Construction site material tracking management method and system

Abstract

The invention discloses a construction site material tracking management method and system, and the method comprises the steps: laying video monitoring points at key nodes of a construction site, and deploying a weight sensor in a material stacking region; when it is detected that the material weight change exceeds a threshold value, key frame images in front and back time periods are obtained, a moving object is extracted through a pixel change rate algorithm, and a target object closest to a stacking point is selected; synchronously calling video data of each monitoring point, extracting a moving object of each point location, matching spatial-temporal characteristics by adopting a clustering algorithm, screening a set with the highest similarity with a target object, and generating a material moving path according to a time sequence; and performing deviation degree analysis on the path and a preset transportation route library, and triggering an abnormal alarm when a threshold value is exceeded. Through multi-source data fusion and track matching, the technical problem of construction site material transportation path offset monitoring is effectively solved, and the real-time performance and accuracy of material management are improved.

Description

Technical Field

[0001] The present invention relates to the field of visual inspection technology, and in particular to a construction site material tracking management method and system. Background Art

[0002] In construction projects, the compliant transportation and safe management of materials on site are core links in ensuring project costs and progress. In existing technologies, supervision of the material transportation process mainly relies on manual registration, fixed-point video surveillance, or single sensor monitoring, which makes it difficult to effectively deal with illegal behaviors such as illegal transportation and detour transportation. For example, traditional solutions lack a real-time and accurate comparison mechanism between the actual movement path of materials and the preset transportation route, and are unable to dynamically track the deviation of the route of materials during transportation. When materials move abnormally, due to the lack of linkage analysis between multi-source monitoring data (such as weight change data and video image data), it is difficult to quickly locate the transportation carrier directly related to the change in material status. This leads to delayed detection of illegal behaviors, frequent material loss and waste, and seriously affects the safety and efficiency of construction site material management.

[0003] Existing technologies have significant deficiencies in restoring material transport trajectories and detecting abnormal behavior. On the one hand, insufficient analysis of the temporal correlation between weight changes in the material stacking area and surrounding video surveillance makes it impossible to accurately capture key moving targets based on weight mutation nodes, resulting in blindness in the initial positioning of the tracking target. On the other hand, the collaborative processing capabilities of multiple video monitoring points are weak, and there is a lack of feature association and trajectory fitting technology for cross-regional moving objects. The generated material movement paths have poor continuity, and the identification of key nodes (such as forks and turns) is fuzzy, making it difficult to accurately calculate the path deviation and trigger abnormal alarms in a timely manner. The above problems result in the existing system being poorly effective in monitoring violations during material transportation, and unable to meet the actual needs of intelligent construction sites for full-process material supervision. Summary of the Invention

[0004] In order to solve at least one of the above-mentioned technical problems, the present invention provides a construction site material tracking and management method and system.

[0005] In a first aspect, the present invention provides a construction site material tracking and management method, the method comprising:

[0006] Based on setting up multiple video monitoring points on the construction site structure and setting up material monitoring points in the material stacking area, real-time video stream data is collected;

[0007] When the material monitoring point detects that the weight change of the material exceeds a first threshold, a plurality of key frame images of a preset time length before and after the change time point are acquired;

[0008] Extract multiple suspected moving objects from the key frame image based on the pixel change rate algorithm, and select the suspected moving object closest to the material stacking area as the target object;

[0009] Retrieve video data from all video monitoring points within a preset time window and extract suspected moving objects at each monitoring point;

[0010] Perform cluster matching on suspected moving objects at each video monitoring point, screen out the target set with the highest feature similarity to the target object, and generate the material movement path based on the time sequence arrangement of the objects in the target set;

[0011] The moving path is compared with the preset transport route library. If the deviation exceeds the second threshold, an abnormal alarm is triggered.

[0012] Preferably, cluster matching is performed on the suspected moving objects at each video monitoring point to screen out a target set with the highest feature similarity to the target object, including:

[0013] extracting a plurality of pixel difference sub-images corresponding to each of the suspected moving objects, wherein the pixel difference sub-images are generated by performing adjacent frame difference processing on the key frame images;

[0014] Mixing the pixel difference sub-image corresponding to the suspected moving object at each video monitoring point with the pixel difference sub-image of the target object to generate a mixed image set;

[0015] Performing clustering processing on the mixed image set using a clustering algorithm to obtain multiple classification sets containing different image feature distributions;

[0016] Calculating the distribution ratio of the pixel difference sub-images of the target object in each classification set, and marking the classification set whose distribution ratio exceeds a third preset threshold as the target object classification set;

[0017] The pixel difference sub-images of the suspected moving objects at each video monitoring point are traversed, and their distribution proportions in the target object classification set are counted, and the suspected moving objects whose proportions exceed a fourth preset threshold are classified into the target set.

