Intelligent building safety supervision system based on big data

Abstract

The invention discloses an intelligent building safety supervision system based on big data, and the system comprises a collection module which comprises a plurality of cameras and collects image data and video data in real time through the cameras; the data acquisition module is used for acquiring image data and video data and transmitting the acquired image data and video data to the data storage module; the data storage module is used for storing the picture data and the video data transmitted by the acquisition module; the material pile supervision module is used for supervising a material pile in the building materials, identifying whether the stacking height of the material pile meets the stacking height requirement or not, and if not, triggering early warning and judging whether the material pile meets the tidy stacking requirement or not. Whether the stacking height of the material pile meets the stacking height requirement or not is identified, whether the material pile meets the tidy stacking requirement or not is judged, the safety accident that materials roll down or collapse due to the fact that the material pile is too high or the material pile is not stacked tidily can be effectively prevented, and the life safety of construction site personnel is guaranteed to the maximum extent.

Description

Technical Field

[0001] The present invention relates to the technical field of building safety supervision, and in particular to an intelligent building safety supervision system based on big data. Background Art

[0002] In today's rapidly developing construction industry, the scale and complexity of construction sites are constantly increasing, and safety management of construction sites has become increasingly important. The safety of material stacking on construction sites is directly related to the safety of construction workers and the smooth progress of the entire project. Once a material rolls or collapses, it is very likely to trigger a chain reaction, resulting in irreversible serious consequences.

[0003] Publication No. CN117765434B discloses an on-site video image processing method and system for construction projects, which divides the construction project site into: a construction site area, an import and export road area, a material stacking area and an equipment room area, and sets the construction site area and the import and export road area as primary monitoring areas, and sets the material stacking area and the equipment room area as secondary monitoring areas. By setting the material stacking area and the equipment room area as secondary monitoring areas, there are fewer people or vehicles in the secondary monitoring area, and the main monitoring of objects and materials is carried out; by grouping continuous image frames of the monitoring image, comparing the deviations of each group of image frames, and calculating the intra-group difference values ​​of the image frame groupings; the monitoring images when the objects are changed can be retained more accurately, so as to more safely ensure the safety of the objects stored in the secondary monitoring area.

[0004] However, the above application still has the following problems: Although it focuses on implementing zoning monitoring of the material stacking area, focusing on the change of items, and paying attention to the dynamic information of materials to a certain extent, it only stays at the level of change monitoring. At the construction site, when the material pile is high or stacked unevenly, if there is a lack of effective early warning and prevention mechanisms, it may deteriorate into a large-scale collapse or rolling accident, which will not only damage valuable materials and cause direct economic losses, but also endanger the lives of on-site construction workers. Summary of the invention

[0005] In order to solve the technical problems existing in the background technology, the present invention proposes an intelligent building safety supervision system based on big data.

[0006] The intelligent building safety supervision system based on big data proposed in the present invention includes: Acquisition module: includes multiple cameras, collects picture data and video data in real time through the cameras; and is used to transmit the collected picture data and video data to the data storage module; Data storage module: used to store the image data and video data transmitted by the acquisition module; Material pile supervision module: used to supervise the material piles in the construction materials, identify whether the stacking height of the material pile meets the stacking height requirements, if not, trigger an early warning, and determine whether the material pile meets the neat stacking requirements, if not, trigger an early warning.

[0007] Preferably, the material pile supervision module includes: Deep learning model generation unit: adopts deep learning algorithm to analyze the image data and video data in the data storage module to identify whether the stacking height of the material pile in the continuous frames meets the stacking height requirements; Neat stacking judgment unit: In the deep learning model generation unit, the deep learning algorithm uses the Canny edge detection algorithm to determine the outline of the material pile, the boundary and stacking direction of the material pile, and the Canny edge detection algorithm to determine whether the material pile is stacked neatly.

[0008] Preferably, the material pile supervision module further includes: Material pile height warning unit: used to determine the material pile height. If it exceeds the set threshold, an early warning is triggered. Preferably, in the deep learning model generation unit, for the image data and video data in the data storage module, pixels belonging to the material pile are directly identified by a deep learning algorithm, the pixels belonging to the material pile are marked as a specific category, and the remaining background pixels are marked as another category; A material pile segmentation map is generated from image data and video data by labeling pixels belonging to the material pile as specific categories.

