Construction scene identification method and device, electronic equipment and readable storage medium

Abstract

The embodiment of the invention provides a construction scene recognition method and device, electronic equipment and a readable storage medium, and the method comprises the steps: recognizing a target construction element contained in a to-be-recognized scene image, and determining a first recognition result based on the target construction element and an element number corresponding to the target construction element; under the condition that the first identification result is to be verified, acquiring a first scene image; an interval between first acquisition time corresponding to the first scene image and second acquisition time corresponding to the to-be-recognized scene image is a preset duration; and determining a target recognition result based on the first construction element contained in the first scene image. Thus, the construction clues are prevented from being collected, integrated and analyzed manually, the construction scene can be more comprehensively and accurately obtained and judged through the embodiment of the invention, the reliability of the construction scene recognition result is improved, and more effective support is provided for monitoring and management of the construction scene.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of engineering construction technology, and in particular to a construction scene recognition method and device, electronic equipment and readable storage medium. BACKGROUND

[0002] With the acceleration of urban construction, the number and scale of construction projects continue to grow, and the safety of construction sites has become a focus. Construction projects involve construction units, construction units and supervision units, and other parties, known as "three-party construction". In the complex construction site environment, it is crucial to ensure the standardization and safety of construction activities.

[0003] In related technologies, in the current construction, the collection, integration and analysis of construction clues often rely on manual operation, which not only consumes time and effort, but also easily leads to decision-making errors due to information transmission lag or errors. SUMMARY

[0004] To overcome the problems in the related art, the present application provides a construction scene recognition method, device, electronic equipment and readable storage medium.

[0005] In a first aspect, the present application provides a construction scene recognition method, comprising:

[0006] identifying a target construction element contained in a to-be-recognized scene image, and determining a first recognition result based on the target construction element and the number of elements corresponding to the target construction element;

[0007] In the case where the first recognition result is to be verified, a first scene image is obtained; the first acquisition time corresponding to the first scene image and the second acquisition time corresponding to the to-be-recognized scene image are separated by a preset time length;

[0008] determining a target recognition result based on a first construction element contained in the first scene image.

[0009] Optionally, the first recognition result is determined based on the target construction element and the number of elements corresponding to the target construction element, comprising:

[0010] In the case where the target construction element contains a fence element, if the fence recognition result corresponding to the fence element is a construction fence, the first recognition result is determined to be a construction scene;

[0011] In the case where the target construction element does not contain the fence element or the fence recognition result is not a construction fence, if the number of element types of the target construction element is at least two, the first recognition result is determined to be a construction scene;

[0012] In a case where the element type quantity of the target construction element is less than two, if the target construction element contains a dynamic construction element, the first recognition result is determined to be to be verified.

[0013] Optionally, after the target construction element contained in the to-be-recognized scene image is recognized, the method further comprises:

[0014] acquiring a historical false positive image corresponding to the to-be-recognized scene image, and recognizing a second construction element contained in the historical false positive image;

[0015] In a case where the second construction element contains a fence element, the second construction element is compared with the target construction element;

[0016] In a case where the second construction element does not satisfy a pre-judgment condition, a recognition result for the to-be-recognized scene image is determined to be a non-construction scene; the pre-judgment condition is that the second construction element is different from the target construction element and the element quantity of the target construction element is greater than the element quantity of the second construction element;

[0017] In a case where the second construction element does not contain a fence element, or the second construction element contains a fence element and the second construction element satisfies the pre-judgment condition, the step of determining the first recognition result based on the target construction element and the element quantity corresponding to the target construction element is performed.

[0018] Optionally, the determining of the target recognition result based on the first construction element contained in the first scene image comprises:

[0019] In a case where the first construction element contains only a fence element, if a fence recognition result corresponding to the fence element is a construction fence, the target recognition result is determined to be a construction scene;

[0020] In a case where the first construction element does not contain a fence element or contains a fence element and other elements at the same time, the target recognition result is determined to be a construction scene;

[0021] In a case where the first construction element is empty, or the fence recognition result corresponding to the fence element is a non-construction fence, the target recognition result is determined to be a non-construction scene.

[0022] In a second aspect, the present application provides a construction scene recognition device, which comprises:

[0023] a first recognition module, configured to recognize a target construction element contained in a to-be-recognized scene image, and determine a first recognition result based on the target construction element and an element quantity corresponding to the target construction element;

[0024] The first acquisition module is configured to acquire a first scene image when the first recognition result is to be verified, wherein a first acquisition time corresponding to the first scene image and a second acquisition time corresponding to the scene image to be recognized are separated by a preset time length.

[0025] The first determination module is configured to determine a target recognition result based on a first construction element contained in the first scene image.

[0026] Optionally, the first recognition module comprises:

[0027] The second determination module is configured to, when the target construction element contains a fence element, determine that the first recognition result is a construction scene if a fence recognition result corresponding to the fence element is a construction fence.

[0028] The third determination module is configured to, when the target construction element does not contain the fence element or the fence recognition result is not a construction fence, determine that the first recognition result is a construction scene if the number of element types of the target construction element is at least two.

[0029] The fourth determination module is configured to, when the number of element types of the target construction element is less than two, determine that the first recognition result is to be verified if the target construction element contains a dynamic construction element.

[0030] Optionally, the apparatus further comprises:

[0031] The second acquisition module is configured to acquire a historical false positive image corresponding to the scene image to be recognized, and recognize a second construction element contained in the historical false positive image.

[0032] The first comparison module is configured to compare the second construction element with the target construction element when the second construction element contains a fence element.