[0018] Preferably, extracting a plurality of pixel difference sub-images corresponding to each suspected moving object includes:

[0019] Grayscale processing is performed on multiple key frame images;

[0020] Calculate the pixel difference matrix between adjacent key frame images and generate a binary difference image based on a preset difference threshold;

[0021] performing corrosion and expansion processing on the binary difference image;

[0022] Connected region analysis is performed on the binary difference image, and connected regions whose areas exceed a preset threshold are marked as pixel difference sub-images.

[0023] Preferably, generating the material movement path according to the time sequence arrangement of the objects in the target set includes:

[0024] Obtain geographic coordinate information of each video monitoring point;

[0025] Establish a temporal index based on the appearance time of the target object in the video of each monitoring point;

[0026] The cubic spline interpolation algorithm is used to fit the coordinates of discrete monitoring points to generate the material movement path.

[0027] Preferably, the video monitoring point is set at a fork in the construction site.

[0028] In a second aspect, the present invention further provides a construction site material tracking and management system, the system comprising:

[0029] The video monitoring module is used to collect video stream data in real time based on multiple video monitoring points set up in the construction site structure and material monitoring points set up in the material storage area; when the material monitoring point detects that the weight change of the material exceeds a first threshold, multiple key frame images with a preset time length before and after the change time point are obtained;

[0030] Suspected object recognition module, used to extract multiple suspected moving objects from the key frame image based on the pixel change rate algorithm, and select the suspected moving object closest to the material stacking area as the target object;

[0031] The global video analysis module is used to retrieve the video data of all video monitoring points within a preset time window and extract the suspected moving objects at each monitoring point;

[0032] The object matching and path generation module is used to cluster and match suspected moving objects at each video monitoring point, screen out the target set with the highest feature similarity with the target object, and generate the material movement path based on the time sequence arrangement of the objects in the target set;

[0033] The path comparison and alarm module is used to compare the moving path with the preset transportation route library. If the deviation exceeds the second threshold, an abnormal alarm is triggered.

[0034] Preferably, the object matching and path generation module is further used to:

[0035] extracting a plurality of pixel difference sub-images corresponding to each of the suspected moving objects, wherein the pixel difference sub-images are generated by performing adjacent frame difference processing on the key frame images;

[0036] Mixing the pixel difference sub-image corresponding to the suspected moving object at each video monitoring point with the pixel difference sub-image of the target object to generate a mixed image set;

[0037] Performing clustering processing on the mixed image set using a clustering algorithm to obtain multiple classification sets containing different image feature distributions;

[0038] Calculating the distribution ratio of the pixel difference sub-images of the target object in each classification set, and marking the classification set whose distribution ratio exceeds a third preset threshold as the target object classification set;

[0039] The pixel difference sub-images of the suspected moving objects at each video monitoring point are traversed, and their distribution proportions in the target object classification set are counted, and the suspected moving objects whose proportions exceed a fourth preset threshold are classified into the target set.

[0040] Preferably, the object matching and path generation module is further used to:

[0041] Obtain geographic coordinate information of each video monitoring point;

[0042] Establish a temporal index based on the appearance time of the target object in the video of each monitoring point;

[0043] The cubic spline interpolation algorithm is used to fit the coordinates of discrete monitoring points to generate the material movement path.

[0044] In a third aspect, the present invention also provides an electronic device comprising a processor and a memory, wherein the memory is used to store computer program code, and the computer program code comprises computer instructions. When the processor executes the computer instructions, the electronic device executes the method as described in the first aspect above and any possible implementation thereof.

[0045] In a fourth aspect, the present invention also provides a computer-readable storage medium, in which a computer program is stored. The computer program includes program instructions, and when the program instructions are executed by a processor of an electronic device, the processor executes the method as described in the first aspect above and any possible implementation method thereof.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] The present invention realizes all-round and real-time monitoring of materials on the construction site by setting up multiple video monitoring points in the construction site structure and setting up material monitoring points in the material stacking area. Combined with the real-time weight detection of high-precision pressure sensors, abnormal changes in materials can be captured in time. Suspected moving objects are extracted based on the pixel change rate algorithm, and the object closest to the material stacking area is selected as the target object, which effectively narrows the tracking range and improves the accuracy of target positioning. The suspected moving objects at each video monitoring point are clustered and matched, and the target set with the highest similarity to the target object characteristics is screened out, and the material movement path is generated according to the time sequence arrangement, so that the movement trajectory of the material is clear and traceable. By comparing the movement path with the preset transportation route library and triggering an abnormal alarm, violations in material transportation, such as illegal transportation, detour transportation, etc., can be discovered in time, effectively preventing material loss and waste, and improving the safety and efficiency of construction site material management.