[0009] Preferably, in the deep learning model generation unit, the material categories in the material pile segmentation map are identified by a deep learning algorithm, and different height thresholds are set according to different material categories.

[0010] Preferably, in the material pile stacking height warning unit, when the deep learning algorithm determines whether the material category in the material pile segmentation map exceeds the height threshold, when performing height recognition on one of the material piles M of a material category, the material pile segmentation map of multiple consecutive frames of the material pile M is used, and the stacking height of the material pile M in each material pile segmentation map is calculated by the deep learning algorithm to obtain a stacking height set S1, and then the Kalman filter algorithm is used to smooth the stacking height set S1 to obtain a stacking height. , and then determine the stacking height Whether it exceeds the set threshold of the material category. If so, an early warning is triggered.

[0011] Preferably, in the stacking neatness determination unit, a Canny edge detection algorithm is used to determine whether the material pile is neatly stacked. If it exceeds a set threshold, an early warning is triggered as follows: After receiving the material pile segmentation map, the input color material pile segmentation map is converted into a grayscale image. The Gaussian filtering algorithm is used to smooth the grayscale image; Through gradient calculation, the image after Gaussian filtering is convolved in the horizontal and vertical directions respectively to obtain the gradient component Gx in the horizontal direction and the gradient component Gy in the vertical direction; Calculate the gradient magnitude G at each pixel position: ; Calculate the gradient direction : ; The value range is 0°-360°; Calculate the gradient direction of each pixel position; Through gradient calculation, the position of pixel grayscale change in the material pile segmentation map is located, the outline of the material pile is outlined, and the stacking boundary of the material pile is determined; Then, based on the edge direction and continuity, the Hough transform technique is used to determine the stacking direction of the material pile; Next, the angle deviation between the boundary line segment of the material pile outline and the straight line of the stacking direction is calculated. If the angle deviation exceeds the set threshold, it is determined that the material pile is not stacked neatly, and an early warning is triggered immediately. Calculate the length consistency deviation between the boundary line segment of the material pile outline and the stacking boundary. If the length consistency deviation exceeds the set threshold, it is determined that the material pile is not stacked neatly and an early warning is triggered immediately. Calculate the straightness of the stacking boundary. If the straightness of the stacking boundary exceeds the set threshold, it is determined that the material pile is not stacked neatly and an early warning is triggered immediately.

[0012] The smart building safety supervision method based on big data includes the following steps: The camera collects image data and video data in real time; and is used to transmit the collected image data and video data to the data storage module for data storage; For the image data and video data in the data storage module, the pixels belonging to the material pile are directly identified through the deep learning algorithm, and the pixels belonging to the material pile are marked as a specific category, and the remaining background pixels are marked as another category; Generate a material pile segmentation map from the image data and the video data by labeling pixels belonging to the material pile as specific categories; Next, the Canny edge detection algorithm is used to determine the material stacking boundary and stacking direction of the material pile segmentation map. The Canny edge detection algorithm is used to determine whether the material pile is stacked neatly. If it exceeds the set threshold, an early warning is triggered; Identify material categories in the material pile segmentation map through deep learning algorithms; set different height thresholds according to different material categories; Then, a deep learning algorithm is used to determine whether the material category in the material pile segmentation map exceeds the height threshold. If so, an early warning is triggered; When the deep learning algorithm determines whether the material category in the material pile segmentation map exceeds the height threshold, when performing height recognition on a material pile M of a material category, the material pile segmentation map of multiple consecutive frames of the material pile M is used, and the stacking height of the material pile M in each material pile segmentation map is calculated by the deep learning algorithm to obtain a stacking height set S1, and then the Kalman filter algorithm is used to smooth the stacking height set S1 to obtain a stacking height. , and then determine the stacking height Whether it exceeds the set threshold of the material category. If so, an early warning is triggered.