[0033] The fifth determination module is configured to determine that the recognition result for the scene image to be recognized is a non-construction scene when the second construction element does not satisfy a pre-judgment condition, wherein the pre-judgment condition is that the second construction element is different from the target construction element and the number of elements in the target construction element is greater than the number of elements in the second construction element.

[0034] The first execution module is configured to execute the step of determining the first recognition result based on the target construction element and the number of elements corresponding to the target construction element when the second construction element does not contain a fence element, or when the second construction element contains a fence element and the second construction element satisfies the pre-judgment condition.

[0035] Optionally, the first determining module comprises:

[0036] The sixth determining module is configured to, in a case where the first construction element only comprises the fence element, if the fence identification result corresponding to the fence element is a construction fence, determine that the target identification result is a construction scene.

[0037] The seventh determining module is configured to, in a case where the first construction element does not comprise the fence element or comprises the fence element and other elements at the same time, determine that the target identification result is a construction scene.

[0038] The eighth determining module is configured to, in a case where the first construction element is empty, or the fence identification result corresponding to the fence element is a non-construction fence, determine that the target identification result is a non-construction scene.

[0039] In a third aspect, the present application provides an electronic device, comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor implements the construction scene identification method of any one of the first aspect when executing the program.

[0040] In a fourth aspect, the present application provides a readable storage medium, when the instructions in the storage medium are executed by the processor of an electronic device, the electronic device can execute the steps in the construction scene identification method of any one of the first aspect.

[0041] In the embodiments of the present application, the target construction element contained in the to-be-identified scene image is identified, and the first identification result is determined based on the target construction element and the number of elements corresponding to the target construction element; in a case where the first identification result is to be verified, a first scene image is acquired; the first acquisition time corresponding to the first scene image and the second acquisition time corresponding to the to-be-identified scene image are separated by a preset time length; and the target identification result is determined based on the first construction element contained in the first scene image. In this way, the target construction element in the to-be-identified scene image is identified first, and the first identification result is determined by comprehensively considering the elements and the number thereof, so that various information in the image is fully utilized, and the scene is analyzed from multiple dimensions. When the first identification result is to be verified, the first scene image collected at an interval of a preset time length is introduced for re-identification. Since the construction scene is dynamic, the images collected at different times can reflect the change of the scene. Therefore, the target identification result is determined by using the first construction element in the first scene image, which avoids misjudgment caused by insufficient image information at a single time, so that the final judgment is more in line with the actual scene state. Furthermore, the construction clues are collected, integrated and analyzed by using the artificial method, and the construction scene can be more comprehensively and accurately known and judged by using the embodiments of the present application, so that the reliability of the construction scene identification result is improved, and more effective support is provided for the monitoring and management of the construction scene. BRIEF DESCRIPTION OF DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed to be used in the embodiments or prior art description will be briefly introduced. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort based on these drawings.

[0043] Figure 1 is a step flow chart of a construction scene recognition method provided by an embodiment of the present application;

[0044] Figure 2 is a specific step flow chart of a construction scene recognition method provided by an embodiment of the present application;

[0045] Figure 3 is a structure diagram of a construction scene recognition device provided by an embodiment of the present application;

[0046] Figure 4 is a structure diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0047] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort fall within the scope of the present application.

[0048] Figure 1 is a step flow chart of a construction scene recognition method provided by an embodiment of the present application, as shown in Figure 1 The method can include:

[0049] Step 101, recognizing a target construction element contained in a to-be-recognized scene image, and determining a first recognition result based on the target construction element and an element quantity corresponding to the target construction element.

[0050] In the embodiment of the present application, in order to realize real-time discovery and monitoring of the construction site of the construction project, an image acquisition device can be deployed to obtain real-time scene images of the site, and an image recognition model can be used to analyze the scene images to determine whether the scene images meet the characteristics of the construction activities, and then determine whether the scene images are construction scenes. When the scene images are identified as construction scenes, an alarm mechanism can be triggered to remind the supervisors to pay attention, so as to take appropriate management measures. The image acquisition device can be a camera, and the image recognition model can be obtained by training a large number of historical construction images after labeling. The image recognition model has the ability to identify various construction elements and determine construction scenes. When labeling the historical construction images, a standardized labeling guide can be established, and an identification specification can be developed in conjunction with the business party. The standardized labeling guide is used to define the labeling standards and rules of various elements in the construction scene images, to ensure that the labeling results of the same image by different labelers are consistent and accurate. The identification specification developed in conjunction with the business party fully considers the actual business needs and scene characteristics, making the identification process more suitable for practical applications. Through this way, the identification quality is guaranteed from two aspects of data labeling and identification rules, and the reliability and usability of the identification results are improved.

[0051] The scene images collected by the image acquisition device are obtained in real time as the to-be-identified scene images. The target construction elements contained in the to-be-identified scene images are identified by the image recognition model. The target construction elements can include fences, engineering vehicles, safety helmets, work clothes, and pipelines, etc. Specifically, when the image recognition model identifies, the image features in the to-be-identified scene images can be matched with the features of the preset construction elements to determine the target construction elements contained in the to-be-identified scene images. The target construction elements can include all or part of the construction elements of the preset construction elements. The preset construction elements can include elements closely related to construction, such as fences, engineering vehicles, safety helmets, work clothes, and pipelines, etc. The construction elements can be divided into dynamic construction elements and static construction elements. The dynamic construction elements can include but are not limited to engineering vehicle elements, safety helmet elements, and work clothes elements. The static construction elements can include but are not limited to pipeline elements.