[0048] It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the background technology, the drawings required for use in the embodiments of the present invention or the background technology will be described below.

[0050] The accompanying drawings herein are incorporated into and constitute a part of the specification. These drawings illustrate embodiments consistent with the present disclosure and, together with the specification, are used to explain the technical solutions of the present disclosure.

[0051] Figure 1 A schematic diagram of a flow chart of a construction site material tracking and management method provided by an embodiment of the present invention;

[0052] Figure 2 This is a structural diagram of a construction site material tracking and management system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0053] In order to enable those skilled in the art to better understand the solutions of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0054] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present invention. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute a separate or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0055] Existing construction site material transportation supervision technology has insufficient linkage analysis of multi-source monitoring data, weak ability to accurately track movement paths and detect abnormal behavior, resulting in difficulty in timely detection of illegal behaviors such as illegal transportation and detour transportation, and low material management safety and efficiency.

[0056] See also Figure 1 , Figure 1 The following is a flow chart of a construction site material tracking and management method provided by an embodiment of the present invention. Figure 1 As shown, the method includes:

[0057] S100, based on setting up multiple video monitoring points on the construction site structure and setting up material monitoring points in the material stacking area, collects video stream data in real time;

[0058] Monitoring point layout planning is based on the construction site's 3D BIM model, ensuring full coverage of material storage areas, main transportation routes, and key work areas. For example, at a deep foundation pit construction site, a rotatable high-definition pan-tilt camera is deployed at the edge of the pit, combined with a fixed camera at the bottom of the pit, to form a three-dimensional monitoring network, effectively eliminating the blind spots in traditional two-dimensional deployments.

[0059] S200, when the material monitoring point detects that the weight change of the material exceeds a first threshold, a plurality of key frame images of a preset time length before and after the change time point are acquired;

[0060] Monitoring points deployed in material storage areas utilize arrays of pressure sensors, which monitor weight changes in real time. Using a weight conversion model, they accurately measure loose materials (such as sand, gravel, and cement). If the data deviation exceeds 5%, a manual verification process is triggered. Using an adaptive keyframe extraction algorithm based on motion entropy, when a material monitoring point detects a weight change, the rate of change is calculated and the time window size is dynamically adjusted based on the rate of change (a larger rate of change results in a smaller window size). For example, if a rapid decrease in steel is detected, 15 seconds of video before and after the weight change is captured; for slowly depleting cement, 60 seconds of video before and after the weight change is captured.

[0061] S300, extract multiple suspected moving objects from the key frame image based on the pixel change rate algorithm, and select the suspected moving object closest to the material stacking area as the target object. In this embodiment, the closest refers to each suspected moving object. Calculate the center of gravity of its connected domain Center with material storage area Euclidean distance

[0062]

[0063] Pick The smallest suspected moving object is taken as the target object. By default, the above distance is calculated in the image coordinate plane; when there is a tie, the weight change trigger moment is preferred. The object that first appears earlier in the preset time window before and after is selected; if they are still tied, the object with the larger circumscribed rectangle is selected to suppress the small noise target;

[0064] When extracting multiple suspected moving objects from keyframe images based on a pixel change rate algorithm, the keyframe images are preprocessed, including Gaussian filtering to remove image noise and calculating pixel differences between adjacent frames to generate a difference image. The difference image is binarized by setting an appropriate threshold, and regions where pixel changes exceed the threshold are marked as suspected moving regions. Morphological operations such as erosion and dilation are performed on the binarized image to eliminate small noisy regions and connect adjacent moving regions. Contour detection and analysis are used to extract the complete suspected moving objects. When selecting the suspected moving object closest to the material storage area as the target object, the Euclidean distance between the center of gravity of each suspected moving object and the center of the material storage area is calculated, and the object with the smallest distance is selected as the target object.

[0065] S400, retrieving video data of all video monitoring points within a preset time window, and extracting suspected moving objects at each monitoring point;

[0066] When retrieving video data from all monitoring points within a preset time window, the system determines the time window before and after the preset time period based on the time point of the material weight change, and retrieves the corresponding video data from the video storage server. When extracting suspected moving objects at each monitoring point, the same pixel change rate algorithm as previously used is used, but the algorithm parameters are adaptively adjusted based on the video characteristics of each monitoring point, such as lighting conditions and viewing angle.