[0013] In the present invention, the proposed intelligent building safety supervision system based on big data has the following beneficial technical effects: 1. In the safety supervision of material piles at construction sites, the deep learning algorithm directly identifies pixels belonging to the material pile, marks the pixels belonging to the material pile as a specific category, and marks the remaining background pixels as another category. Then, the pixels belonging to the material pile are marked as a specific category, and a material pile segmentation map is generated from the image data and video data; the irrelevant interference information of the complex buildings, equipment, and vegetation in the background of the image data and video data is filtered out. In this way, the height characteristics of the material pile itself can be focused, and the stacking height can be judged more accurately.

[0014] 2. When judging the height of the material pile, the material pile segmentation map of multiple continuous frames of the material pile is used, and the stacking height of each frame is calculated using the deep learning algorithm to form a stacking height set S1, and then the Kalman filter algorithm is used to smooth it to obtain the final stacking height. This processing method effectively reduces the height data fluctuation caused by factors such as instantaneous errors in image acquisition and light changes, increases the accuracy of height judgment, makes the judgment results more stable and reliable, and enhances the credibility and effectiveness of safety supervision.

[0015] 3. Use the Canny edge detection algorithm to determine the material stacking boundary and direction of the material pile segmentation diagram, and judge in real time whether the material pile is neatly stacked. Once the set threshold is exceeded, an early warning is triggered immediately. This automated detection method allows construction management personnel to grasp the material stacking status in a timely manner, quickly adjust the uneven stacking situation, avoid imbalance and collapse of materials caused by messy stacking, reduce potential safety hazards at the construction site, and maintain the orderliness of the construction site.

[0016] In summary, the present application can effectively prevent safety accidents caused by material piles that are too high or unevenly stacked, resulting in material rolling or collapse, by identifying whether the stacking height of the material pile meets the stacking height requirements and judging whether the material pile meets the requirements for neat stacking, thereby ensuring the safety of the lives of construction site personnel to the greatest extent.

[0017] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a principle block diagram of the system of the present invention; Figure 2 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0019] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar symbols throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0020] like Figure 1 The big data-based intelligent building safety supervision system shown includes: a collection module: including multiple cameras, collecting picture data and video data in real time through the cameras; and used to transmit the collected picture data and video data to the data storage module; Data storage module: used to store the image data and video data transmitted by the acquisition module; Material pile supervision module: used to supervise the material pile in the construction materials, identify whether the stacking height of the material pile meets the stacking height requirements, and determine whether the material pile meets the neat stacking requirements. If not, an early warning is triggered to prevent the materials in the material pile from rolling down or collapsing.

[0021] The material pile supervision module includes: Deep learning model generation unit: adopts deep learning algorithm to analyze the image data and video data in the data storage module to identify whether the stacking height of the material pile in the continuous frames meets the stacking height requirements; In the deep learning model generation unit, for the image data and video data in the data storage module, the pixels belonging to the material pile are directly identified through the deep learning algorithm, and the pixels belonging to the material pile are marked as a specific category, and the remaining background pixels are marked as another category; Generate a material pile segmentation map from the image data and the video data by labeling pixels belonging to the material pile as specific categories; After obtaining the material pile segmentation map, in the subsequent stacking height recognition process, the original image area corresponding to the material pile pixels in the segmentation map can be simply extracted for analysis, and irrelevant information such as complex background buildings, equipment, and vegetation can be filtered out, so that the deep learning model can focus on the height characteristics of the material pile itself, and the stacking height can be judged more accurately; In the safety supervision of material piles at construction sites, the deep learning algorithm directly identifies pixels belonging to the material pile, marks the pixels belonging to the material pile as a specific category, and marks the remaining background pixels as another category. Then, the pixels belonging to the material pile are marked as a specific category, and a material pile segmentation map is generated from the image data and video data; the irrelevant interference information of the complex buildings, equipment, and vegetation in the background of the image data and video data is filtered out. In this way, the height characteristics of the material pile itself can be focused on, and the stacking height can be judged more accurately.

