CN115180522A

CN115180522A - Safety monitoring method and system for hoisting device construction site

Info

Publication number: CN115180522A
Application number: CN202210613319.6A
Authority: CN
Inventors: 朱常玉; 莫绪军; 单建华; 王晶; 成卫琴; 俞琦莺
Original assignee: Pin Ming Technology Co ltd
Current assignee: Pin Ming Technology Co ltd
Priority date: 2022-05-31
Filing date: 2022-05-31
Publication date: 2022-10-14

Abstract

The application relates to safety monitoring of a hoisting device construction site, wherein the method comprises the following steps: acquiring a historical image set of a construction site of the hoisting device, and marking a target object in the historical image set; constructing a detection network and a classification network, respectively training the detection network and the classification network based on the labeled historical image set, and respectively obtaining a detection model and a classification model; collecting a current image of a construction site, detecting a human body object and a lifting hook in the current image through a detection model, and identifying a specific human body object in the human body object through a classification model; determining a target area in the current image according to the hook and the specific human body object in the current image; and tracking the specific human body object in real time, acquiring a coordinate corresponding to the specific human body object in the current image, judging whether the coordinate is in a target area within a preset time period, and if not, indicating to send a safety alarm. Through this application, promoted the safety monitoring ability of tower crane job site, further ensured constructor's safety.

Description

Safety monitoring method and system for hoisting device construction site

Technical Field

The application relates to the field of security and protection, in particular to a safety monitoring method and system for a hoisting device construction site, computer equipment and a computer readable storage medium.

Background

Along with the modernization construction is constantly carried out, the tower crane becomes indispensable mechanical equipment commonly used on the building site, because the tower crane is higher apart from ground, lifts by crane the object quality great, consequently, is crucial to the safety precaution and the warning of lifting by crane the process.

In the related art, chinese patent CN109019335a provides a hoisting safety distance detection method based on deep learning, which describes detection of a hook and a pedestrian under the hook, and determines whether to alarm by calculating an actual distance according to a detection result. Chinese patent CN111062373A provides a hoisting process danger identification method and system based on deep learning, by monitoring a lifting hook, a worker who correctly wears a safety helmet and a worker who does not correctly wear the safety helmet, whether the behavior of the worker meets the safety operation requirement of a hoisting construction site or not is judged, and further, whether the worker is located in the predicted path range of the lifting hook or not can be judged, so that the safety of the hoisting construction site is improved.

However, in the process of hoisting heavy objects, it is not enough to prevent safety accidents, and only a machine is used for detection, so that a signaler still needs to be in the field for supervision all the time, even if an accident happens, the signaler can deal with the accident in time, and if the signaler is separated from a supervision post, a great safety risk is caused.

Disclosure of Invention

The embodiment of the application provides a safety monitoring method, a safety monitoring system, computer equipment and a computer readable storage medium for a hoisting device construction site, and aims to at least solve the problem of poor safety guarantee of a hoisting device monitoring method in the related art.

In a first aspect, an embodiment of the present application provides a safety monitoring method for a hoisting device construction site, where the method includes:

acquiring a historical image set of a construction site of a hoisting device, and labeling a target object in the historical image set, wherein the target object comprises a lifting hook, a specific human body object and a non-specific human body object;

constructing a detection network and a classification network, respectively training the detection network and the classification network based on the labeled historical image set, and respectively obtaining a detection model and a classification model;

collecting a current image of the construction site, detecting a human body object and a lifting hook in the current image through the detection model, and identifying a specific human body object in the human body objects through the classification model;

determining a target area in the current image according to the hook in the current image;

and tracking the specific human body object in real time, acquiring a coordinate corresponding to the specific human body object in the current image, judging whether the coordinate is in the target area within a preset time period, and if not, indicating to send a safety alarm.

In some of these embodiments, the specific human subject is a signaler wearing red headgear, and the non-specific human is a person other than the signaler.

In some embodiments, the detecting the human object and the hook in the current image by a detection model includes:

extracting features in the current image through a backbone network, wherein the backbone network is of a Ghostnet structure,

fusing the features through a neck layer to obtain a feature map, wherein the neck layer comprises a top-down sampling layer and a bottom-up sampling layer, the top-down sampling layer acquires semantic information, the bottom-up sampling layer acquires positioning information, and the features are fused based on the semantic information and the positioning information;

and detecting the characteristic diagram through the detection head to obtain the human body object and the lifting hook in the current image.