[0067] S500: Cluster matching is performed on the suspected moving objects at each video monitoring point, and the target set with the highest feature similarity to the target object is screened out. The material movement path is generated based on the time sequence arrangement of the objects in the target set;

[0068] When clustering and matching suspected moving objects at each video monitoring point, the feature vector of each suspected moving object is extracted, including color histogram, texture features, and shape features. The K-means clustering algorithm is then used to cluster the feature vectors, grouping similar objects into the same category. By calculating the similarity between the target object's feature vector and each category center, the target set with the highest feature similarity to the target object is selected. When generating the material movement path based on the time-series arrangement of the objects in the target set, a time-space interpolation algorithm is used to generate a continuous material movement path, combining the geographic location information of each video monitoring point and the chronological order of the objects' appearance at different monitoring points.

[0069] S600: Compare the moving path with a preset transport route library. If the deviation exceeds a second threshold, an abnormality alarm is triggered.

[0070] When comparing the movement path with a library of pre-set transport routes, the system first calculates the geometric similarity between the actual movement path and each pre-set route, using metrics such as Hausdorff distance or Fréchet distance. If the deviation exceeds a second threshold, an abnormality alarm is triggered. Abnormality alarms include audible and visual alarms, SMS notifications, and system push notifications, ensuring that relevant personnel are promptly notified of any abnormal material movement.

[0071] In this embodiment, by setting up multiple video monitoring points in the construction site structure and material monitoring points in the material storage area, all-round, real-time monitoring of construction site materials is achieved. The 360-degree rotation and zoom functions of the high-definition panoramic camera, combined with the real-time weight detection of the high-precision pressure sensor, can promptly capture abnormal changes in materials. Based on the pixel change rate algorithm, suspected moving objects are extracted, and the object closest to the material storage area is selected as the target object, which effectively narrows the tracking range and improves the accuracy of target positioning. The suspected moving objects at each video monitoring point are clustered and matched, and the target set with the highest feature similarity to the target object is screened out. The material movement path is generated according to the time sequence arrangement, making the material movement trajectory clear and traceable. By comparing the movement path with the preset transportation route library and triggering an abnormal alarm, violations in material transportation, such as illegal transportation and detour transportation, can be discovered in a timely manner, effectively preventing material loss and waste, and improving the safety and efficiency of construction site material management.

[0072] Preferably, cluster matching is performed on the suspected moving objects at each video monitoring point to screen out a target set with the highest feature similarity to the target object, including:

[0073] extracting a plurality of pixel difference sub-images corresponding to each of the suspected moving objects, wherein the pixel difference sub-images are generated by performing adjacent frame difference processing on the key frame images;

[0074] Mixing the pixel difference sub-image corresponding to the suspected moving object at each video monitoring point with the pixel difference sub-image of the target object to generate a mixed image set;

[0075] Performing clustering processing on the mixed image set using a clustering algorithm to obtain multiple classification sets containing different image feature distributions;

[0076] Calculating the distribution ratio of the pixel difference sub-images of the target object in each classification set, and marking the classification set whose distribution ratio exceeds a third preset threshold as the target object classification set;

[0077] The pixel difference sub-images of the suspected moving objects at each video monitoring point are traversed, and their distribution proportions in the target object classification set are counted, and the suspected moving objects whose proportions exceed a fourth preset threshold are classified into the target set.

[0078] To extract multiple pixel difference sub-images corresponding to each suspected moving object, multiple keyframe images are grayscaled, converting color images into grayscale images to simplify subsequent processing steps. To calculate the pixel difference matrix between adjacent keyframe images, the absolute difference between the grayscale values ​​of the two frames is calculated pixel by pixel to obtain the difference matrix. A binary difference image is generated based on a preset difference threshold. Pixels in the difference matrix greater than the threshold are set to 255 (white), and pixels less than or equal to the threshold are set to 0 (black). Erosion and dilation are performed on the binary difference image. Erosion is first performed using a 3×3 structuring element to remove small noise points. Dilation is then performed to restore the eroded target area, resulting in a more complete outline of the moving object. Connected component analysis is performed on the binary difference image. Using a connected component marking algorithm, interconnected white pixel regions are marked as connected regions, and the area of ​​each connected region is calculated. Connected regions with an area exceeding the preset threshold are marked as pixel difference sub-images to ensure that meaningful moving object regions are extracted.