[0022] In the deep learning model generation unit, the material categories in the material pile segmentation map are identified by a deep learning algorithm, and different height thresholds are set according to different material categories; Different height thresholds can be set manually according to different material categories; Neat stacking determination unit: In the deep learning model generation unit, the deep learning algorithm uses the Canny edge detection algorithm to determine the outline of the material pile, the boundary and stacking direction of the material pile, and the Canny edge detection algorithm to determine whether the material pile is neatly stacked. Material pile height warning unit: used to determine the material pile height. If it exceeds the set threshold, an early warning is triggered. In the deep learning model generation unit, the deep learning algorithm uses the Kalman filter algorithm in the material pile stacking height recognition. For the stacking height recognition results of the deep learning algorithm in continuous frames, the Kalman filter algorithm smoothes the estimation of the stacking height, and then judges whether the stacking height obtained by the Kalman filter algorithm exceeds the set threshold of the material category. If it exceeds, an early warning is triggered; In an optional embodiment, in the material pile stacking height warning unit, when the deep learning algorithm determines whether the material category in the material pile segmentation map exceeds the height threshold, when performing height recognition on one of the material piles M of a material category, the material pile segmentation map of multiple consecutive frames of the material pile M is used, and the stacking height of the material pile M in each material pile segmentation map is calculated by the deep learning algorithm to obtain a stacking height set S1, and then the Kalman filter algorithm is used to smooth the stacking height set S1 to obtain a stacking height. , and then determine the stacking height Whether it exceeds the set threshold of the material category. If so, an early warning is triggered.

[0023] When judging the height of the material pile, the material pile segmentation map of multiple continuous frames of the material pile is used, and the stacking height of each frame is calculated using the deep learning algorithm to form a stacking height set S1, and then the Kalman filter algorithm is used to smooth it to obtain the final stacking height. This processing method effectively reduces the height data fluctuation caused by factors such as instantaneous errors in image acquisition and light changes, increases the accuracy of height judgment, makes the judgment results more stable and reliable, and enhances the credibility and effectiveness of safety supervision.

[0024] In an optional embodiment, in the stacking neatness determination unit, a Canny edge detection algorithm is used to determine whether the material pile is neatly stacked. If it exceeds a set threshold, an early warning is triggered as follows: After receiving the material pile segmentation map, the input color material pile segmentation map is converted into a grayscale image. The Gaussian filtering algorithm is used to smooth the grayscale image; Through gradient calculation, the image after Gaussian filtering is convolved in the horizontal and vertical directions respectively to obtain the gradient component Gx in the horizontal direction and the gradient component Gy in the vertical direction; Calculate the gradient magnitude G at each pixel position: ; Calculate the gradient direction : ; The value range is 0°-360°; arctan is the inverse tangent function; Calculate the gradient direction of each pixel position; Through gradient calculation, the position of pixel grayscale change in the material pile segmentation map is located, the outline of the material pile is outlined, and the stacking boundary of the material pile is determined; Then, based on the edge direction and continuity, the Hough transform technique is used to determine the stacking direction of the material pile; Next, the angle deviation between the boundary line segment of the material pile outline and the straight line of the stacking direction is calculated. If the angle deviation exceeds the set threshold, it is determined that the material pile is not stacked neatly, and an early warning is triggered immediately. Calculate the length consistency deviation between the boundary line segment of the material pile outline and the stacking boundary. If the length consistency deviation exceeds the set threshold, it is determined that the material pile is not stacked neatly and an early warning is triggered immediately. Calculate the straightness of the stacking boundary. If the straightness of the stacking boundary exceeds the set threshold, it is determined that the material pile is not stacked neatly and an early warning is triggered immediately.

[0025] The Canny edge detection algorithm is used to determine the material stacking boundary and direction of the material pile segmentation map, and to judge in real time whether the material pile is neatly stacked. Once the set threshold is exceeded, an early warning is triggered immediately. This automated detection method allows construction management personnel to grasp the material stacking status in a timely manner and quickly adjust the uneven stacking to avoid imbalance and collapse of materials caused by messy stacking, reduce potential safety hazards at the construction site, and maintain the orderliness of the construction site.

