A monitoring method for steel bar management at construction site

Abstract

The present invention provides a monitoring method for steel bar management at a construction site, which relates to the technical field of steel bar management, and aims to achieve more accurate, timely and efficient steel bar management, including using colored wireframes to divide the steel bar storage area in a steel bar warehouse, and setting a type mark for each steel bar storage area to calibrate the steel bar type of the corresponding steel bar storage area; arranging multiple image acquisition devices to acquire images of each steel bar storage area; identifying the placement position, quantity and type of steel bars in each steel bar stacking area through the top view and side view of the acquired image; judging whether there are abnormal conditions, and the abnormal conditions include the placement position of the steel bar exceeding the preset range, the wrong area for the steel bar placement and the number of steel bars of a certain type being lower than the corresponding quantity threshold. The present invention has the advantages of intelligent, efficient, high-quality and safer management of steel bars.

Description

Technical Field

[0001] The present invention relates to the technical field of steel bar management, and in particular to a monitoring method for steel bar management at a construction site. Background Art

[0002] Rebar refers to steel used for reinforced concrete and prestressed reinforced concrete. Its cross-section is round or sometimes square with rounded corners.

[0003] In the construction process, a large number of steel bars are needed. Generally, for the convenience of use, the stored steel bars need to be classified and stored and managed. Since the required steel bars are generally moved in batches according to the placement area, if there is a placement error, it is very likely that the wrong type of steel bars will be taken away or even used, which may cause serious consequences. Therefore, it is usually necessary to manually record and detect the classification, placement and quantity of steel bars, so as to purchase insufficient types of steel bars as needed and deal with incorrectly placed steel bars. The management intensity of steel bars is high, and this management method is overly dependent on manual labor and is relatively mechanized, which wastes human resources and is prone to errors.

[0004] By designing an intelligent management method for image processing, more accurate, timely and efficient steel bar management can be achieved. Summary of the invention

[0005] The purpose of the present invention is to provide a monitoring method for steel bar management at a construction site, so as to achieve more accurate, timely and efficient steel bar management.

[0006] The embodiments of the present invention are implemented by the following technical solutions:

[0007] A monitoring method for steel bar management at a construction site comprises the following steps:

[0008] The steel bar storage area in the steel bar warehouse is divided into zones using colored wireframes. Each steel bar storage area is stacked with steel bars of one type. Each steel bar storage area is provided with a type identification to mark the type of steel bars in the corresponding steel bar storage area.

[0009] Multiple image acquisition devices are set to acquire images of each steel bar storage area, and the acquired viewing angles include a top view and a side view, wherein the side view acquires the end faces of the stacked steel bars;

[0010] The placement, quantity and type of steel bars in each steel bar stacking area are identified through the top view and side view of the collected images;

[0011] Determine whether there are any abnormal situations. If not, do not perform any operation. If yes, mark and warn the steel bars with abnormal situations. The abnormal situations include the placement of the steel bars exceeding the preset range, the wrong area for placement of the steel bars, and the number of a certain type of steel bars being lower than the corresponding quantity threshold.

[0012] Preferably, the image acquisition device adopts a CCD image sensor.

[0013] Preferably, the colored wireframes of different colors are used to delineate multiple areas, and the colored wireframes are used as the category identification.

[0014] Preferably, the method for identifying the placement position, quantity and type of steel bars in each steel bar stacking area by using the top view and the side view of the collected video includes:

[0015] Extracting an image as a check frame from the acquired image;

[0016] Performing noise reduction processing on the extracted inspection frame;

[0017] Perform corrosion and dilation on the inspection frame;

[0018] In the inspection frame of the side view, the end face of each steel bar in the inspection frame is segmented by disassembling the connected domain;

[0019] The inspection frame based on the side view angle determines the number and type of steel bars through image recognition;

[0020] The inspection frame based on the top-down perspective uses image recognition to determine whether the placement of the steel bars exceeds the preset range.