[0079] The pixel-difference sub-images corresponding to the suspected moving objects at each video monitoring point are blended with the pixel-difference sub-images of the target object to generate a blended image set. During the blending process, the identification information of the monitoring point from which each sub-image originates is added to facilitate subsequent analysis. When clustering the blended image set, the DBSCAN density clustering algorithm is selected. This algorithm automatically determines the number of clusters based on the density of data points and identifies noise points. By setting appropriate neighborhood radius and minimum point count parameters, the sub-images in the blended image set are divided into multiple classification sets containing different image feature distributions.

[0080] When calculating the distribution ratio of the pixel difference sub-image of the target object in each classification set, count the number of times the sub-image of the target object appears in each classification set, and divide it by the total number of target object sub-images to obtain the distribution ratio. The classification sets whose distribution ratio exceeds the third preset threshold are marked as target object classification sets. These classification sets contain sub-images with similar characteristics to the target object. Traverse the pixel difference sub-images of the suspected mobile objects at each video monitoring point, count their distribution ratios in the target object classification set, and classify the suspected mobile objects whose ratio exceeds the fourth preset threshold into the target set. In this way, suspected mobile objects with high similarity to the target object characteristics can be accurately screened out, thereby improving the accuracy of material tracking.

[0081] In this embodiment, by graying the key frame image, calculating pixel differences, binarizing, corroding and dilating, and analyzing connected regions, accurate pixel difference sub-images are extracted, providing a reliable data basis for subsequent cluster matching. The pixel difference sub-images of the suspected mobile objects at each monitoring point are mixed with the sub-images of the target object, and processed using a clustering algorithm, which can effectively identify suspected mobile objects with similar features to the target object. The distribution ratio of the target object sub-image in each classification set is calculated, and the classification set with a ratio exceeding a threshold is marked as the target object classification set, further improving the accuracy of the screening. The suspected mobile object sub-images of each monitoring point are traversed, and their distribution ratio in the target object classification set is counted, and objects with a ratio exceeding a threshold are classified into the target set, ensuring that the objects in the target set have a high degree of feature similarity with the target object, thereby improving the reliability of material tracking.

[0082] Preferably, extracting a plurality of pixel difference sub-images corresponding to each suspected moving object includes:

[0083] Grayscale processing is performed on multiple key frame images;

[0084] Calculate the pixel difference matrix between adjacent key frame images and generate a binary difference image based on a preset difference threshold;

[0085] performing corrosion and expansion processing on the binary difference image;

[0086] Connected region analysis is performed on the binary difference image, and connected regions whose areas exceed a preset threshold are marked as pixel difference sub-images.

[0087] When grayscaling multiple keyframe images, a weighted average method is used to convert color images into grayscale images. The pixel difference matrix between adjacent keyframe images is calculated pixel by pixel. For two adjacent images I1(x,y) and I2(x,y), the pixel difference matrix D(x,y) is calculated as D(x,y) = |I1(x,y) - I2(x,y)|. When generating a binary difference image based on a preset difference threshold, each pixel value in the difference matrix D(x,y) is compared with a preset threshold T. If D(x,y) > T, the corresponding pixel in the binary image B(x,y) is set to 255 (white), indicating a significant change; otherwise, it is set to 0 (black). When eroding and dilating the binary difference image, a 3×3 square structuring element is used. The erosion operation slides the structuring element across the binary image. The center pixel of the structuring element remains white only when it is completely contained within the foreground (white pixels); otherwise, it becomes black. The dilation operation turns the central pixel of a structuring element white when it intersects the foreground area. By performing erosion followed by dilation, small noise points can be effectively removed and adjacent foreground areas can be connected, making the outline of the moving object more complete. A scanline-based connected region labeling algorithm is used for connected region analysis of the binary difference image. Starting from the upper left corner of the image, the image is scanned line by line. Whenever a white pixel is encountered, a label value is assigned to it, and the connected region in which the pixel resides is recorded. During the scanning process, if adjacent white pixels are found to belong to different labels, these labels are merged into the same connected region. After the scan is completed, the area of ​​each connected region is calculated, and connected regions with an area exceeding a preset threshold are marked as pixel difference sub-images. This method can accurately extract the key feature regions of moving objects, providing a reliable data foundation for subsequent cluster matching.

[0088] In this embodiment, the key frame image is grayscaled, which simplifies the subsequent image processing steps while retaining the main brightness information of the image. By calculating the pixel difference matrix between adjacent frames and generating a binary difference image, the moving area in the image can be highlighted, facilitating subsequent analysis. The binary difference image is eroded and expanded to effectively remove small noise points and connect adjacent moving areas, making the outline of the moving object more complete. The binary difference image is subjected to connected area analysis, and the connected areas with an area exceeding a preset threshold are marked as pixel difference sub-images, ensuring that meaningful moving object areas are extracted and eliminating irrelevant background interference. The comprehensive application of these technical means improves the accuracy and reliability of pixel difference sub-image extraction, providing high-quality data for subsequent cluster matching and material tracking.

