and height of the image are set to 1280 pixels and 720 pixels to adapt to the commonly used high-definition standards; after calling the grabber.start() (video grabber) method, the grabber starts working and prepares to grab real-time video frames; Then, the system creates an instance for video recording. When setting the recorder parameters, the system ensures that the width and height of the recorder are consistent with the grabber, and sets the frame rate to be the same as the grabber. The video encoding format selects H.264, the video bit rate is 3,000,000 bits per second, and the audio part uses MP3 encoding with the number of channels set to 1. After all parameters are adjusted, call recorder.start() (start the recorder) to prepare for video recording.

[0026] Within the set time (15 seconds), the system enters a loop to continuously grab video frames and record them. In each loop, the system calls grabber.grabFrame() to get the current frame, and then records the frame to the specified file through recorder.record(frame); at the end, when the recording time is over, the system needs to stop the grabber and recorder; at this point, call grabber.stop() and recorder.stop() to ensure that all resources are properly released, thus ending the video acquisition and recording operations.

[0027] Use the trained YOLOv8 model to classify and count devices in the video stream. During the device classification and counting process, to ensure that the same device is counted only once in consecutive frames, cut the video stream into several frames, determine the uniqueness of the frames based on the similarity and timestamp between each frame, and remove duplicate frames from the several frames cut from the video stream, so as to ensure that the device is identified in different frames and only count the devices in the video stream under the unique frame. Since the service object of the present invention is not limited to one tower, but the equipment on multiple different towers is observed, it is impossible to ensure that the video stream transmitted back by the camera on each tower is of uniform length (because it is necessary to ensure that all cameras have completely collected the equipment images on the tower, but individual cameras may have slightly faster rotation speeds or different rotation times, resulting in differences in the length of the video streams collected by all cameras). Therefore, when detecting the number of devices in the video stream transmitted back by each camera, the device may appear repeatedly and be repeatedly identified. To avoid repeated identification and counting of devices, the following method is used: A confidence mechanism is introduced to assign a confidence level to a device. A confidence assessment is performed on each detected device to analyze whether the device is a new target. When the confidence level exceeds the preset threshold, it is considered that a new device is detected, and the count increases by 1. When the confidence level does not exceed the preset threshold, it is considered that the device comes from a video stream that has been detected or that the device appears repeatedly, and the count remains unchanged.

[0028] The image containing only the device area in the static image is extracted by removing the background environment image of the non-device area image in the static image using the background subtraction algorithm, and then the acquired device area image is selected with a rectangular frame. The data enhancement technology includes flipping, cropping, rotating, and scaling the image, and the sizes of all images in the image data set to be processed are standardized.

[0029] The device counting dictionary includes the signal box counting dictionary and the antenna box counting dictionary; the page number is generated by a complete real-time video stream, and is stored in segments according to the time it takes for the camera installed on the tower to rotate one circle, that is, the counting result of the device within the time interval of one rotation of the camera is recorded as a page number; adjacent page numbers are page numbers generated by two pages of device counting results of one rotation of the camera stored in sequence; the data stored in each page number is the counting situation of the device when the camera rotates one circle, that is, the device status detected in a complete device monitoring cycle.

[0030] If an abnormal change in the number of devices is detected (such as a sudden decrease in the number of antenna boxes), the system will automatically trigger an alarm to notify maintenance personnel to handle it in time. In addition, all detected devices and their counting results will be updated in real time to the device counting dictionary to ensure the accuracy and timeliness of the data.

[0031] Data analysis is an important step after equipment counting, which aims to use the stored data for in-depth analysis and decision support. All equipment counting and identification results are stored in a structured database for quick query and analysis. Using statistical methods, the frequency of use and failure rate of equipment can be identified, and detailed equipment usage reports can be generated to help operators understand the status of equipment. Through the visual system interface, the analysis results are displayed in the form of charts and dashboards, allowing decision makers to quickly understand the data and take action. Based on these analyses, the system provides optimization suggestions, such as adjusting maintenance plans or resource allocation, to enhance the efficiency and reliability of equipment management. Finally, the data analysis results will be fed back to the system to optimize the equipment counting algorithm and model training process to ensure future identification and counting accuracy.

[0032] See also Figure 2 and Figure 3 , which are respectively the detection result of the number of operator equipment on the tower of the present invention and the alarm result diagram generated by the change of the number of operator equipment of the present invention; from Figure 2 The camera installation location and the number of detected devices can be clearly seen and accurately classified and counted; Figure 3 In the video stream capture clip, we can see the part framed by the frame line, which is the antenna box in the target device to be detected, and the changes in the antenna box and the signal box of the tower are effectively counted; therefore, when there is doubt about the automatic statistical data, the staff can also check the camera feedback video to verify the number of devices by themselves.

[0033] The above is a very detailed description of one or more embodiments of the present invention, but the content is only a specific example of the present invention and cannot be considered to limit the scope of application of the present invention. Any other methods and changes proposed based on the content of the present invention should fall within the scope of patent protection of the present invention.