[0021] Preferably, the method for judging whether the placement position of the steel bar exceeds a preset range is:

[0022] A RGB value set B within a color threshold range is set based on the RGB value of the outer surface of the steel bar;

[0023] Obtain the colored wireframe of each of the steel bar storage areas;

[0024] Determine whether the colored wireframe is blocked;

[0025] If it is not blocked, the placement position of the steel bar does not exceed the preset range;

[0026] If it is blocked, extract the pixel cluster of the blocked part, check whether the RGB value of the pixel cluster belongs to B, if it does, the placement position of the steel bar exceeds the preset range, if not, the placement position of the steel bar does not exceed the preset range;

[0027] If a pixel cluster whose RGB value belongs to B is detected in any area outside all the steel bar storage areas, the placement position of the steel bars does not exceed the preset range.

[0028] Preferably, the method of segmenting the end face of each steel bar in the inspection frame by disassembling the connected domain is:

[0029] Initialize the label value to 1, set a set of RGB values ​​A within the color threshold range based on the RGB value of the steel bar end surface, and traverse each pixel from the first row of the inspection frame from left to right. When the first pixel whose RGB value belongs to A is detected, the label value is its belonging number.

[0030] Traverse the remaining pixels in the first row in sequence. When the RGB value of the currently traversed pixel belongs to A, check whether the RGB value of the pixel on its left belongs to A. If so, the current pixel has the same belonging number as the pixel on its left. If not, update l abe l = l abe l + 1, and the l abe l value is its belonging number.

[0031] Starting from the second row, traverse each pixel point from left to right and then from top to bottom. When the RGB value of the currently traversed pixel point belongs to A, check whether at least one of the RGB values ​​of the pixel points above, to the upper left, and to the upper right belongs to A. If so, extract the belonging number of any pixel point whose RBG value belongs to A and assign it to the current pixel point. If not, update l abel = l abel + 1, and assign the l abel value to the current pixel point as its belonging number;

[0032] After the traversal is completed, the pixels with the same belonging number are divided into an initial end face partition. Each initial end face partition is traversed from left to right and then from top to bottom. Check whether there are pixels that do not belong to the steel bars between each initial end face partition and its adjacent initial end face partition. If not, the two end face partitions are merged to obtain the updated terminal end face partition. If so, no operation is performed.

[0033] Preferably, the method for identifying the number of the steel bars is: sequentially numbering each of the final end face partitions starting from 1 in order from left to right and then from top to bottom, and the total number of the final numbers is the number of steel bars.

[0034] Preferably, the method for identifying the type of the steel bar is:

[0035] Collect end face data of multiple steel bars of multiple types, obtain key feature values ​​through feature histogram analysis, and use the key feature values ​​and corresponding steel bar types to train a type recognition model;

[0036] Obtain each of the final end surface partitions and extract key feature values ​​of its graphics;

[0037] Substitute the key eigenvalues ​​into the type identification model for type identification to obtain the steel bar type corresponding to each final end face partition.

[0038] Preferably, the method for determining the regional error of the steel bar placement is:

[0039] Compare and determine in sequence whether the type of steel bars corresponding to each of the final end face partitions is consistent with the type identification of the steel bar storage area;

[0040] If they are inconsistent, the final end face partition is marked as an area placement error. If they are consistent, no operation is performed.

[0041] Preferably, a face recognition module is provided at the door of the steel bar warehouse, the face data of the warehouse manager is entered into the face data storage library, and face verification is performed when someone enters the steel bar warehouse.

[0042] The technical solution of the embodiment of the present invention has at least the following advantages and beneficial effects:

[0043] The present invention performs intelligent management of steel bars based on image recognition, and management is no longer solely dependent on manual labor, which is more efficient;

[0044] The present invention solves the problem of visual recognition errors in naked eye observation, and can more accurately ensure that the area where the steel bars are placed is correct;

[0045] The present invention also solves the problem that it is difficult to measure the number of steel bars by manual management, and can manage the number more accurately and replenish the steel bars more timely;

[0046] The present invention directly marks the type of steel bars by colored wireframes, which simplifies the setting, does not require additional props, and makes the arrangement simpler;