[0052] After the target construction element in the to-be-identified scene image is identified, a first identification result is determined based on the target construction element and the number of elements corresponding to the target construction element. The first identification result is determined by judging whether the target construction element contains a fence element, whether the target construction element contains at least two construction elements other than the fence element, and whether the target construction element contains at least one dynamic construction element. The first identification result can include a construction scene, a non-construction scene, and to-be-verified. Specifically, when the target construction element contains a fence element and the fence element is a construction element, the first identification result is determined to be a construction scene. When the target construction element does not contain a fence element but contains at least two construction elements other than the fence element, the first identification result is determined to be a construction scene. When the target construction element does not contain a fence element and does not contain at least two construction elements other than the fence element, but contains at least one dynamic construction element, the first identification result is determined to be to-be-verified.

[0053] Optionally, step 101 can include the following steps:

[0054] In the case that the target construction element contains a fence element, if the fence identification result corresponding to the fence element is a construction fence, the first identification result is determined to be a construction scene.

[0055] In the construction element identification process in the embodiment of the application, the fence element is often misjudged due to factors such as the diversity of its appearance form and the complexity of the environment in which it is located. For example, some temporary obstructions and advertising fence elements in non-construction scenes may be very similar to construction fences in image features. Traditional image recognition models are difficult to accurately distinguish in such cases, resulting in errors in construction scene judgment. Therefore, in the process of construction scene identification, the embodiment of the application first independently judges the fence element. In the case that the target construction element contains a fence element, the type of the fence element is further judged to determine whether the fence type is a construction fence. Specifically, by introducing an image-text recognition link, the image-related text or possible text information in the image of the to-be-identified scene image is analyzed to determine whether it contains a preset keyword related to construction, such as the words “construction”, “engineering”, and “construction”.

[0056] In the case that the target construction element contains the fence element, further optical character recognition is performed on the identified fence element. The optical character recognition can be performed in the following manner: whether the fence element contains text information is obtained, the text information is compared with a preset keyword, and in the case that the text information matches the preset keyword, the fence element is determined as a construction fence. It can be understood that the specific manner of determining the type of the fence element can be set according to actual requirements. For example, whether the type of the fence element is a construction fence can be determined based on specific colors and specific styles of the construction fence, and the embodiments of the present application do not limit this.

[0057] Since the construction fence is usually a clear sign of construction activities, its existence often means that construction activities are being carried out or will be carried out soon, so once the fence element is determined as a construction fence, the current scene can be quickly and accurately determined as a construction scene.

[0058] In the case that the target construction element does not contain the fence element or the fence recognition result is not a construction fence, if the number of element types of the target construction element is at least two, the first recognition result is determined as a construction scene.

[0059] In the embodiments of the present application, if the target construction element does not contain the fence element or although it contains the fence element but the fence recognition result is not a construction fence, further judgment is performed according to the number of element types of the target construction element. The element type refers to different construction element categories such as engineering vehicles, safety helmets, work clothes, and pipe materials. When the number of element types of the target construction element is at least two, i.e., two or more different types of construction elements exist, such as engineering vehicles and safety helmets, it is indicated that the scene has multiple elements closely related to construction activities. From the probability and actual situation, the scene is very likely to be a construction scene, so the first recognition result can be determined as a construction scene. For example, engineering vehicles and work clothes appear at the same time, indicating that construction equipment and construction personnel are present, which meets the characteristics of the construction scene. This judgment manner fully utilizes the comprehensive information provided by the simultaneous appearance of multiple construction elements, and improves the accuracy of the judgment of the construction scene.

[0060] In the case that the number of element types of the target construction element is less than two, if the target construction element contains a dynamic construction element, the first recognition result is determined as to be verified.

[0061] In the embodiment of the present application, if the target construction element does not contain a fence element, or although it contains a fence element, it is judged that the fence recognition result is not a construction fence, and the number of element types of the target construction element is less than two, at this time, it is further needed to judge whether the target construction element contains a dynamic construction element. The dynamic construction element can include an engineering vehicle, a safety helmet and a work uniform, and the dynamic construction element is often closely related to the dynamic activity in the construction process. For example, the movement of the engineering vehicle, the construction personnel wearing the safety helmet for operation, etc. If the target construction element contains the dynamic construction element, due to the existence of the dynamic construction element, it may mean that the construction activity is in the process, but only according to the current information, it is not enough to completely determine the construction scene, so it can be determined that the first recognition result is to be verified, and subsequent further analysis and judgment can be combined with other time collected image information.

[0062] In the case that the target construction element does not contain a fence element or the fence recognition result is not a construction fence, if the number of element types of the target construction element is less than two, and the target construction element does not contain a dynamic construction element, it can be determined that the first recognition result is a non-construction scene.

[0063] In the embodiment of the present application, by separately analyzing the fence element and combining the image-text recognition to judge whether it is a construction fence, the construction scene can be more accurately recognized, and the misjudgment can be reduced. At the same time, considering the number of types of the target construction element, when the number of element types is at least two, it is determined to be a construction scene, which comprehensively considers the information of multiple construction elements, so that the judgment is more reliable. For the case that the number of element types is small but contains a dynamic construction element, it is set to be verified, which leaves room for further verification in the future, and avoids incorrect judgment due to insufficient information.

[0064] Step 102, in the case that the first recognition result is to be verified, a first scene image is acquired; the first acquisition time corresponding to the first scene image and the second acquisition time corresponding to the to-be-recognized scene image are separated by a preset time length.