In some of these embodiments, the method further comprises:

in the training process of the detection model, the historical data set is expanded through mixup data enhancement,

BCEWithLoitsLoss is used as a classification loss function, and IoU loss is used as a regression loss function.

In some embodiments, the classification model is a mobilenetv3 model, and positive and negative samples are equalized by a focalloss loss function during training of the classification model.

In some embodiments, determining a target region in the current image based on the hook in the current image comprises:

acquiring the pixel length of the hook in the current image and acquiring the actual length of the hook;

determining a conversion ratio of the actual length to the pixel length and a preset moving radius value of the signal processor in a real space;

converting the preset active radius value into a pixel radius value in the current image according to the conversion ratio;

and determining the target area in the current image by taking the center of the lifting hook as a circle center according to the pixel radius value.

In some embodiments, before obtaining the historical image set of the crane construction site, the method further comprises:

the method comprises the steps of collecting a video of a construction site of the lifting device through a camera device, dividing the video into a plurality of groups of single-frame images, and forming the historical image set by the plurality of groups of single-frame images, wherein the camera device is installed on a large arm of the lifting device, and the camera angle is a vertical depression angle.

In a second aspect, an embodiment of the present application provides a safety monitoring system for a hoisting device construction site, the system includes: the device comprises a preprocessing module, a training module and a detection module, wherein the preprocessing module is used for preprocessing the received signal;

the preprocessing module is used for acquiring a historical image set of a construction site of the hoisting device and marking target objects in the historical image set, wherein the target objects comprise a lifting hook, specific human body objects and unspecific human body objects;

the training module is used for constructing a detection network and a classification network, respectively training the detection network and the classification network based on the labeled historical image set, and respectively obtaining a detection model and a classification model;

the detection module is used for collecting the current image of the construction site, detecting the human body object and the lifting hook in the current image through the detection model, identifying the specific human body object in the human body object through the classification model, and

determining a target area in the current image based on the hook in the current image, an

In a third aspect, an embodiment of the present application provides a computer device, which includes a memory, a processor, and a computer program stored on the memory and executable on the processor, and the processor, when executing the computer program, implements the method according to the first aspect.

In a fourth aspect, embodiments of the present application provide a computer-readable storage medium, on which a computer program is stored, which when executed by a processor, implements the method according to the first aspect described above.

Compared with the related technology, the method for monitoring the safety of the construction site provided by the embodiment of the application obtains a plurality of groups of historical images of the construction site and marks the target objects in the historical images; and training a detection model and a classification model based on the plurality of groups of historical images after labeling. Further, in an actual detection link, the human body object and the lifting hook in the current image are detected through the trained detection model and classification model, and the specific human body object in the human body object is identified. And finally, tracking a specific human body object in real time, judging whether the specific human body object is in a target area within a preset time period, if not, indicating that the specific human body leaves the working position and is overtime, and indicating other equipment to send a safety alarm. The problem that safety guarantee of a monitoring method of a lifting device in the related art is poor is solved, detection and tracking of a signal person in charge of supervision are achieved, and an alarm signal is output when the safety person leaves a preset moving range below a lifting hook for a certain time, so that safety monitoring capability of a tower crane construction site is improved, and safety of construction personnel is further guaranteed.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiment(s) of the application and together with the description serve to explain the application and not to limit the application. In the drawings:

fig. 1 is a schematic application environment diagram of a safety monitoring method for a construction site of a hoisting device according to an embodiment of the application;

FIG. 2 is a flow chart of a method of monitoring safety of a lifting apparatus construction site according to an embodiment of the present application;

FIG. 3 is a schematic diagram of feature fusion at the hack layer according to an embodiment of the present application;

FIG. 4 is a block diagram of a safety monitoring system of a lifting device construction site according to an embodiment of the application;

FIG. 5 is a schematic diagram of an internal structure of an electronic device according to an embodiment of the application;

fig. 6 is an internal structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be described and illustrated below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments provided in the present application without any inventive step are within the scope of protection of the present application.