[0026] like Figure 2 The smart building safety supervision method based on big data includes the following steps: The camera collects image data and video data in real time; and is used to transmit the collected image data and video data to the data storage module for data storage; For the image data and video data in the data storage module, the pixels belonging to the material pile are directly identified through the deep learning algorithm, and the pixels belonging to the material pile are marked as a specific category, and the remaining background pixels are marked as another category; Generate a material pile segmentation map from the image data and the video data by labeling pixels belonging to the material pile as specific categories; Next, the Canny edge detection algorithm is used to determine the material stacking boundary and stacking direction of the material pile segmentation map. The Canny edge detection algorithm is used to determine whether the material pile is stacked neatly. If it exceeds the set threshold, an early warning is triggered; Basis for judging neat stacking: After the algorithm extracts the boundary of the material pile, the boundary of the neatly stacked material pile should have a regular geometric shape or present a regular arrangement. For example, if the material is in the shape of a cuboid and is neatly stacked, in an ideal front view, the boundary obtained by the Canny edge detection algorithm should be parallel straight lines with uniform spacing. From the perspective of stacking direction, if the material pile has specific stacking requirements, such as stacking in sequence in a certain direction, the edge detection algorithm can determine the direction of the edge of the material pile, thereby judging whether it meets the requirements of the stacking direction. Once the boundary shape is irregular or the stacking direction does not meet the set standard, and this difference exceeds the set threshold, it can be judged that the material pile is not stacked neatly.

[0027] Identify material categories in the material pile segmentation map through deep learning algorithms; set different height thresholds according to different material categories; Then, a deep learning algorithm is used to determine whether the material category in the material pile segmentation map exceeds the height threshold. If so, an early warning is triggered; When the deep learning algorithm determines whether the material category in the material pile segmentation map exceeds the height threshold, when performing height recognition on a material pile M of a material category, the material pile segmentation map of multiple consecutive frames of the material pile M is used, and the stacking height of the material pile M in each material pile segmentation map is calculated by the deep learning algorithm to obtain a stacking height set S1, and then the Kalman filter algorithm is used to smooth the stacking height set S1 to obtain a stacking height. , and then determine the stacking height Whether it exceeds the set threshold of the material category. If so, an early warning is triggered.

[0028] In summary, the present application can effectively prevent safety accidents caused by material piles that are too high or unevenly stacked, resulting in material rolling or collapse, by identifying whether the stacking height of the material pile meets the stacking height requirements and judging whether the material pile meets the requirements for neat stacking, thereby ensuring the safety of the lives of construction site personnel to the greatest extent.

[0029] Through this intelligent supervision system that integrates multiple advanced algorithms, the supervision of material stacking and safety protection facilities at construction sites can be fully automated and intelligent, greatly improving the efficiency of construction safety management and eliminating safety hazards in a timely manner.

[0030] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0031] In the embodiments provided by the present invention, it should be understood that the disclosed system or method can be implemented in other ways. For example, the above-described embodiments of the invention are only illustrative, for example, the division of modules is only a logical function division, and there may be other division methods in actual implementation.

[0032] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, and may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0033] In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0034] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the basic features of the present invention.

[0035] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. Intelligent building safety supervision system based on big data, characterized by: include: Acquisition module: includes multiple cameras, which collect image data and video data in real time; And used to transmit the collected image data and video data to the data storage module; Data storage module: used to store the image data and video data transmitted by the acquisition module; Material pile supervision module: used to supervise the material piles in the construction materials, identify whether the stacking height of the material pile meets the stacking height requirements, if not, trigger an early warning, and determine whether the material pile meets the neat stacking requirements, if not, trigger an early warning.

2. The intelligent building safety supervision system based on big data according to claim 1 is characterized in that: The material pile supervision module includes: Deep learning model generation unit: adopts deep learning algorithm to analyze the image data and video data in the data storage module to identify whether the stacking height of the material pile in the continuous frames meets the stacking height requirements; Neat stacking judgment unit: In the deep learning model generation unit, the deep learning algorithm uses the Canny edge detection algorithm to determine the outline of the material pile, the boundary and stacking direction of the material pile, and the Canny edge detection algorithm to determine whether the material pile is stacked neatly.

3. The intelligent building safety supervision system based on big data according to claim 2 is characterized in that: The stockpile supervision module also includes: Material pile height warning unit: used to determine the material pile height. If it exceeds the set threshold, an early warning is triggered.

4. The intelligent building safety supervision system based on big data according to claim 2 or 3 is characterized in that: In the deep learning model generation unit, for the image data and video data in the data storage module, the pixels belonging to the material pile are directly identified through the deep learning algorithm, and the pixels belonging to the material pile are marked as a specific category, and the remaining background pixels are marked as another category; A material pile segmentation map is generated from image data and video data by labeling pixels belonging to the material pile as specific categories.