[0065] In the embodiment of the present application, after obtaining the first recognition result, the first recognition result is judged. If the first recognition result is to be verified, it is needed to enter a specific process for further verification, that is, it is needed to perform the identification operation again after a certain time interval. At this time, the first scene image is acquired, the first acquisition time corresponding to the first scene image and the second acquisition time corresponding to the to-be-recognized scene image are separated by a preset time length, and the first acquisition time is after the second acquisition time. The preset time length can be determined comprehensively according to the change law of the actual construction scene and the identification demand and other factors, for example, the preset time length can be set to 8 minutes, so as to ensure that the scene image with reference value is acquired after a suitable time interval. For example, 8 minutes after the to-be-recognized scene image is acquired, the first scene image is acquired again through the image acquisition device.

[0066] In step 103, a target recognition result is determined based on first construction elements contained in the first scene image.

[0067] In the embodiment of the present application, after the first scene image is acquired, the first scene image is recognized to identify the first construction elements contained therein. The first construction elements also cover various elements related to the construction scene, which can be similar to the target construction elements, including fences, engineering vehicles, safety helmets, work clothes, and pipes. Based on the first construction elements identified in the first scene image, the target recognition result is further determined. For example, if multiple construction elements are identified in the first scene image, the final target recognition result is determined in combination with the previous identification information, which can be used to determine the specific state of the construction scene, whether it is a construction fence related situation, etc., to provide a basis for subsequent decision-making or operation.

[0068] By combining the scene images collected at different times, the element situation in the construction scene can be more comprehensively and accurately determined, and the reliability of the recognition result can be improved to provide more effective support for the monitoring and management of the construction scene.

[0069] In summary, in the embodiment of the present application, the target construction elements contained in the to-be-recognized scene image are identified, and the first identification result is determined based on the target construction elements and the number of elements corresponding to the target construction elements. In the case that the first identification result is to be verified, the first scene image is acquired. The first acquisition time corresponding to the first scene image and the second acquisition time corresponding to the to-be-recognized scene image are separated by a preset time interval. The target recognition result is determined based on the first construction elements contained in the first scene image. In this way, the target construction elements in the to-be-recognized scene image are first identified, and the first identification result is determined by comprehensively considering the elements and their quantities, which fully utilizes the various information in the image and analyzes the scene from multiple dimensions. When the first identification result is to be verified, the first scene image collected at an interval of a preset time is introduced for re-recognition. Since the construction scene is dynamic, the images collected at different times can reflect the changes in the scene. Therefore, the target recognition result is determined based on the first construction elements in the first scene image, which avoids misjudgment caused by insufficient image information at a single time, and makes the final judgment more consistent with the actual scene state. Furthermore, by avoiding the collection, integration, and analysis of construction clues by manual means, the construction scene can be more comprehensively and accurately known and determined by the embodiment of the present application, the reliability of the construction scene recognition result is improved, and more effective support is provided for the monitoring and management of the construction scene.

[0070] The embodiment of the present application can combine reasonable image acquisition device layout, image recognition model training development, real-time monitoring and analysis, detailed information recording and emergency plan designation, clearly determine the response measures when the third-party construction activity is found, form a closed loop from capturing to disposing the construction scene, and effectively improve the management level and safety of the construction site.

[0071] Optionally, after "identifying the target construction element contained in the to-be-identified scene image", the embodiment of the present application can further include the following steps:

[0072] Step 301: acquiring a historical false alarm image corresponding to the to-be-identified scene image, and identifying a second construction element contained in the historical false alarm image.

[0073] In the embodiment of the present application, the historical false alarm image corresponding to the to-be-identified scene image can include an image that is wrongly determined as a construction scene when identifying a similar scene in the past, for example, an image that is wrongly determined as a construction scene by the image acquisition device corresponding to the to-be-identified scene image before. That is, the historical false alarm image is an image that is determined to be a construction scene but is actually a non-construction scene after verification. Taking the historical false alarm image as a reference can quickly determine whether the to-be-identified scene image is a construction scene.

[0074] The historical false alarm image is acquired and the second construction element contained in the historical false alarm image is identified. The second construction element can include an engineering vehicle, a safety helmet, a work uniform, a fence and other elements related to a construction scene.

[0075] Step 302: comparing the second construction element with the target construction element in the case that the second construction element contains a fence element.

[0076] In the embodiment of the present application, the second construction element can be compared with the target construction element identified from the to-be-identified scene image in the case that the second construction element contains a fence element. The comparison includes the types, quantities and positions of the elements in the image and other information. Through this comparison, the difference between the construction element situation in the current to-be-identified scene image and the construction element situation in the historical false alarm image can be analyzed, thereby providing a reference for determining whether the current scene is a true construction scene.

[0077] Step 303: determining that the identification result for the to-be-identified scene image is a non-construction scene in the case that the second construction element does not satisfy a pre-judgment condition; the pre-judgment condition is that the second construction element is different from the target construction element and the number of elements in the target construction element is greater than the number of elements in the second construction element.

[0078] In the embodiment of the present application, if the second construction element does not meet the pre-judgment condition, it is determined that the recognition result for the to-be-recognized scene image is a non-construction scene. The pre-judgment condition can be that the second construction element is different from the target construction element, and the number of elements in the target construction element is greater than the number of elements in the second construction element. The pre-judgment condition can be set based on in-depth analysis of the dynamic change law of the construction scene.

[0079] If the second construction element does not meet the pre-judgment condition, that is, the second construction element is the same as the target construction element, or the second construction element is different from the target construction element, but the number of elements in the target construction element is not greater than the number of elements in the second construction element, it can be directly determined that the recognition result for the to-be-recognized scene image is a non-construction scene.

