Supervision method and system based on construction project quality detection

Abstract

The invention discloses a supervision method and system based on construction project quality detection, and relates to the technical field of project quality detection. Comprises: obtaining an engineering project set; obtaining a real-time engineering project progress, re-planning detection planning time through a pre-estimation calculation method, and obtaining detection pre-estimation time; according to the detection estimation time, uploading an original photo for detection; detecting the original picture according to a picture focusing distribution detection method to obtain a detection result; and if the detection is qualified, carrying out the next detection until the engineering project is finished. According to the invention, the original pictures are detected through the picture focusing distribution detection method, the detection result is obtained, data recording is carried out, and the focusing constant is calculated according to the actual situation of each original picture, so that the feature value of the position with high importance degree in the original feature map is enhanced, and focusing detection of different degrees is realized. Therefore, the detection accuracy is improved, and the detection capability of small-scale defects can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of engineering quality detection, and in particular to a supervision method and system based on construction engineering quality detection. Background Art

[0002] The project quality inspection is divided into: material inspection, foundation inspection, main body inspection and decoration inspection.

[0003] Concrete walls are an important part of building structures. In the main inspection, concrete wall inspection occupies most of the inspection items. Its quality and performance are directly related to the safety and service life of the building. Due to problems such as aging and damage of concrete, concrete walls may face various safety hazards. Therefore, conducting concrete wall inspections in engineering quality inspections and timely discovering and dealing with wall problems are necessary measures to ensure building safety and extend service life.

[0004] In the process of building construction, it is necessary to use fragmentation settings, such as the image data-based rapid detection method for concrete cracks with patent publication number CN116958182A, which can obtain edge segments in the grayscale image of concrete walls; divide the edge segments based on the position distribution of pixel points on the edge segments to obtain each sub-segment; perform wavelet decomposition on the sub-segments to obtain the approximation coefficient, detail coefficient and wavelet spectrum of each sub-segment at different levels; determine the optimal decomposition level of each sub-segment according to the approximation coefficient and detail coefficient of each sub-segment at adjacent levels; obtain the crack edge confidence level corresponding to each sub-segment according to the relative distance between each sub-segment and the nearest sub-segment, the difference in wavelet spectrum, the difference in sensitivity and the corresponding optimal decomposition level, and then obtain the crack area. The accuracy of concrete crack detection results is improved.

[0005] There are many types of defect detection for concrete walls, including honeycombs, rough surfaces, holes, cracks, etc. Various defects have different impacts on concrete walls. Concrete wall inspection during engineering quality inspection and timely discovery and treatment of wall problems are necessary measures to ensure building safety and extend service life. The above method can only analyze and process the image as a whole, but cannot specify the importance of different spatial positions in the image, cannot focus on the spatial area where the defects are located, and has poor effect on small-scale defect detection. Summary of the invention

[0006] The purpose of the present invention is to provide a supervision method and system based on construction project quality inspection to solve the problems raised in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solution: a supervision method based on construction project quality inspection, the method comprising:

[0008] Obtain a collection of engineering projects, which includes the estimated time of project progress, the inspection planning time, and the inspection data storage set;

[0009] Obtain real-time project progress, re-plan the inspection time through the estimated calculation method, and obtain the estimated inspection time;

[0010] According to the re-planned estimated inspection time, prompt inspection is carried out in advance and the original inspection photos are uploaded to the inspection data storage set;

[0011] The original photos are inspected according to the image focus allocation detection method, the detection results are obtained, and the data is recorded. The focus constant is calculated according to the actual situation of each original photo, so that the feature values ​​of the positions with high importance of the original feature map will be enhanced, while the feature values ​​of the positions with low importance will be suppressed, so as to achieve different degrees of focus detection, thereby improving the accuracy of detection and the detection ability of small-scale defects;

[0012] When the test fails, make corrections and record the correction process. Similar problems can be matched and referenced in the future. If the test fails after correction, the last correction process can be queried to facilitate the re-formulation of the correction plan. When the test passes, the above test steps can be cycled and the next test can be carried out until the project is completed.