[0047] The invention has reasonable design, convenient data processing, low cost and is easy to promote and apply. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0049] Figure 1 A schematic diagram of a process flow provided in Example 1 of the present invention;

[0050] Figure 2A schematic diagram of an on-site steel bar storage area provided in Example 1 of the present invention;

[0051] Figure 3 A schematic diagram of the structure of a data processing device provided in Example 1 of the present invention. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0053] Example 1

[0054] See also Figure 1 This embodiment provides a monitoring method for steel bar management at a construction site, comprising the following steps:

[0055] Step S1: Use colored wireframes to divide the steel bar storage area in the steel bar warehouse. Each steel bar storage area is stacked with a type of steel bar. Each steel bar storage area is set with a type identifier to mark the steel bar type of the corresponding steel bar storage area. For details, please refer to Figure 2 ,The measured areas 1-N in the figure are different tendon storage areas;

[0056] Step S2: multiple image acquisition devices are set to acquire images of each steel bar storage area, and the acquired viewing angles include a top view and a side view, wherein the side view acquires the end face of the stacked steel bars; similarly, refer to Figure 2 ,Here, a CCD image sensor is used, which is installed in top and side view to extract the required image;

[0057] Step S3: identifying the placement position, quantity and type of steel bars in each steel bar stacking area through the top view and side view of the collected image;

[0058] Step S4: Determine whether there are any abnormal situations. If not, do not perform any operation. If yes, mark and warn the steel bars with abnormal situations. The abnormal situations include the placement of the steel bars exceeding the preset range, the wrong area for placement of the steel bars, and the number of a certain type of steel bars being lower than the corresponding quantity threshold.

[0059] In this embodiment, the image acquisition device adopts a CCD image sensor.

[0060] On the other hand, the colored wireframes of different colors are used to delimit multiple areas, and the colored wireframes are used as the category identification. Here, the colored wireframes that delimit the storage range can be used as category identification directly, and the colors are easy to identify, which can also simplify the construction on site, and there is no need to arrange traditional identification such as QR codes.

[0061] The overall embodiment of this invention is to divide different steel bar storage areas to store different types of steel bars, and to manage the storage of steel bars based on technologies such as image acquisition and image recognition, including directly identifying the number and type of steel bars in each steel bar storage area, and identifying whether the placement of the steel bars exceeds the range. Because when the number of steel bars is insufficient and there is no timely procurement, it may affect the progress of the entire project. If the steel bars are placed in the wrong steel bar storage area, when a certain type of steel bar is pulled away in batches during the construction process, the wrongly placed steel bars may be taken away or even used incorrectly, which may also cause serious consequences. Whether the placement of the steel bars exceeds the range will cause messy management of the steel bar warehouse.

[0062] When an abnormality occurs, an early warning is issued to prompt the staff to handle it. For example, when the number of a certain type of steel bars is lower than the corresponding quantity threshold, the staff is reminded to purchase them. When a steel bar is placed in the wrong storage area or the position of the steel bar is out of range, the staff is reminded to adjust it. Image recognition can manage steel bars more efficiently, timely and accurately.

[0063] It should be noted here that neat stacking is generally a basic requirement for steel bar management when storing steel bars. This aspect is also conducive to the accuracy of steel bar management by image recognition.

[0064] Finally, the data processing device of this embodiment can refer to Figure 3 Taking the CCD image sensor, that is, the industrial CCD sensor, as an example, the data collected by it is transmitted to the industrial computer as negative feedback. The industrial computer performs actions based on its data and can control the LED through the signal generator as a warning.

[0065] Example 2

[0066] This embodiment is based on the technical solution of Embodiment 1 and further illustrates the specific identification and management method of the steel bars involved in Steps S3-S4.

[0067] In this embodiment, the method for identifying the placement position, quantity and type of steel bars in each steel bar stacking area through the top view and side view of the collected video includes:

[0068] Extracting an image as a check frame from the acquired image;

[0069] Performing noise reduction processing on the extracted inspection frame;

[0070] Perform corrosion and dilation on the inspection frame;

[0071] In the inspection frame of the side view, the end face of each steel bar in the inspection frame is segmented by disassembling the connected domain;

[0072] The inspection frame based on the side view angle determines the number and type of steel bars through image recognition;

[0073] The inspection frame based on the top-down perspective uses image recognition to determine whether the placement of the steel bars exceeds the preset range.