[0080] From the perspective of the stability of the construction scene, if the second construction element is the same as the target construction element, it means that the construction element in the scene does not change at different times or in different identification stages, and remains relatively stable. In a normal non-construction scene, the change of construction elements does not occur frequently, so the historical false image represents a non-construction scene, and the stability of the second construction element and the target construction element meets the characteristics of the non-construction scene, so it can be directly determined that the recognition result for the to-be-recognized scene image is a non-construction scene.

[0081] At the same time, when the number of elements in the target construction element is not greater than the number of elements in the second construction element, from the logical point of view of construction activities, the construction scene usually shows a state of gradually increasing or maintaining a certain scale of construction elements. If the number of target construction elements does not increase or even decrease, it does not meet the rule of element accumulation in the process of construction activities, and is more inclined to be a non-construction scene. At the same time, it can avoid misjudgment as a construction scene due to the appearance of a small amount of construction elements caused by some accidental factors, improve the rigor and accuracy of the judgment, and also save the calculation resources and time cost of further analysis.

[0082] In step 304, if the second construction element does not contain the fence element, or the second construction element contains the fence element and the second construction element meets the pre-judgment condition, the step of determining the first identification result based on the target construction element and the number of elements corresponding to the target construction element is performed.

[0083] In the embodiment of the present application, if the second construction element does not contain the fence element, or the second construction element contains the fence element and the second construction element does not satisfy the pre-judgment condition, the step of determining the first identification result based on the target construction element and the number of elements corresponding to the target construction element in step 101 is performed. The pre-judgment condition is that the second construction element is different from the target construction element and the number of elements in the target construction element is greater than the number of elements in the second construction element. When the second construction element in the historical false positive image is different from the target construction element in the current scene image to be identified, and the number of construction elements in the current image is more, the pre-judgment condition is satisfied, and the determination of the first identification result needs to be performed according to the target construction element and the number of elements in the target construction element according to the conventional process.

[0084] Further, since the fence element has an important identification role in construction scene recognition, when the second construction element does not contain the fence element, it means that the current judgment of the construction scene lacks this key reference information, and at this time it is necessary to further analyze the target construction element and the number of elements to accurately determine whether the scene is a construction scene. When the second construction element does not satisfy the pre-judgment condition, it means that there is a difference between the construction elements obtained by the previous identification and the target construction element, and the number of elements in the target construction element is more, which may imply that the scene state has changed or the previous identification has a certain deviation, so it is necessary to determine the first identification result based on the target construction element and the number of elements in the target construction element again to ensure the accuracy of the identification result.

[0085] In the embodiment of the present application, by obtaining the historical false positive image corresponding to the scene image to be identified and identifying the second construction element in the historical false positive image, a reference basis is provided for the current scene judgment. If the second construction element contains the fence element, comparison can be performed to more accurately analyze the changes of scene characteristics and avoid misjudgment due to similarity with historical false positive situations. At the same time, when the second construction element does not contain the fence element or satisfies a certain pre-judgment condition, the step of determining the first identification result based on the target construction element is performed. This avoids unnecessary complex calculations in obviously non-construction scene situations, reasonably allocates computing resources, and makes the identification process more efficient. In addition, the embodiment of the present application can adapt to construction scenes of different types and complexity, whether there is a fence element or other construction elements, and can make judgments according to reasonable logic, providing reliable technical support for construction scene monitoring, management and other work, and helping to discover real construction scenes in time and ensure the smooth development of related work.

[0086] Optionally, step 103 can include the following steps:

[0087] Step 401, in the case where the first construction element only contains the fence element, if the fence identification result corresponding to the fence element is a construction fence, it is determined that the target identification result is a construction scene.

[0088] In the embodiment of the present application, when it is identified that the first construction element in the first scene image only contains the fence element, the fence element needs to be further accurately identified. It can be understood that the further identification of the fence element can refer to step 201, which will not be repeated here. When it is determined that the fence recognition result corresponding to the fence element is a construction fence, it can be directly determined that the target recognition result is a construction scene.

[0089] In step 402, in the case that the first construction element does not contain the fence element or contains the fence element and other elements at the same time, it is determined that the target recognition result is a construction scene.

[0090] In the embodiment of the present application, if the identification result of the scene image to be identified is to be verified, it means that the initial identification cannot clearly determine whether the scene is a construction scene, and the scene nature has a certain ambiguity. In this case, the first scene image still appears the phenomenon that the first construction element does not contain the fence element or contains the fence element and other elements at the same time. From the logic and characteristics of construction activities, the construction scene often has continuity and dynamics. Within a preset time length, even if the initial identification result is uncertain, the appearance of other construction elements such as engineering vehicles, construction tools, building materials, etc. in the first scene image, or the existence of the fence element and other construction elements at the same time, fully indicates that there are still construction-related activities in the scene within the preset time length. For example, the existence of engineering vehicles may mean that material transportation or site leveling work is being carried out; the appearance of construction tools and building materials directly reflects that construction operations are being carried out. Therefore, based on the information contained in the first construction element, it can be determined that the target recognition result is a construction scene. By fully considering the dynamic change characteristics of the construction scene, it can effectively avoid misjudgment due to the uncertainty of the initial identification, and improve the accuracy and reliability of the judgment of the construction scene.

[0091] In step 403, in the case that the first construction element is empty, or the fence recognition result corresponding to the fence element is a non-construction fence, it is determined that the target recognition result is a non-construction scene.