Claims

1. A method for detecting the number of operator equipment on a tower based on a tower camera, characterized in that: The steps include: S1. Collect and preprocess a variety of static images to obtain a dataset of images to be processed; S2. Use the YOLOv8 model to train the image data set to be processed, ensure that the YOLOv8 model has the ability to identify devices, specifically, to identify the signal box and the antenna box, and save the YOLOv8 model that has the ability to identify the signal box and the antenna box; S3, combining the YOLOv8 model and the dynamic counting strategy, identifying and counting the signal boxes and antenna boxes in the real-time video stream, and storing the identification and counting results in the database; S4. Generate corresponding page numbers in different device counting dictionaries based on the identification and counting results in the database, and produce a detailed device report based on the data recorded in adjacent page numbers of the same type of device counting dictionary.

2. The method for detecting the number of operator equipment on a tower based on a tower camera according to claim 1, characterized in that: The specific method of S1 is as follows: Obtain static images of the signal box and antenna box taken at different angles, extract images containing only the equipment area from the static images and label them, and classify them into signal box static image dataset and antenna box static image dataset by category; The signal box static image dataset and the antenna box static image dataset are divided into corresponding training sets, validation sets and test sets respectively; The division rule is: randomly select data from the divided data set and classify them, wherein 70% of the data is the training set, 15% of the data is the validation set, and 15% of the data is the test set.

3. The method for detecting the number of operator equipment on a tower based on a tower camera according to claim 1, characterized in that: The specific steps of S2 are as follows: S2-1, the image dataset to be processed obtained in S1 after the YOLOv8 model is imported; S2-2. Configure the hyperparameters of the YOLOv8 model, specifically: the initial learning rate is set to 0.001, the batch size is set to 32, and the training rounds are set to 80; S2-3. The architecture of the YOLOv8 model selects the Nano lightweight minimum version, and combines data enhancement technology, loss function, optimization algorithm, and hyperparameters of the YOLOv8 model to perform multiple rounds of training on the YOLOv8 model to minimize the error between the output of the YOLOv8 model and the real image; S2-4. After each training round, the YOLOv8 model uses the validation set obtained in step S1 to calculate the precision and recall of the model, judge the detection effect of the model, and adjust the initial learning rate and batch size in the hyperparameters of the YOLOv8 model according to the detection effect; repeat the training until the set upper limit of the training round is reached to complete the training; S2-5. After the training is completed, the test set obtained in step S1 is used to bring into the YOLOv8 model to calculate whether the error between the precision and recall rate and the precision and recall rate obtained in S2-4 does not exceed ±0.05%; if yes, the current model parameters are retained; if not, the number of training rounds is increased, and steps S2-3 to S2-5 are repeated until the error between the current precision and recall rate and the precision and recall rate obtained in S2-4 does not exceed ±0.05%, and the current model parameters are retained.

4. The method for detecting the number of operator equipment on a tower based on a tower camera according to claim 1, characterized in that: The specific method of S3 is as follows: Start the camera installed on the tower to capture the video stream in real time, and use the YOLOv8 model trained in step S2 to classify and count the devices in the video stream; In the process of device classification and counting, to ensure that the same device is counted only once in consecutive frames, the video stream is cut into several frames, and the uniqueness of the frames is determined based on the similarity and timestamp between each frame. The duplicate frames in the several frames cut from the video stream are removed to ensure that the device is identified in different frames; Then, a confidence mechanism is introduced to assign a confidence value to the device identified in each frame, and then the confidence value is evaluated; When the confidence exceeds the preset threshold, it is considered that a new device is detected and the count increases by 1; When the confidence does not exceed the preset threshold, the device is considered to be from a frame that has been detected, and the count remains unchanged.

5. The method for detecting the number of operator equipment on a tower based on a tower camera according to claim 2, characterized in that: The extraction of the image containing only the device area in the static image is specifically performed by removing the background environment image of the non-device area image in the static image by a background subtraction algorithm, and then selecting the acquired device area image with a rectangular frame.

6. The method for detecting the number of operator equipment on a tower based on a tower camera according to claim 3, characterized in that: The data enhancement technology includes flipping, cropping, rotating, and scaling the image, and standardizing the sizes of all images in the image data set to be processed.

7. The method for detecting the number of operator equipment on a tower based on a tower camera according to claim 1, characterized in that: In said S4, the device count dictionary includes a signal box count dictionary and an antenna box count dictionary; The page number is generated by a complete real-time video stream, and is specifically stored in segments according to the time it takes for a camera installed on the tower to rotate one circle, that is, the counting result of the device within the time interval of one rotation of the camera is recorded as a page number; The adjacent page numbers are page numbers generated by the counting results of the device when the two cameras rotate one circle and are stored in sequence; The equipment reports include: equipment increase and decrease reports, equipment failure rate reports and service resource configuration reports.

8. The method for detecting the number of operator equipment on a tower based on a tower camera according to claim 7, characterized in that: The equipment failure rate report is specifically obtained by calculating the frequency of equipment replacement within a set fixed time range; The service resource configuration report is specifically obtained based on the total number of signal boxes and antenna boxes currently installed on the tower, combined with the user density in the service area of ​​the tower.

9. A tower operator equipment quantity detection system based on a tower camera, characterized in that: include: Security cameras; The surveillance camera is equipped with a communication module; The communication module communicates wirelessly with the computing device to feed back the video stream to the computing device; The computing device is used to implement the steps and methods according to any one of claims 1 to 7.

Images