[0089] Preferably, the material movement path is generated according to the time sequence arrangement of the objects in the target set, including:

[0090] Obtain geographic coordinate information of each video monitoring point;

[0091] Establish a temporal index based on the appearance time of the target object in the video of each monitoring point;

[0092] The cubic spline interpolation algorithm is used to fit the coordinates of discrete monitoring points to generate the material movement path.

[0093] To obtain the geographic coordinates of each video monitoring point, GPS positioning technology combined with total station measurement is used. A GPS receiver is used to obtain the approximate longitude and latitude coordinates of each monitoring point. A total station is then used to perform precise measurements, determining the precise location of the monitoring point within the construction site coordinate system with millimeter-level accuracy. This geographic coordinate information is stored in the system database for subsequent use in generating material movement paths. To create a temporal index based on the time of appearance of the target object in the video at each monitoring point, the timestamps of the first and last appearance of the target object at each monitoring point are recorded and sorted chronologically. This temporal index clearly defines the target object's movement sequence and time interval between different monitoring points, providing a temporal basis for generating an accurate movement path. A cubic spline interpolation algorithm is used to fit the path to the discrete monitoring point coordinates. The geographic coordinates of each monitoring point are used as control points, and the coefficients of the cubic spline function are calculated to generate a smooth curve. The cubic spline interpolation algorithm has second-order continuous derivatives, ensuring that the generated path curve exhibits excellent smoothness and continuity at the control points. During the interpolation process, the target object's travel time between monitoring points is considered and the path is time-weighted, making the generated movement path more consistent with the actual situation. The material movement path generated by this method not only accurately reflects the material's actual movement trajectory, but also provides a reliable basis for subsequent path analysis and anomaly detection.

[0094] In this embodiment, the geographic coordinate information of each video monitoring point is obtained to provide a spatial reference for generating an accurate material movement path. A time series index is established based on the appearance time of the target object in the video of each monitoring point, which clearly reflects the movement order and time interval of the material, and provides a time basis for path fitting. The cubic spline interpolation algorithm is used to fit the path of discrete monitoring point coordinates. The generated smooth curve can not only accurately connect each monitoring point, but also reflect the actual movement trajectory of the material between the monitoring points. The path is time-weighted considering the movement time of the target object, so that the generated movement path is more in line with the actual situation. The combination of these technical means generates a continuous, smooth and accurate material movement path, which provides a reliable basis for subsequent path analysis and abnormality judgment, and helps to improve the level of refinement of construction site material management.

[0095] Preferably, the video monitoring point is set at a fork in the construction site.

[0096] When setting up video monitoring points at construction site forks, we consider that forks are critical nodes in material transportation routes, where materials may choose different transportation directions. Therefore, high-definition cameras are installed above or to the side of each fork to ensure clear capture of the movement of material transport vehicles or personnel. For example, at a T-shaped fork, cameras are installed above the roads in three directions, forming a comprehensive monitoring network. The deployment of these video monitoring points effectively improves the accuracy and completeness of material tracking. As materials pass through the fork, cameras in different directions record their movement from multiple angles, providing richer visual information. Analysis of this multi-view video data enables more accurate determination of the material's direction of transport, avoiding misjudgments caused by the limitations of a single perspective. Furthermore, in the subsequent cluster matching and path generation processes, the data provided by the fork monitoring points can better connect material movement information from different road sections, making the generated material movement path more coherent and accurate. The placement of video monitoring points at construction site forks also enables timely detection of unusual stops or diversions in material transport, providing strong support for efficient material management on the site.

[0097] In summary, the method provided in this embodiment can at least achieve the following effects:

[0098] The present invention realizes all-round and real-time monitoring of materials on the construction site by setting up multiple video monitoring points in the construction site structure and setting up material monitoring points in the material stacking area. Combined with the real-time weight detection of high-precision pressure sensors, abnormal changes in materials can be captured in time. Suspected moving objects are extracted based on the pixel change rate algorithm, and the object closest to the material stacking area is selected as the target object, which effectively narrows the tracking range and improves the accuracy of target positioning. The suspected moving objects at each video monitoring point are clustered and matched, and the target set with the highest similarity to the target object characteristics is screened out, and the material movement path is generated according to the time sequence arrangement, so that the movement trajectory of the material is clear and traceable. By comparing the movement path with the preset transportation route library and triggering an abnormal alarm, violations in material transportation, such as illegal transportation, detour transportation, etc., can be discovered in time, effectively preventing material loss and waste, and improving the safety and efficiency of construction site material management.