5. The intelligent building safety supervision system based on big data according to claim 2 is characterized in that: In the deep learning model generation unit, the material categories in the material pile segmentation map are identified through a deep learning algorithm, and different height thresholds are set according to different material categories.

6. The intelligent building safety supervision system based on big data according to claim 3 is characterized in that: In the material pile stacking height warning unit, when the deep learning algorithm determines whether the material category in the material pile segmentation map exceeds the height threshold, when performing height recognition on a material pile M of a material category, the material pile segmentation map of multiple consecutive frames of the material pile M is used, and the stacking height of the material pile M in each material pile segmentation map is calculated by the deep learning algorithm to obtain a stacking height set S1, and then the Kalman filter algorithm is used to smooth the stacking height set S1 to obtain a stacking height. , and then determine the stacking height Whether it exceeds the set threshold of the material category. If so, an early warning is triggered.

7. The intelligent building safety supervision system based on big data according to claim 2 is characterized in that: In the stacking neatness judgment unit, the Canny edge detection algorithm is used to judge whether the material pile is neatly stacked. If it exceeds the set threshold, an early warning is triggered as follows: After receiving the material pile segmentation map, the input color material pile segmentation map is converted into a grayscale image. The Gaussian filtering algorithm is used to smooth the grayscale image; Through gradient calculation, the image after Gaussian filtering is convolved in the horizontal and vertical directions respectively to obtain the gradient component Gx in the horizontal direction and the gradient component Gy in the vertical direction; Calculate the gradient magnitude G at each pixel position: ； Calculate the gradient direction : ; The value range is 0°-360°; Calculate the gradient direction of each pixel position; Through gradient calculation, the position of pixel grayscale change in the material pile segmentation map is located, the outline of the material pile is outlined, and the stacking boundary of the material pile is determined; Then, based on the edge orientation and continuity, the Hough transform technique is used to determine the stacking direction of the material pile; Next, the angle deviation between the boundary line segment of the material pile outline and the straight line of the stacking direction is calculated. If the angle deviation exceeds the set threshold, it is determined that the material pile is not stacked neatly, and an early warning is triggered immediately. Calculate the length consistency deviation between the boundary line segment of the material pile outline and the stacking boundary. If the length consistency deviation exceeds the set threshold, it is determined that the material pile is not stacked neatly and an early warning is triggered immediately. Calculate the straightness of the stacking boundary. If the straightness of the stacking boundary exceeds the set threshold, it is determined that the material pile is not stacked neatly and an early warning is triggered immediately.

8. The method for intelligent building safety supervision based on big data according to any one of claims 1 to 7 is characterized in that: The following steps are involved: Collect image data and video data in real time through cameras; and is used to transmit the collected image data and video data to the data storage module for data storage; For the image data and video data in the data storage module, the pixels belonging to the material pile are directly identified through the deep learning algorithm, and the pixels belonging to the material pile are marked as a specific category, and the remaining background pixels are marked as another category; Generate a material pile segmentation map from the image data and the video data by labeling pixels belonging to the material pile as specific categories; Next, the Canny edge detection algorithm is used to determine the material stacking boundary and stacking direction of the material pile segmentation map. The Canny edge detection algorithm is used to determine whether the material pile is stacked neatly. If it exceeds the set threshold, an early warning is triggered; Identify material categories in the material pile segmentation map through deep learning algorithms; set different height thresholds according to different material categories; Then, a deep learning algorithm is used to determine whether the material category in the material pile segmentation diagram exceeds the height threshold. If so, an early warning is triggered; When the deep learning algorithm determines whether the material category in the material pile segmentation map exceeds the height threshold, when performing height recognition on a material pile M of a material category, the material pile segmentation map of multiple consecutive frames of the material pile M is used, and the stacking height of the material pile M in each material pile segmentation map is calculated by the deep learning algorithm to obtain a stacking height set S1, and then the Kalman filter algorithm is used to smooth the stacking height set S1 to obtain a stacking height. , and then determine the stacking height Whether it exceeds the set threshold of the material category. If so, an early warning is triggered.

Images