[0092] In the embodiment of the present application, when the first construction element is empty, it means that no construction-related element is identified in the scene after the preset time length, which obviously does not meet the characteristics of the construction scene, so the target recognition result is determined to be a non-construction scene. Moreover, if the fence recognition result corresponding to the fence element is a non-construction fence, even if the fence element exists in the scene, since the fence does not have the characteristics of the construction fence, it cannot be used as an effective identifier of the construction scene. Combined with other conditions that may exist in the scene, it is determined that the scene does not belong to the construction scene, and therefore the target recognition result is determined to be a non-construction scene.

[0093] In further judgment on the to-be-verified scene, if the first construction element in the first scene image does not contain a fence element but contains other construction elements, or contains both a fence and other elements, combined with the persistence and dynamics of construction activities, the scene can be more accurately determined as a construction scene. This is because construction scenes often have persistence and dynamics, and the presence of these construction elements within a preset time period indicates that there are still construction-related activities in the scene. When the first construction element is empty or the fence element is not a construction fence, it can be determined as a non-construction scene. This dynamic judgment logic fully considers the actual situation of the construction scene, effectively avoids misjudgment due to initial identification ambiguity, and improves the overall identification accuracy.

[0094] In the embodiments of the present application, in view of the characteristics of the construction scene having persistent and dynamic changes, when the identification result of the to-be-identified scene image is to be verified, the first construction element in the first scene image after a preset time period is analyzed for further judgment. If the first construction element does not contain a fence element but contains other construction elements, or contains both a fence and other elements, combined with the persistence and dynamics of construction activities, the scene can be more accurately determined as a construction scene, avoiding misjudgment due to initial identification ambiguity and effectively improving the overall identification accuracy. At the same time, when the first construction element is empty or the fence element is not a construction fence, it can be quickly determined as a non-construction scene, avoiding unnecessary calculation and analysis, reasonably allocating computing resources, and optimizing identification efficiency.

[0095] In a possible implementation, the QUIC / UDP protocol can be used for data transmission of scene images. In the data transmission process, the QUIC / UDP protocol can reduce the handshake link, quickly establish a connection, significantly reduce the handshake delay, and speed up the data transmission speed. At the same time, the QUIC / UDP protocol supports data fragmentation transmission, and for larger data such as construction scene images, it can be split into multiple small fragments for parallel transmission, improving transmission efficiency and ensuring that image data can be transmitted to the processing end in a timely and complete manner, providing timely data support for subsequent identification work.

[0096] The image recognition model can be a lightweight model, which can greatly reduce the parameter quantity and calculation quantity of the model while ensuring a certain recognition accuracy. At the same time, the lightweight model can be run on an edge computing node, and the edge computing node can sink the computing power to the place close to the data source, reducing the distance and time of data transmission to the cloud server, and being able to quickly perform preliminary identification and analysis on the construction scene image, balancing the recognition accuracy and real-time performance. It can not only give the identification result in time, but also ensure the accuracy of the identification to a certain extent, meeting the needs of real-time monitoring of construction scenes.

[0097] Through continuous data accumulation, a large amount of construction scene image data and related identification results, scene change information, etc. are collected. Using these rich data, data analysis and machine learning algorithms are used to further mine potential rules and risk factors in the construction scene. For example, the correlation between the changes in image features in different construction stages and construction risks is analyzed to enhance the ability to predict construction risks. Potential safety hazards can be detected in advance to provide strong support for construction management and safety protection, and to achieve a shift from passive identification to proactive prevention.

[0098] Exemplary, Figure 2 A specific step flowchart of a construction scene identification method is provided, as Figure 2 As shown, first, image recognition is performed on the to-be-identified scene image to identify whether it contains preset construction elements such as fences, construction vehicles, safety helmets, work clothes, and pipes, etc. to obtain target construction elements contained in the to-be-identified scene image. It is determined whether the identification result generated by the historical camera for the collected historical image is a false alarm. If so, the historical false alarm image is obtained, and the second construction elements contained in the historical false alarm image are identified. In the case where the second construction elements contain fence elements, if the second construction elements are the same as the target construction elements or the number of elements in the target construction elements is less than or equal to the number of elements in the second construction elements, i.e., there is no new construction element in the to-be-identified scene image relative to the historical false alarm image, it can be directly determined that the identification result is a non-construction scene. If the second construction elements are different from the target construction elements and the number of elements in the target construction elements is greater than the number of elements in the second construction elements, i.e., there is a new construction element in the to-be-identified scene image relative to the historical false alarm image, or the second construction elements do not contain fence elements, it is determined whether the to-be-identified scene image contains a fence. If the to-be-identified scene image contains a fence, the type of the fence element is determined. If it is a construction fence, it is determined that the identification result is a construction scene. If the to-be-identified scene image does not contain a fence or the fence element is not a construction fence, it is determined whether the number of element types of the target construction elements is at least two. If so, it is determined that the identification result is a construction scene. If not, it is determined whether the target construction elements contain dynamic construction elements. If not, it is determined that the identification result is a non-construction scene. If so, it is determined that the identification result is to be verified and a first scene image after a preset time period is further obtained. It is determined whether the first scene image contains a fence element. If it contains a fence, the type of the fence element is determined. If it is a construction fence, it is determined that the identification result is a construction scene. If the to-be-identified scene image does not contain a fence or contains a fence element and other elements, it is determined that the identification result is a construction scene. If the first scene image does not contain any construction element or contains a fence element but is not a construction fence, it is determined that the identification result is a non-construction scene.

[0099] Figure 3is a structural schematic diagram of a construction scene recognition device provided by an embodiment of the present application, as shown in the figure, the device can specifically include: Figure 3

[0100] The first identification module 501 is configured to identify a target construction element contained in a to-be-identified scene image, and determine a first identification result based on the target construction element and an element quantity corresponding to the target construction element.