[0099] See also Figure 2 In one embodiment, a construction site material tracking and management system is provided, the system comprising:

[0100] The video monitoring module 100 is configured to collect video stream data in real time by setting up multiple video monitoring points on the construction site structure and setting up material monitoring points in the material storage area; when the material monitoring point detects that the weight change of the material exceeds a first threshold, it obtains multiple key frame images of a preset time length before and after the change time point;

[0101] Suspected object identification module 200, for extracting multiple suspected moving objects from the key frame image based on a pixel change rate algorithm, and selecting the suspected moving object closest to the material stacking area as the target object;

[0102] The global video analysis module 300 is used to retrieve the video data of all video monitoring points within a preset time window and extract the suspected moving objects at each monitoring point;

[0103] The object matching and path generation module 400 is used to perform cluster matching on the suspected moving objects at each video monitoring point, screen out the target set with the highest feature similarity to the target object, and generate the material movement path based on the time sequence arrangement of the objects in the target set;

[0104] The path comparison and alarm module 500 is used to compare the moving path with the preset transportation route library, and trigger an abnormal alarm if the deviation exceeds a second threshold.

[0105] Preferably, the object matching and path generation module 400 is further configured to:

[0106] extracting a plurality of pixel difference sub-images corresponding to each of the suspected moving objects, wherein the pixel difference sub-images are generated by performing adjacent frame difference processing on the key frame images;

[0107] Mixing the pixel difference sub-image corresponding to the suspected moving object at each video monitoring point with the pixel difference sub-image of the target object to generate a mixed image set;

[0108] Performing clustering processing on the mixed image set using a clustering algorithm to obtain multiple classification sets containing different image feature distributions;

[0109] Calculating the distribution ratio of the pixel difference sub-images of the target object in each classification set, and marking the classification set whose distribution ratio exceeds a third preset threshold as the target object classification set;

[0110] The pixel difference sub-images of the suspected moving objects at each video monitoring point are traversed, and their distribution proportions in the target object classification set are counted, and the suspected moving objects whose proportions exceed a fourth preset threshold are classified into the target set.

[0111] Preferably, the object matching and path generation module 400 is further configured to:

[0112] Obtain geographic coordinate information of each video monitoring point;

[0113] Establish a temporal index based on the appearance time of the target object in the video of each monitoring point;

[0114] The cubic spline interpolation algorithm is used to fit the coordinates of discrete monitoring points to generate the material movement path.

[0115] It can be understood that the functions or modules included in the system provided in this embodiment can be used to execute the method described in the above method embodiment. Its specific implementation can refer to the description of the above method embodiment. For the sake of brevity, it will not be repeated here.

[0116] The present invention also provides an electronic device, including a processor and a memory, wherein the memory is used to store computer program code, and the computer program code includes computer instructions. When the processor executes the computer instructions, the electronic device executes a method as described in any one of the possible implementation modes.

[0117] The present invention also provides a computer-readable storage medium, in which a computer program is stored. The computer program includes program instructions. When the program instructions are executed by a processor of an electronic device, the processor executes a method as described in any one of the possible implementation methods described above.

[0118] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present invention.

[0119] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here. Those skilled in the art will also clearly understand that the descriptions of the various embodiments of the present invention have different focuses. For the convenience and brevity of description, the same or similar parts may not be repeated in different embodiments. Therefore, for parts not described or not described in detail in a certain embodiment, reference can be made to the descriptions of other embodiments.

Claims

1. A construction site material tracking and management method, characterized in that: The method comprises: Based on setting up multiple video monitoring points on the construction site structure and setting up material monitoring points in the material stacking area, real-time video stream data is collected; When the material monitoring point detects that the weight change of the material exceeds a first threshold, a plurality of key frame images of a preset time length before and after the change time point are acquired; Extract multiple suspected moving objects from the key frame image based on the pixel change rate algorithm, and select the suspected moving object closest to the material stacking area as the target object; Retrieve video data from all video monitoring points within a preset time window and extract suspected moving objects at each monitoring point; Perform cluster matching on suspected moving objects at each video monitoring point, screen out the target set with the highest feature similarity to the target object, and generate the material movement path based on the time sequence arrangement of the objects in the target set; The moving path is compared with the preset transport route library. If the deviation exceeds the second threshold, an abnormal alarm is triggered.