[0074] In the above steps, noise reduction can be achieved using a variety of filters to improve the accuracy of subsequent image recognition. Image erosion can make the gaps between objects in the image larger, and image expansion can make individual objects more obvious. Here, you can expand first and then erode, so that the gaps between the stacked steel bars are larger when the steel bar body is clearer, which is helpful for subsequent image recognition. Since the steel bars are stacked, it is more convenient to find the number of end faces of a single steel bar through the stacked end faces to calculate the number of steel bars. On the other hand, the placement position is easier to judge from the top view.

[0075] Specifically, the method for judging whether the placement position of the steel bar exceeds the preset range is:

[0076] Based on the RGB value of the outer surface of the steel bar, a set B of RGB values ​​within the color threshold range is set; this step is set to take into account the color deviation caused by light, etc., but because the image has been processed by noise reduction in advance, the deviation is generally not too large, and there may be only a few values ​​in the set B. The RGB value of the outer surface of the steel bar used as a benchmark is also a value extracted from the image that can widely represent the color of the steel bar.

[0077] Obtain the colored wireframe of each of the steel bar storage areas;

[0078] Determine whether the colored wireframe is blocked;

[0079] If it is not blocked, the placement position of the steel bar does not exceed the preset range;

[0080] If it is blocked, extract the pixel cluster of the blocked part and check whether the RGB value of the pixel cluster belongs to B. If it does, the placement position of the steel bar exceeds the preset range; if it does not, the placement position of the steel bar does not exceed the preset range. Here, it is considered that other dropped objects may block the colored wireframe. This situation is not due to the placement of the steel bars, but this situation can still be used to alert the staff to check, thereby improving the accuracy of the judgment.

[0081] If a pixel cluster whose RGB value belongs to B is detected in any area outside all the steel bar storage areas, the placement position of the steel bars does not exceed the preset range.

[0082] Here, the placement beyond the preset range is divided into two situations: the first is generally due to the uneven placement, resulting in a few steel bars protruding from the colored wireframe, and the second is that the steel bars may be accidentally scattered on the ground of the steel warehouse. The steel bars that are placed beyond the preset range are marked in the image and an early warning is issued to remind the staff to deal with it.

[0083] In addition, the method of segmenting the end face of each steel bar in the inspection frame by disassembling the connected domain is:

[0084] Initialize the label value to 1, set a set of RGB values ​​A within the color threshold range based on the RGB value of the steel bar end surface, and traverse each pixel from the first row of the inspection frame from left to right. When the first pixel whose RGB value belongs to A is detected, the label value is its belonging number.

[0085] Traverse the remaining pixels in the first row in sequence. When the RGB value of the currently traversed pixel belongs to A, check whether the RGB value of the pixel on its left belongs to A. If so, the current pixel has the same belonging number as the pixel on its left. If not, update l abe l = l abe l + 1, and the l abe l value is its belonging number.

[0086] Starting from the second row, traverse each pixel point from left to right and then from top to bottom. When the RGB value of the currently traversed pixel point belongs to A, check whether at least one of the RGB values ​​of the pixel points above, to the upper left, and to the upper right belongs to A. If so, extract the belonging number of any pixel point whose RBG value belongs to A and assign it to the current pixel point. If not, update l abel = l abel + 1, and assign the l abel value to the current pixel point as its belonging number;

[0087] After the traversal is completed, the pixels with the same assigned number are divided into a final end face partition.

[0088] Since the expansion operation is performed before, the gap between the steel bars is enlarged, which is beneficial to the segmentation of the single end face in this embodiment, so it can be basically guaranteed that there must be other pixels between the stacked steel bars, so it can be divided by traversing according to the pixels, and it can be ensured that the final end face partition can represent a complete steel bar end face. When judging the pixel points, check whether at least one of the RGB values ​​of the pixels above, on the upper left, and on the upper right belongs to A. This is because the end face may be a polygon or a circle, so it cannot be done by only checking the pixels above it in the usual way. According to the method of this implementation, the amount of data processing increases slightly, but only in this way can a single steel bar end face be divided more accurately.