[0101] The first acquisition module 502 is configured to acquire a first scene image in a case where the first identification result is to be verified; a first acquisition time corresponding to the first scene image and a second acquisition time corresponding to the to-be-identified scene image are separated by a preset time length.

[0102] The first determination module 503 is configured to determine a target identification result based on a first construction element contained in the first scene image.

[0103] Optionally, the first identification module 501 includes:

[0104] The second determination module is configured to, in a case where the target construction element contains a fence element, determine that the first identification result is a construction scene if a fence identification result corresponding to the fence element is a construction fence.

[0105] The third determination module is configured to, in a case where the target construction element does not contain the fence element or the fence identification result is not a construction fence, determine that the first identification result is a construction scene if an element type quantity of the target construction element is at least two.

[0106] The fourth determination module is configured to, in a case where the element type quantity of the target construction element is less than two, determine that the first identification result is to be verified if the target construction element contains a dynamic construction element.

[0107] Optionally, the device further includes:

[0108] The second acquisition module is configured to acquire a historical false positive image corresponding to the to-be-identified scene image, and identify a second construction element contained in the historical false positive image.

[0109] The first comparison module is configured to compare the second construction element with the target construction element in a case where the second construction element contains a fence element.

[0110] ​The fifth determining module is configured to determine that the identification result of the to-be-identified scene image is a non-construction scene when the second construction element does not satisfy a pre-judgment condition, the pre-judgment condition being that the second construction element is different from the target construction element and the number of elements in the target construction element is greater than the number of elements in the second construction element.

[0111] The first executing module is configured to execute the step of determining the first identification result based on the target construction element and the number of elements corresponding to the target construction element when the second construction element does not contain a fence element, or the second construction element contains a fence element and the second construction element satisfies the pre-judgment condition.

[0112] Optionally, the first determining module 503 comprises:

[0113] The sixth determining module is configured to determine that the target identification result is a construction scene when the fence identification result corresponding to the fence element is a construction fence if the first construction element contains only the fence element.

[0114] The seventh determining module is configured to determine that the target identification result is a construction scene when the first construction element does not contain a fence element or contains a fence element and other elements at the same time.

[0115] The eighth determining module is configured to determine that the target identification result is a non-construction scene when the first construction element is empty or the fence identification result corresponding to the fence element is a non-construction fence.

[0116] The present application also provides an electronic device, referring to Figure 4 , comprising a processor 601, a memory 602, and a computer program 6021 stored in the memory and executable on the processor, and the processor implements the construction scene identification method of the foregoing embodiments when executing the program.

[0117] The present application also provides a readable storage medium, when the instructions in the storage medium are executed by the processor of the electronic device, the electronic device can execute the construction scene identification method of the foregoing embodiments.

[0118] For the device embodiment, it is basically similar to the method embodiment, so the description is relatively simple, and the related parts are described in the part of the method embodiment.

[0119] The algorithms and displays presented herein are not inherently related to any particular computer, virtual system, or other apparatus. Various general purpose systems can be used with programs in accordance with the teachings herein, or it can prove convenient to construct more specialized apparatus to perform the required method steps. The required structure for a variety of these systems will be apparent from the description above. In addition, the present application is not intended to be limited to any particular programming language. It will be appreciated that there are many programming languages that can be used to implement the teachings herein, and any specific language can be chosen for use in this application.

[0120] In the description provided herein, numerous specific details are set forth. However, it is understood that embodiments of the application can be practiced without these specific details. In some instances, well-known methods, structures and techniques have not been described in detail in order to not obscure the understanding of this description.

[0121] Similarly, it is to be understood that the mechanical details of the application sometimes are presented in terms of certain spatially-related or geometrical configurations and / or descriptions. It will be appreciated by those skilled in the art that the application can be practiced without such specifically-named spatial / geometrical configurations and / or descriptions. Furthermore, it is to be understood that the use of certain specific language or terminology in the specification is merely to further aid a state of the art and is not intended to limit the scope of the application.

[0122] Those skilled in the art will appreciate that the modules in the apparatuses in the embodiments can be adapted and placed in one or more apparatuses other than the embodiments. The modules or units or components in the embodiments can be combined into one module or unit or component, and further can be divided into more sub-modules or sub-units or sub-components. Any combination of all the features disclosed in the specification (including the accompanying claims, abstract and drawings), and any method or apparatus so disclosed, can be made unless specifically stated otherwise. Each feature disclosed in the specification (including the accompanying claims, abstract and drawings) can be replaced by alternative features serving the same, equivalent or similar purpose unless specifically stated otherwise.

[0123] Various component embodiments of the present application can be implemented in hardware, or as software modules running in one or more processors, or in combinations thereof. As will be apparent to those of ordinary skill in the art, some or all of the functionality of some or all of the components in the sequencing apparatus according to the present application can be implemented using a microprocessor or a digital signal processor (DSP) in practice. The present application can also be implemented as a program for executing, in whole or in part, the methods described herein on a device or apparatus. Such a program embodying the present application can be stored on a computer-readable medium, or can be in the form of signals on one or more carriers. Such signals can be downloaded from an Internet website, or provided on a carrier, or in any other form.

[0124] It should be noted that the above-mentioned embodiments illustrate rather than limit the application, and that one skilled in the art will be able to design many alternative embodiments without departing from the scope of the appended claims. In the claims, any reference signs placed between parentheses shall not be construed as limiting the claim. The word 'comprising' does not exclude the presence of elements or steps not listed in a claim. The word 'a' or 'an' preceding an element does not exclude the presence of a plurality of such elements. The application can be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In the claims, the word 'first','second', 'third', etc. does not limit the number for these elements. These words are only used to distinguish between alternative claims. The mere fact that different claims depend on the same independent claim is not a reason for assuming that the dependent claims are mutually identical.