It is obvious that the drawings in the following description are only examples or embodiments of the present application, and that it is also possible for a person skilled in the art to apply the present application to other similar contexts on the basis of these drawings without inventive effort. Moreover, it should be appreciated that such a development effort might be complex and tedious, but would nevertheless be a routine undertaking of design, fabrication, and manufacture for those of ordinary skill having the benefit of this disclosure, given the benefit of this disclosure, without departing from the scope of this disclosure.

Reference in the specification to "an embodiment" means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the specification. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are separate or alternative embodiments mutually exclusive of other embodiments. Those of ordinary skill in the art will explicitly and implicitly appreciate that the embodiments described herein may be combined with other embodiments without conflict.

Unless otherwise defined, technical or scientific terms referred to herein should have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. Reference to "a," "an," "the," and similar words throughout this application are not to be construed as limiting in number, and may refer to the singular or the plural. The use of the terms "including," "comprising," "having," and any variations thereof herein, is meant to cover a non-exclusive inclusion; for example, a process, method, system, article, or apparatus that comprises a list of steps or modules (elements) is not limited to the listed steps or elements, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus. Reference to "connected," "coupled," and the like in this application is not intended to be limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The term "plurality" as referred to herein means two or more. "and/or" describes an association relationship of associated objects, meaning that three relationships may exist, for example, "A and/or B" may mean: a exists alone, A and B exist simultaneously, and B exists alone. The character "/" generally indicates that the former and latter associated objects are in an "or" relationship. Reference herein to the terms "first," "second," "third," and the like, are merely to distinguish similar objects and do not denote a particular ordering for the objects.

The safety monitoring method for the hoisting device construction site can be applied to the application environment shown in fig. 1, fig. 1 is an application environment schematic diagram of the safety monitoring method for the hoisting device construction site according to the embodiment of the application, as shown in fig. 1, the image of the construction site is collected in real time by the camera device 10 installed on the tower crane large arm, further, the server 11 arranged in a monitoring room or a cloud receives the image, the human body objects in the image are identified through the built-in detection model, and the specific human body objects (such as annunciators wearing red safety caps) are identified from all human bodies through the classification model. And finally, tracking the specific human body object in real time, and identifying whether the signaler in charge of safety supervision is in an on-duty state by judging whether the specific human body object is positioned in a preset moving range under the lifting hook, so that the construction safety monitoring capability is optimized, and the safety of constructors is guaranteed.

Fig. 2 is a flowchart of a safety monitoring method for a hoisting device construction site according to an embodiment of the present application, and as shown in fig. 2, the flowchart includes the following steps:

s201, acquiring a video of a construction site of a hoisting device through a camera device, dividing the video into a plurality of groups of single-frame images, and forming a history image set by the plurality of groups of single-frame images, wherein the camera device is arranged on a large arm of the hoisting device, and the camera angle is a vertical depression angle;

in this embodiment, the imaging device may be any one of a conventional camera, an IPC camera, a depth camera, and an infrared camera.

Optionally, the camera device may be fixedly or slidably mounted on the boom of the lifting device, specifically: during sliding installation, the trolley capable of sliding back and forth is arranged on the large arm of the lifting device, and the camera device is installed on the trolley, so that the shooting position is adjusted according to the actual situation on site.

Further, the shooting angle of the camera is a vertical depression angle, and the camera is used for shooting a live image which is less distorted and closer to a real scene.

It should be noted that, in the historical image set acquired in this step, the number of images should meet a certain strength, so as to meet the minimum requirement of model training.

S202, acquiring a historical image set of a construction site of the hoisting device, and labeling target objects in the historical image set, wherein the target objects comprise a lifting hook, specific human body objects and non-specific human body objects;

the historical image set consists of a plurality of single-frame images with different contents, and the single-frame images can be similar or completely different; for example, the human body has clear or fuzzy outlines, large or small outlines, occlusion or non-occlusion, and overlap or non-overlap.

It should be noted that, in the above historical image set, the types of the single-frame images must be abundant, so as to ensure the detection model obtained subsequently, and obtain the identification effect capable of meeting the requirements.

Further, the number of the labeled images should reach a certain order of magnitude, wherein for the deep network model, under the same condition, if the data volume applied in the training process is larger and the data types are richer, the obtained model has a better effect. However, correspondingly, the increase of the data amount will also result in the increase of the computation amount, and the ordinary device is likely not to have the computation requirement. In this example, the labeled specific human body object image and the non-specific human body object image are 1w each, as the case may be.