[0013] The picture focus allocation detection method comprises:

[0014] S1: According to the number of pixels in the image and the number of channels of each pixel, the focusing constant of each pixel is obtained through the focusing constant calculation method to reflect the importance of the target pixel. All the focusing constants constitute a focusing characteristic map;

[0015] S2: Obtain a grayscale image based on the color image, and perform edge sharpening through a convolution algorithm to obtain an original feature map, thereby improving the clarity of the grayscale image and obtaining more accurate edge information;

[0016] S3: Convolution calculation is performed on the original feature map and the focus characteristic map to obtain a feature magnification map, and the number of pixels in the feature magnification map whose feature values ​​reach the threshold is calculated. When the number of pixels exceeds the set standard value, the detection is judged to be unqualified, and these pixels are marked on the original image. The marked image is output to facilitate the determination of the defect location.

[0017] Preferably, the focusing constant calculation method includes:

[0018] S1: Compare the values ​​of several channels of the same pixel and record the channel item with the largest value;

[0019] S2: Get the number of pixels with the same maximum channel item and the total number of pixels of the original image, and then calculate according to the formula:

[0020]

[0021]

[0022] in is the weight coefficient of channel e, is the number of pixels in the original image with channel e as the maximum channel item, is the total number of pixels of the original image, is the focusing constant at the coordinate (i, j) in the focusing characteristic diagram, representing the focusing degree of the position at the coordinate (i, j) in the original image. is the element value at coordinate (i, j) in the original image and on channel e, is the number of channels for a single pixel in the original image.

[0023] Preferably, the focusing constant calculation method includes:

[0024] L1: Sequentially number the multiple original images obtained, obtain the grayscale image of each image, calculate the average grayscale value, and then calculate according to the formula:

[0025]

[0026] in; is the quality coefficient of the original image, numbered according to the order of the original images, The average gray value of the first original image, is the average gray value of the pth original image, To obtain the total number of original images, the original image quality coefficient of the first original image is calculated by comparing multiple original images to reduce the impact of external factors such as the shooting angle and overall brightness of a single image on the detection.

[0027] L2: According to the sequence number of the original images, compare the values ​​of several channels of each pixel in the first image, and record the channel item with the largest value;

[0028] L3: Get the number of pixels with the same maximum channel item and the total number of pixels of the original image, and then calculate according to the formula:

[0029]

[0030]

[0031] in is the original image quality coefficient, is the weight coefficient of channel e, is the number of pixels in the original image with channel e as the maximum channel item, is the total number of pixels of the original image, is the focusing constant at the coordinate (i, j) in the focusing characteristic diagram, representing the focusing degree of the position at the coordinate (i, j) in the original image. is the element value at coordinate (i, j) in the original image and on channel e, It is the number of channels of a single pixel in the original image. By comparing multiple original images, the influence of a single original photo on the calculation results is reduced, and the reliability of the detection results is further improved.

[0032] Preferably, the estimation calculation method includes:

[0033] G1: Based on the multiple real-time project progresses obtained, record the last real-time project progress obtained and compare it with the estimated time of the project progress;

[0034] G2: Calculate the estimated detection time according to the formula:

[0035]

[0036] in Estimated time for the current test item. Plan time for the current test project. The estimated time of the project progress corresponding to the last acquired real-time project progress in the engineering project collection. The time when the real-time project progress was last obtained. is the time fluctuation coefficient of the rth process in the engineering project set, indicating the degree of time fluctuation that may occur in this process in actual production. It is set by the technical staff. is the total number of processes before the current inspection item, In order to obtain the number of processes before the last real-time project progress, different time fluctuation coefficients are assigned to each project in the engineering project set, so as to determine the sensitivity of each project to the completion time and more accurately calculate the estimated inspection time of the current inspection project.