[0089] Furthermore, since a final end face partition represents a complete end face of a steel bar, the number of final end face partitions is actually the number of steel bars. The method for identifying the number of steel bars is to sequentially number each final end face partition from 1 in order from left to right and then from top to bottom, and the total number of final numbers is the number of steel bars. Here, dynamic tracking of the number of steel bars in each steel bar stacking area can be achieved, and when the number of steel bars in a steel bar stacking area is lower than the set quantity threshold, an early warning will be issued to remind the staff to purchase supplies.

[0090] Furthermore, since all the stacked steel bars can only be observed through the end face, the type identification is mainly based on the side view. Here, the method for identifying the type of the steel bar is as follows:

[0091] Collect end face data of multiple steel bars of multiple types, obtain key feature values ​​through feature histogram analysis, and use the key feature values ​​and corresponding steel bar types to train a type recognition model;

[0092] Obtain each of the final end surface partitions and extract key feature values ​​of its graphics;

[0093] Substitute the key eigenvalues ​​into the type identification model for type identification to obtain the steel bar type corresponding to each final end face partition.

[0094] In the above operations, the feature histogram analysis is to filter out the parameters in the end face data that are most conducive to judging the type of steel bar. It can be mainly used for the frequency distribution of the feature vector, which can help to better understand the features. Specifically, according to the collected data set, multiple types of feature values ​​can be pre-selected, and the frequency distribution of each feature value can be calculated and expressed as a feature histogram. By observing the feature histogram, the feature value with the most obvious feature representation can be selected as the key feature value. Here, according to the commonly used types of steel bars, the shape, maximum width, and maximum length of the end face can be used as feature values. The type of each steel bar can be determined by training and substituting it into the type recognition model.

[0095] Based on the above, the method for judging the regional error of the steel bar placement is:

[0096] Compare and determine in sequence whether the type of steel bars corresponding to each of the final end face partitions is consistent with the type identification of the steel bar storage area;

[0097] If they are inconsistent, the final end face partition is marked as an area placement error. If they are consistent, no operation is performed.

[0098] Here, the types of steel bars that should be stored in the steel bar stacking area can be marked with colored wireframes. After collecting the end face data, on the one hand, the color of the colored wireframe in the steel bar stacking area can be obtained, and the system can know what type of steel bars should be stored here. On the other hand, the end faces of all the steel bars can be obtained to determine the type of each steel bar and compare whether it is the type that should be stored here. If not, it will be marked in the image and an early warning will be issued to remind the staff to deal with it.

[0099] Example 3

[0100] This embodiment is based on the technical solution of implementation 1 and further illustrates the safety management of the steel bar warehouse.

[0101] As a preferred solution of this embodiment, a face recognition module is set at the door of the steel bar warehouse, the face data of the warehouse manager is entered into the face data storage library, and face verification is performed when someone enters the steel bar warehouse.

[0102] Simply put, the facial data of all warehouse managers are entered into the data repository. When a person is identified entering the steel warehouse, the individual is identified. If the facial data of this person is not found in the facial data being identified, an alarm is triggered to prompt other personnel to verify and handle the situation. This provides further security for our steel warehouse.