In this embodiment, the labeling process includes a human body frame labeling process and a classification labeling process, where the human body frame labeling process is to frame out an area covered by a human body in an original image, and a frame range should include the whole part of the human body; further, in the classification labeling process, the labeled human body block diagram area is cut out to obtain a human body block diagram, and then the human body block diagram is classified according to whether the human body block diagram is a specific human body object or not.

Note that, in the present embodiment, the specific human object is a signaler responsible for safety supervision. According to the requirements of the current construction industry specifications, the marked characteristics of the significances of the annunciators are usually expressed as wearing red safety helmets; of course, in some special scenarios, the annunciator may also be provided with other identifying features, such as wearing a blue helmet, wearing a red coat, and the like.

It should be understood that, under the condition that the signaler has the unified identifier and the identifier can be identified and processed by the algorithm model, the technical solution of the present application may utilize the unified identifier to achieve the expected technical effect, and therefore, in the embodiment of the present application, what kind of identifier feature is adopted for the specific human body object is not specifically limited.

S203, constructing a detection network and a classification network, respectively training the detection network and the classification network based on the labeled historical image set, and respectively obtaining a detection model and a classification model;

wherein, above-mentioned detection network combines hoisting accessory job site characteristics on the basis of yolov5 algorithm, obtains through further improving, and is specific:

the backhaul in the original yolov5 algorithm is replaced by a Ghostnet structure, and as the Ghostnet structure is lighter than the backhaul, more characteristics can be generated by using smaller parameters, so that the overall detection speed is improved;

a PAN structure is added in a feature fusion layer of the model, and further feature fusion from top to bottom is carried out on the basis of the original bottom-up feature fusion, so that a better feature map is obtained;

optionally, mixup enhanced data may be adopted during training, a BCEWithLogitsLoss may be adopted as a main classification loss function, an IoU loss may be adopted as a regression loss function, and finally, the output heads are still kept as yolov5 output heads yolo1, yolo2 and yolo3, optionally, 200 epochs (epochs) are trained and iterated.

Further, the classification network may use any common two-class recognition algorithm, and optionally, a mobilenetv3 algorithm model may be used, which is small and fine, and has an input resolution of 224 × 224, in this embodiment, 100 epochs (epochs) are iterated for the classification model training.

S204, collecting a current image of a construction site, detecting a human body object and a lifting hook in the current image through a detection model, and identifying a specific human body object in the human body object through a classification model;

through the above steps S201 to S203, the trained detection model and classification model are obtained, and in this step, the model may be deployed on a construction site for monitoring. The hook and the human body objects in the field real-time image are identified and obtained by using the detection model, and the specific human body objects are identified from the human body objects through the classification model.

It should be noted that, because the image capturing device captures images at a vertical depression angle and the capturing distance is relatively long, the identification feature of a specific human object (such as a red helmet) in the image only occupies a small range in the whole image. Therefore, if the method of directly applying the detection model to detect the specific human body object in the whole picture in the related art is adopted, there is a great risk of false detection, for example: the non-specific human body is recognized as a specific human body, and the non-human body object is recognized as the specific human body object.

In the embodiment of the present application, compared with the related art, the specific human body object is not directly detected from the image, but all human body objects (including the specific human body and the non-specific human body) are first obtained from the real-time image through the detection model to form the human body block diagram, and further, all the human body object block diagrams are input into the classification model, and through the classification model, the two classifications are performed within the smaller range of all the human body object block diagrams to obtain the specific human body object. Therefore, the false detection risk is greatly avoided, and the detection accuracy is improved.

S205, determining a target area in the current image according to the hook in the current image, tracking the specific human body object in real time, acquiring the coordinate of the specific human body object in the current image, judging whether the coordinate is in the target area within a preset time period, and if not, indicating to send a safety alarm.

It should be noted that the process of tracking a specific human body object can be implemented by using kalman filtering in combination with the hungarian algorithm.

Alternatively, the preset time period may be 10 minutes, and when the specific human body object is identified not to be located in the target area under the hook in the image within the preset time period, it is equivalent to that the required time for the annunciator to leave the supervision station has been exceeded.