[0037] Preferably, the estimation calculation method includes:

[0038] Record the real-time project progress obtained multiple times, and calculate the estimated detection time through the formula:

[0039]

[0040]

[0041] in is the number of times the implementation project progress is obtained, and ≥2, To calculate from the back to the front, The time to obtain the real-time project progress. To calculate from the back to the front, The estimated time of the project progress corresponding to the real-time project progress obtained in the project collection, To calculate from the back to the front, The impact operator of the real-time project progress obtained, Estimated time for the current test item. Plan the time for the inspection of the current inspection project. According to the number of times the real-time project progress is obtained, the later the real-time project progress is, the greater the impact on the estimated inspection time of the current inspection project. According to the size of the influence, fully consider its influence on the estimated inspection time of the current inspection project, so that the estimated inspection time of the current inspection project is more accurate.

[0042] Preferably, there are several thresholds, each threshold is set with a corresponding standard value, when the number of pixels reaching the threshold reaches the standard value of the target threshold, the detection is judged to be unqualified, and the positions where the characteristic values ​​in the feature magnification image exceed different thresholds are marked with different colors at the corresponding positions on the original image, so that different defect forms can be marked differently, which is convenient for inspection personnel to make rectifications.

[0043] Preferably, the convolution operation algorithm includes:

[0044] B1: The grayscale image is convolved using the Sobel operator, and a greater weight is given to the central pixel when calculating the gradient, which can better highlight the edge information and has relatively good anti-noise performance;

[0045] B2: Detect the grayscale value on the grayscale image and compare the grayscale value of each pixel with the four grayscale values ​​of adjacent pixels. When the ratio of the grayscale value of a pixel to the grayscale value of its adjacent pixels is above a set multiple, the grayscale value of the target pixel is set to the average of the grayscale values ​​of the four adjacent pixels, which can effectively remove noise.

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

[0047] The original photos are inspected by the image focus allocation detection method, the detection results are obtained, and the data is recorded. The focus constant is calculated according to the actual situation of each original photo so that the feature values ​​of the original feature map at the positions with high importance in the focus characteristic map will be enhanced, while the feature values ​​of the positions with low importance will be suppressed, so as to achieve different degrees of focus detection, thereby improving the detection accuracy and the detection capability of small-scale defects;

[0048] When the test fails, make corrections and record the correction process. Similar problems can be matched and referenced in the future. If the test fails after correction, the last correction process can be queried to facilitate the re-formulation of the correction plan. When the test passes, the above test steps can be cycled and the next test can be carried out until the project is completed.

[0049] At the same time, when calculating the focusing constant, the influence of a single original photo on the calculation result is reduced by comparing multiple original pictures, and the detection subject part on the original picture can be further strengthened to prevent the non-detection part in the original picture from affecting the detection result, thereby further improving the reliability of the detection result;

[0050] In addition, the subsequent inspection planning time can be re-planned according to the real-time project progress obtained during the construction process, and reminders can be given before the inspection planning time, which can ensure the accuracy of the reminder time and prevent the reminder from being too late, resulting in missed inspections, etc. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a flow chart of the construction project quality inspection and supervision method of the present invention;

[0052] Figure 2 A schematic diagram of the flow of the image focus allocation detection method of the present invention;

[0053] Figure 3 Schematic diagram of the flow of the focusing constant calculation method (single original image) of the present invention;

[0054] Figure 4 It is a schematic flow chart of the focusing constant calculation method (multiple original images) of the present invention;

[0055] Figure 5 It is a flow chart of the estimation calculation method (setting time fluctuation coefficient) of the present invention;

[0056] Figure 6 It is a flow chart of the estimation calculation method (without setting the time fluctuation coefficient) of the present invention. DETAILED DESCRIPTION

[0057] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments 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 creative work are within the scope of protection of the present invention.

[0058] In the present application, for ease of understanding, the method steps used do not need to be executed in the order of the steps in this embodiment during actual operation. In other embodiments, these steps may be performed simultaneously or in a different order.