2. The construction site material tracking and management method according to claim 1, characterized in that: The cluster matching of the suspected moving objects at each video monitoring point is performed to screen out the target set with the highest feature similarity to the target object, including: extracting a plurality of pixel difference sub-images corresponding to each of the suspected moving objects, wherein the pixel difference sub-images are generated by performing adjacent frame difference processing on the key frame images; Mixing the pixel difference sub-image corresponding to the suspected moving object at each video monitoring point with the pixel difference sub-image of the target object to generate a mixed image set; Performing clustering processing on the mixed image set using a clustering algorithm to obtain multiple classification sets containing different image feature distributions; Calculating the distribution ratio of the pixel difference sub-images of the target object in each classification set, and marking the classification set whose distribution ratio exceeds a third preset threshold as the target object classification set; The pixel difference sub-images of the suspected moving objects at each video monitoring point are traversed, and their distribution proportions in the target object classification set are counted, and the suspected moving objects whose proportions exceed a fourth preset threshold are classified into the target set.

3. The construction site material tracking and management method according to claim 2, characterized in that: The extracting of a plurality of pixel difference sub-images corresponding to each of the suspected moving objects includes: Grayscale processing is performed on multiple key frame images; Calculate the pixel difference matrix between adjacent key frame images and generate a binary difference image based on a preset difference threshold; performing corrosion and expansion processing on the binary difference image; Connected region analysis is performed on the binary difference image, and connected regions whose areas exceed a preset threshold are marked as pixel difference sub-images.

4. The construction site material tracking and management method according to claim 1, characterized in that: Generating a material movement path according to the time sequence arrangement of objects in the target set includes: Obtain geographic coordinate information of each video monitoring point; Establish a temporal index based on the appearance time of the target object in the video of each monitoring point; The cubic spline interpolation algorithm is used to fit the coordinates of discrete monitoring points to generate the material movement path.

5. The construction site material tracking and management method according to claim 1, characterized in that: The video monitoring point is set at the fork in the construction site.

6. A construction site material tracking and management system, characterized in that: The system comprises: The video monitoring module is used to collect video stream data in real time based on multiple video monitoring points set up in the construction site structure and material monitoring points set up in the material storage area; when the material monitoring point detects that the weight change of the material exceeds a first threshold, multiple key frame images with a preset time length before and after the change time point are obtained; Suspected object recognition module, used to extract multiple suspected moving objects from the key frame image based on the pixel change rate algorithm, and select the suspected moving object closest to the material stacking area as the target object; The global video analysis module is used to retrieve the video data of all video monitoring points within a preset time window and extract the suspected moving objects at each monitoring point; The object matching and path generation module is used to cluster and match suspected moving objects at each video monitoring point, screen out the target set with the highest feature similarity with the target object, and generate the material movement path based on the time sequence arrangement of the objects in the target set; The path comparison and alarm module is used to compare the moving path with the preset transportation route library. If the deviation exceeds the second threshold, an abnormal alarm is triggered.

7. The construction site material tracking and management system according to claim 6, characterized in that: The object matching and path generation module is also used to: extracting a plurality of pixel difference sub-images corresponding to each of the suspected moving objects, wherein the pixel difference sub-images are generated by performing adjacent frame difference processing on the key frame images; Mixing the pixel difference sub-image corresponding to the suspected moving object at each video monitoring point with the pixel difference sub-image of the target object to generate a mixed image set; Performing clustering processing on the mixed image set using a clustering algorithm to obtain multiple classification sets containing different image feature distributions; Calculating the distribution ratio of the pixel difference sub-images of the target object in each classification set, and marking the classification set whose distribution ratio exceeds a third preset threshold as the target object classification set; The pixel difference sub-images of the suspected moving objects at each video monitoring point are traversed, and their distribution proportions in the target object classification set are counted, and the suspected moving objects whose proportions exceed a fourth preset threshold are classified into the target set.

8. The construction site material tracking and management system according to claim 6, characterized in that: The object matching and path generation module is also used to: Obtain geographic coordinate information of each video monitoring point; Establish a temporal index based on the appearance time of the target object in the video of each monitoring point; The cubic spline interpolation algorithm is used to fit the coordinates of discrete monitoring points to generate the material movement path.

9. An electronic device, characterized in that: include: A processor and a memory, the memory is used to store computer program code, the computer program code includes computer instructions, when the processor executes the computer instructions, the electronic device executes the construction site material tracking and management method according to any one of claims 1 to 5.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which includes program instructions. When the program instructions are executed by a processor of an electronic device, the processor executes the construction site material tracking and management method according to any one of claims 1 to 5.

Images