In this case, the server instructs the alarm device, which may be a voice alarm device such as a jobsite radio, a portable communications device, etc., to send a safety alarm to prompt the annunciator to quickly return to a supervisory station, or to notify the administrator for further decision making.

Through the steps S201 to S205, compared with a safety detection method of a hoisting device construction site in the related art, all human body objects are obtained through the detection model, and furthermore, the signalers wearing red safety caps and responsible for construction supervision are identified from all the human body objects through the classification model. Therefore, after the fact that the annunciator exceeds the moving range below the lifting hook for a certain time is recognized, an alarm signal is sent to remind the annunciator to rapidly return to the supervision position. The embodiment of the application realizes the identification of specific crowds, improves the safety monitoring capability of the tower crane construction site, and further guarantees the safety of constructors.

In some of these embodiments, the process of detecting the target object by the detection model includes: extracting features in the historical image or the current image through a backbone network, wherein the backbone network is of a Ghostnet structure, fusing the features through a neck layer to obtain a feature map, the neck layer is of a PAN structure, and detecting the feature map through a detection head to obtain a target object.

The Ghost module can be used as a plug-and-play component to upgrade the existing convolutional neural network, and the stack based on the Ghost module establishes a lightweight Ghost network structure, specifically:

in the Ghostnet network structure, a common convolution layer is divided into two parts, wherein the first part relates to common convolution, but the total number of the convolution layer is strictly controlled; further, given the intrinsic profiles of the first portion, a simple linear operation is applied to generate more profiles. Thus, the total number of parameters and computational complexity required for the Ghostnet network architecture is reduced compared to conventional Bckbone convolutional neural networks without changing the size of the output feature map.

Further, fig. 3 is a schematic diagram of feature fusion at the hack layer according to an embodiment of the present application, as shown in fig. 3,

the traditional FPN structure only executes top-down sampling, and a feature map is obtained by fusing upper-layer features and lower-layer features.

In the present embodiment, a bottom-up feature pyramid structure (PAN structure) is added on the basis of the FPN layer. Through the combination operation on two kinds of sampling layers, by FPN layer from the top downward conveying strong semantic feature, the pyramid structure then from the bottom upward conveying locate feature, combine together semantic feature and locate feature, carry out parameter polymerization to the feature that different detection layers obtained from different backbone layers to obtain the better characteristic map of effect.

Fig. 4 is a structural diagram of a detection model according to an embodiment of the present application, and as shown in fig. 4, a construction site image is input into the model, a feature map is obtained by extracting features through Ghostnet and PAN fusion features, and output yolo1, yolo2 and yolo3 are obtained by processing through a detection head.

In some embodiments, in the training process of the detection model, the history data set is extended by the mixup enhanced data, and further, bcewithlogitss is used as a classification loss function, and IoU loss is used as a regression loss function.

It should be noted that, due to the image corresponding to the actual construction process, there are various different types of situations, specifically, there are human body occlusion or human body overlapping images, images with a long or short human body distance, images with a large or small human body contour, and the like. And because the number of the collected and labeled images is limited, the historical data set for training can not cover all the construction sites, so that the image data similar to but not identical with the original image is generated by fusing and overlapping the existing images through a data enhancement mechanism, thereby filling the blank among the existing images, enriching the diversity of the data and improving the generalization capability of the model.

In the embodiment of the application, by using the property of mixup data enhancement, the diversity of data is enriched, and meanwhile, the statistical mean and variance can be better when BN (Batch Normalization) operation is performed.

In some embodiments, determining the target region in the current image based on the hook in the current image comprises: acquiring the pixel length of the lifting hook in the current image and acquiring the actual length of the lifting hook; determining a conversion ratio of the actual length to the pixel length and a preset movable radius value of the signal worker in a real space; converting the preset movable radius value into a pixel radius value in the current image according to the conversion ratio; and determining a target area in the current image by taking the center of the lifting hook as a circle center according to the pixel radius value.

It should be noted that the steps illustrated in the above-described flow diagrams or in the flow diagrams of the figures may be performed in a computer system, such as a set of computer-executable instructions, and that, although a logical order is illustrated in the flow diagrams, in some cases, the steps illustrated or described may be performed in an order different than here.