[0059] like Figure 1-Figure 6 As shown, the present invention provides a technical solution: a supervision method based on construction project quality inspection, comprising:

[0060] Obtain a collection of engineering projects, which includes the estimated time of project progress, the inspection planning time, and the inspection data storage set;

[0061] Obtain real-time project progress, re-plan the inspection time through the estimated calculation method, and obtain the estimated inspection time;

[0062] According to the re-planned estimated inspection time, prompt inspection is carried out in advance and the original inspection photos are uploaded to the inspection data storage set;

[0063] The original photos are inspected according to the image focus allocation detection method, the detection results are obtained, and the data is recorded. The focus constant is calculated according to the actual situation of each original photo, so that the feature values ​​of the positions with high importance in the original feature map will be enhanced, while the feature values ​​of the positions with low importance will be suppressed, so as to achieve different degrees of focus detection, thereby improving the accuracy of detection and the detection ability of small-scale defects;

[0064] When the test fails, rectification is carried out and the rectification process is recorded. Similar problems that arise in the future can be matched and referenced. If the test still fails after rectification, the previous rectification process can be queried to facilitate the reformulation of the rectification plan. When the test passes, the above inspection steps can be looped and the next inspection can be carried out until the project is completed.

[0065] It should be noted that the estimated project progress and inspection planning time can be obtained from the construction plan, and the original photos can be taken by the operator using a mobile phone, camera or drone. However, the original pictures obtained from multiple inspection projects in the entire construction project need to be processed in advance to ensure that the pixel value of the original picture is a specific value and maintains a specific scale with the actual size of the project building to ensure the accuracy of the inspection results.

[0066] like Figure 2 As shown, the image focus allocation detection method includes:

[0067] S1: According to the number of pixels in the image and the number of channels of each pixel, the focusing constant of each pixel is obtained through the focusing constant calculation method to reflect the importance of the target pixel. All the focusing constants constitute a focusing characteristic map;

[0068] S2: Obtain a grayscale image based on the color image, and perform edge sharpening through a convolution algorithm to obtain an original feature map, thereby improving the clarity of the grayscale image and obtaining more accurate edge information;

[0069] S3: Multiply the original feature map and the focus characteristic map element by element to obtain the feature magnification map, calculate the number of pixels in the feature magnification map whose feature values ​​reach the threshold, and when the number of pixels exceeds the set standard value, the detection is judged to be unqualified, and these pixels are marked on the original image, and the marked image is output to facilitate the determination of the defect location.

[0070] It should be noted that the focusing constant is calculated according to the actual situation of each original photo so that the eigenvalues ​​of the original feature map at the positions with high importance in the focusing characteristic map will be enhanced, while the eigenvalues ​​of the positions with low importance will be suppressed, thereby achieving different degrees of focusing detection, thereby improving the accuracy of detection and the ability to detect small-scale defects.

[0071] Moreover, several thresholds can be set, and each threshold has a corresponding standard value. When the number of pixels reaching the threshold reaches the standard value of the target threshold, the detection is judged to be unqualified. The positions where the feature values ​​in the feature magnification image exceed different thresholds are marked with different colors at the corresponding positions on the original image. Different defect forms can be marked differently, which is convenient for detection personnel to make rectifications.

[0072] During detection, multiple thresholds can be set. The larger the threshold, the more serious the defect. They can correspond to different defects respectively. When there are many locations with serious defects, they are directly judged as unqualified, which can further improve the detection accuracy.

[0073] It should be noted that in order to facilitate understanding and simplify calculations, the simulation data is set as follows:

[0074] The set thresholds and their corresponding standard values ​​are shown in Table 1:

[0075] Table 1

[0076] Threshold 260 300 330 350 380 500 Standard value (ten thousand) 360 340 320 300 225 30

[0077] In the set feature enlargement image, the number of pixels with an eigenvalue of 260 is 3.52 million, the number of pixels with an eigenvalue of 300 is 3.3 million, the number of pixels with an eigenvalue of 330 is 3.1 million, the number of pixels with an eigenvalue of 350 is 2.98 million, the number of pixels with an eigenvalue of 350 is 2.1 million, and the number of pixels with an eigenvalue of 500 is 320,000.