[0125] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0126] It should be noted that all the actions of obtaining signals, information or data in the present application are carried out in accordance with the corresponding data protection regulations and policies of the country where the device is located, and with the authorization of the owner of the corresponding device.

[0127] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. Any modifications, equivalent replacements and improvements made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0128] The above only describes the specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which shall be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A method of identifying a construction scene, characterized in that, The method comprises: identifying a target construction element contained in a to-be-identified scene image, and determining a first identification result based on the target construction element and an element quantity corresponding to the target construction element; in a case where the first identification result is to be verified, acquiring a first scene image; a first acquisition time corresponding to the first scene image and a second acquisition time corresponding to the to-be-identified scene image are separated by a preset time length; determining a target identification result based on a first construction element contained in the first scene image.

2. The method of claim 1, wherein, The method comprises: in a case where the target construction element contains a fence element, if a fence identification result corresponding to the fence element is a construction fence, it is determined that the first identification result is a construction scene; in a case where the target construction element does not contain the fence element or the fence identification result is not a construction fence, if an element type quantity of the target construction element is at least two, it is determined that the first identification result is a construction scene; in a case where the element type quantity of the target construction element is less than two, if the target construction element contains a dynamic construction element, it is determined that the first identification result is to be verified.

3. The method of claim 1, wherein, After identifying the target construction element contained in the to-be-identified scene image, the method further comprises: acquiring a historical false positive image corresponding to the to-be-identified scene image, and identifying a second construction element contained in the historical false positive image; in a case where the second construction element contains a fence element, comparing the second construction element with the target construction element; in a case where the second construction element does not satisfy a pre-judgment condition, determining that an identification result for the to-be-identified scene image is a non-construction scene; the pre-judgment condition is that the second construction element is different from the target construction element and an element quantity of the target construction element is greater than an element quantity of the second construction element; in a case where the second construction element does not contain a fence element, or the second construction element contains a fence element and the second construction element satisfies the pre-judgment condition, performing the step of determining the first identification result based on the target construction element and the element quantity corresponding to the target construction element.

4. The method of claim 1, wherein, The method comprises: in a case where the first construction element contains only a fence element, if a fence identification result corresponding to the fence element is a construction fence, it is determined that the target identification result is a construction scene; in a case where the first construction element does not contain a fence element or contains a fence element and other elements at the same time, it is determined that the target identification result is a construction scene; in a case where the first construction element is empty or a fence identification result corresponding to the fence element is a non-construction fence, it is determined that the target identification result is a non-construction scene.

5. An apparatus for identifying a construction scene, characterized by The device comprises: The first identification module is configured to identify a target construction element contained in the to-be-identified scene image, and determine a first identification result based on the target construction element and an element quantity corresponding to the target construction element. The first acquisition module is configured to acquire a first scene image in a case where the first identification result is to be verified; a first acquisition time corresponding to the first scene image and a second acquisition time corresponding to the to-be-identified scene image are separated by a preset time length. The first determination module is configured to determine a target identification result based on a first construction element contained in the first scene image.

6. The apparatus of claim 5, wherein, The first identification module includes: The second determination module is configured to, in a case where the target construction element contains a fence element, determine that the first identification result is a construction scene if a fence identification result corresponding to the fence element is a construction fence. The third determination module is configured to, in a case where the target construction element does not contain the fence element or the fence identification result is not a construction fence, determine that the first identification result is a construction scene if an element type quantity of the target construction element is at least two. The fourth determination module is configured to, in a case where the element type quantity of the target construction element is less than two, determine that the first identification result is to be verified if the target construction element contains a dynamic construction element.

7. The apparatus of claim 5, wherein, The device further includes: The second acquisition module is configured to acquire a historical false positive image corresponding to the to-be-identified scene image, and identify a second construction element contained in the historical false positive image. The first comparison module is configured to compare the second construction element with the target construction element in a case where the second construction element contains a fence element. The fifth determination module is configured to, in a case where the second construction element does not satisfy a pre-judgment condition, determine that an identification result for the to-be-identified scene image is a non-construction scene; the pre-judgment condition is that the second construction element is different from the target construction element and an element quantity in the target construction element is greater than an element quantity of the second construction element. The first execution module is configured to, in a case where the second construction element does not contain a fence element, or the second construction element contains a fence element and the second construction element satisfies the pre-judgment condition, execute the step of determining a first identification result based on the target construction element and an element quantity corresponding to the target construction element.

8. The apparatus of claim 5, wherein, The first determination module includes: The sixth determination module is configured to, in a case where the first construction element contains only a fence element, determine that the target identification result is a construction scene if a fence identification result corresponding to the fence element is a construction fence. The seventh determination module is configured to, in a case where the first construction element does not contain a fence element or contains a fence element and other elements at the same time, determine that the target identification result is a construction scene. The eighth determination module is configured to, in a case where the first construction element is empty or a fence identification result corresponding to the fence element is a non-construction fence, determine that the target identification result is a non-construction scene.

9. An electronic device, comprising: The device further includes: A processor, a memory, and a computer program stored on the memory and executable on the processor, the processor implementing the construction site scene identification method according to any one of claims 1-4 when executing the program.

10. A readable storage medium, characterized by, The instructions in the storage medium, when executed by the processor of the electronic device, enable the electronic device to perform the construction site scene identification method according to any one of claims 1-4.

Images