[0103] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A monitoring method for steel bar management at a construction site, characterized in that: The following steps are involved: The steel bar storage area in the steel bar warehouse is divided into zones using colored wireframes. Each steel bar storage area is stacked with steel bars of one type. Each steel bar storage area is provided with a type identification to mark the type of steel bars in the corresponding steel bar storage area. Multiple image acquisition devices are set to acquire images of each steel bar storage area, and the acquired viewing angles include a top view and a side view, wherein the side view acquires the end faces of the stacked steel bars; The placement, quantity and type of steel bars in each steel bar stacking area are identified through the top view and side view of the collected images; Determine whether there are any abnormal conditions. If not, do not perform any operation. If yes, mark and warn the steel bars with abnormal conditions. The abnormal conditions include the placement of the steel bars beyond the preset range, the wrong area for placement of the steel bars, and the number of a certain type of steel bars being lower than the corresponding threshold; The method for identifying the placement position, quantity and type of steel bars in each steel bar stacking area through the top view and side view of the collected video includes: Extracting an image as a check frame from the acquired image; Performing noise reduction processing on the extracted inspection frame; Perform corrosion and dilation on the inspection frame; In the inspection frame of the side view, the end face of each steel bar in the inspection frame is segmented by disassembling the connected domain; The inspection frame based on the side view angle determines the number and type of steel bars through image recognition; The inspection frame based on the top-down perspective determines whether the placement of the steel bars exceeds the preset range through image recognition; The method for judging whether the placement position of the steel bar exceeds the preset range is: A RGB value set B within a color threshold range is set based on the RGB value of the outer surface of the steel bar; Obtain the colored wireframe of each of the steel bar storage areas; Determine whether the colored wireframe is blocked; If it is not blocked, the placement position of the steel bar does not exceed the preset range; If it is blocked, extract the pixel cluster of the blocked part, check whether the RGB value of the pixel cluster belongs to B, if it does, the placement position of the steel bar exceeds the preset range, if not, the placement position of the steel bar does not exceed the preset range; If a pixel cluster with RGB value belonging to B is detected in any area outside all the steel bar storage areas, the placement position of the steel bars does not exceed the preset range; The method for segmenting the end face of each steel bar in the inspection frame by disassembling the connected domain is: Initialize the label value to 1, set a set of RGB values ​​A within the color threshold range based on the RGB value of the steel bar end surface, and traverse each pixel from the first row of the inspection frame from left to right. When the first pixel whose RGB value belongs to A is detected, the label value is its belonging number; Traverse the remaining pixels in the first row one by one. When the RGB value of the currently traversed pixel belongs to A, check whether the RGB value of the pixel on its left belongs to A. If so, the current pixel has the same belonging number as the pixel on its left. If not, update label = label + 1, and the label value is its belonging number. Starting from the second row, traverse each pixel point from left to right and then from top to bottom. When the RGB value of the currently traversed pixel point belongs to A, check whether at least one of the RGB values ​​of the pixel points above, to the upper left, and to the upper right belongs to A. If so, extract the belonging number of any pixel point whose RBG value belongs to A and assign it to the current pixel point. If not, update label = label + 1 and assign the label value to the current pixel point as its belonging number; After the traversal is completed, the pixels with the same belonging number are divided into a final end face partition; The method for identifying the number of the steel bars is: sequentially numbering each of the final end face partitions starting from 1 in order from left to right and then from top to bottom, and the total number of the final numbers is the number of the steel bars; The method for identifying the type of the steel bar is: Collect end face data of multiple steel bars of multiple types, obtain key feature values ​​through feature histogram analysis, and use the key feature values ​​and corresponding steel bar types to train a type recognition model; Obtain each of the final end surface partitions and extract key feature values ​​of its graphics; Substitute the key characteristic values ​​into the type identification model to perform type identification and obtain the steel bar type corresponding to each final end face partition; The method for judging the regional error of the steel bar placement is: Compare and determine in sequence whether the type of steel bars corresponding to each of the final end face partitions is consistent with the type identification of the steel bar storage area; If they are inconsistent, the final end face partition is marked as an area placement error. If they are consistent, no operation is performed.

2. A monitoring method for steel bar management at a construction site according to claim 1, characterized in that: The image acquisition device adopts a CCD image sensor.

3. A monitoring method for steel bar management at a construction site according to claim 1, characterized in that: The colored wireframes of different colors are used to delineate a plurality of regions, and the colored wireframes are used as the category identification.

4. A monitoring method for steel bar management at a construction site according to claim 1, characterized in that: A face recognition module is set at the door of the steel bar warehouse, and the face data of the warehouse manager is entered into the face data storage library, and face verification is performed when someone enters the steel bar warehouse.

Images