The embodiment also provides a safety monitoring system for the construction site of the lifting device, which is used for implementing the above embodiments and preferred embodiments, and the description of the system is omitted. As used hereinafter, the terms "module," "unit," "subunit," and the like may implement a combination of software and/or hardware for a predetermined function. Although the means described in the embodiments below are preferably implemented in software, an implementation in hardware, or a combination of software and hardware is also possible and contemplated.

Fig. 5 is a block diagram of a safety monitoring system of a hoisting device construction site according to an embodiment of the present application, and as shown in fig. 5, the system includes: a preprocessing module 50, a training module 51, and a detection module 52, wherein;

the preprocessing module 50 is configured to obtain a historical image set of a construction site of the hoisting device, and label a target object therein, where the target object includes a hook, a specific human body object, and an unspecific human body object;

the training module 51 is configured to construct a detection network and a classification network, train the detection network and the classification network based on the labeled historical image set, and obtain a detection model and a classification model respectively;

the detection module 52 is configured to collect a current image of a construction site, detect a human body object and a hook in the current image through a detection model, identify a specific human body object in the human body object through a classification model, determine a target area in the current image according to the hook and the specific human body object in the current image, track the specific human body object in real time, obtain a coordinate corresponding to the specific human body object in the current image, determine whether the coordinate is in the target area within a preset time period, and instruct to send a safety alarm if the coordinate is not in the target area within the preset time period.

In one embodiment, a computer device is provided, which may be a terminal. The computer device comprises a processor, a memory, a network interface, a display screen and an input device which are connected through a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device comprises a nonvolatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by the processor to realize the safety monitoring method of the construction site of the hoisting device. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment can be a touch layer covered on the display screen, a key, a track ball or a touch pad arranged on the shell of the computer equipment, an external keyboard, a touch pad or a mouse and the like.

In an embodiment, fig. 6 is a schematic internal structure diagram of an electronic device according to an embodiment of the present application, and as shown in fig. 6, there is provided an electronic device, which may be a server, and its internal structure diagram may be as shown in fig. 6. The electronic device comprises a processor, a network interface, an internal memory and a non-volatile memory connected by an internal bus, wherein the non-volatile memory stores an operating system, a computer program and a database. The processor is used for providing calculation and control capacity, the network interface is used for being connected and communicated with an external terminal through a network, the internal memory is used for providing an environment for the operation of the operating system and the computer program, the computer program is executed by the processor to realize the safety monitoring method for the construction site of the hoisting device, and the database is used for storing data.

Those skilled in the art will appreciate that the configuration shown in fig. 6 is a block diagram of only a portion of the configuration associated with the present application, and does not constitute a limitation on the electronic device to which the present application is applied, and a particular electronic device may include more or less components than those shown in the drawings, or may combine certain components, or have a different arrangement of components.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in the embodiments provided herein may include non-volatile and/or volatile memory, among others. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), rambus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

The above examples only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the invention. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present patent shall be subject to the appended claims.

Claims

1. A safety monitoring method for a construction site of a hoisting device is characterized by comprising the following steps:

2. The method of claim 1, wherein the specific human subject is an annunciator wearing red safety helmets, and the non-specific human subject is a person other than the annunciator.

3. The method of claim 1, wherein detecting the human object and the hook in the current image by a detection model comprises:

4. The method of claim 1, further comprising:

5. The method of claim 1, wherein the classification model is a mobilenetv3 model, and wherein positive and negative samples are equalized by a focalloss loss function during the training of the classification model.

6. The method of claim 2, wherein determining a target region in the current image based on the hook in the current image comprises:

and determining the target area in the current image according to the pixel radius value by taking the center of the lifting hook as a circle center.

7. The method of claim 1, wherein prior to obtaining the historical image set of the crane construction site, the method further comprises:

8. A safety monitoring system for a hoisting device construction site, the system comprising: the device comprises a preprocessing module, a training module and a detection module, wherein the preprocessing module is used for preprocessing the signal;

the training module is used for constructing a detection network and a classification network, and respectively training the detection network and the classification network based on the labeled historical image set to respectively obtain a detection model and a classification model;

determining a target area in the current image based on the hook and the specific human object in the current image, an

9. A computer device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method according to any of claims 1 to 7 when executing the computer program.

10. Readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1 to 7.