[0078] According to the above data, the number of pixels with a characteristic value of 500 is 320,000, which is greater than the standard value of 300,000. This means that the area of ​​pixels with serious defects exceeds the specified range. Although the area of ​​defective pixels is not large, the area of ​​items with serious defects is large, and they are judged as unqualified.

[0079] Convolution algorithms include:

[0080] B1: The grayscale image is convolved using the Sobel operator, and a greater weight is given to the central pixel when calculating the gradient, which can better highlight the edge information and has relatively good anti-noise performance;

[0081] B2: Detect the grayscale value on the grayscale image and compare the grayscale value of each pixel with the four grayscale values ​​of adjacent pixels. When the ratio of the grayscale value of a pixel to the grayscale value of its adjacent pixels is above a set multiple, the grayscale value of the target pixel is set to the average of the grayscale values ​​of the four adjacent pixels, which can effectively remove noise.

[0082] It should be noted that the Sobel operator is simple to calculate, fast, and has a certain smoothing effect on noise. In practical applications, it can achieve better edge detection effects on images with grayscale gradients and more noise. It is an existing technology to perform convolution operations on images using the Sobel operator, which will not be described in detail here. After the convolution operation, the grayscale values ​​are compared, and pixels that are significantly higher than the surrounding grayscale values ​​or lower than the surrounding grayscale values ​​are modified, which can further reduce the detection error caused by the image shooting quality.

[0083] Embodiment 1:

[0084] like Figure 3 As shown, the focusing constant calculation method includes:

[0085] S1: Compare the values ​​of several channels of the same pixel and record the channel item with the largest value. The more channels there are, the more accurate the calculation result will be.

[0086] S2: Get the number of pixels with the same maximum channel item and the total number of pixels of the original image, and then calculate according to the formula:

[0087]

[0088]

[0089] in is the weight coefficient of channel e, is the number of pixels in the original image with channel e as the maximum channel item, is the total number of pixels of the original image, is the focusing constant at the coordinate (i, j) in the focusing characteristic diagram, representing the focusing degree of the position at the coordinate (i, j) in the original image. is the element value at coordinate (i, j) in the original image and on channel e, is the number of channels for a single pixel in the original image.

[0090] It should be noted that for ease of understanding, the simulation data is set as follows:

[0091] Set to 3. The pixel channels of ordinary mobile phone pictures are 3, namely red, green and blue. In this embodiment, they are marked as channel 0, channel 1 and channel 2 respectively, and the value of each channel is usually between 0 and 225.

[0092] Set to 25 million; In this embodiment, they are , , , and their values ​​are 16 million, 4.5 million, and 4.5 million respectively. In this embodiment, they are , , , according to the above data and formula , , , The calculated values ​​are 0.64, 0.18, and 0.18;

[0093] Take any pixel on the original image, its position is (i, j), the value of its channel 0, channel 1, channel 2 , , They are 200, 50, and 50 respectively.

[0094] Substitute into the formula , we can calculate The value of is 308, and the calculation method for other positions on the original image is the same. The value reflects the importance of the position in the original image. The weight coefficient of channel e can further strengthen the importance of the detection subject in the original image, reduce the impact of the useless detection parts on the image on the detection results, and further improve the reliability of the detection results.

[0095] Embodiment 2:

[0096] like Figure 4 As shown, when the number of acquired original images is large, the focusing constant calculation method includes:

[0097] L1: Sequentially number the multiple original images obtained, obtain the grayscale image of each image, calculate the average grayscale value, and then calculate according to the formula:

[0098]

[0099] in; is the quality coefficient of the original image, numbered according to the order of the original images, The average gray value of the first original image, is the average gray value of the pth original image, To obtain the total number of original images, the original image quality coefficient of the first original image is calculated by comparing multiple original images to reduce the impact of external factors such as the shooting angle and overall brightness of a single image on the detection.

[0100] L2: According to the sequence number of the original images, compare the values ​​of several channels of each pixel in the first image, and record the channel item with the largest value;

[0101] L3: Get the number of pixels with the same maximum channel item and the total number of pixels of the original image, and then calculate according to the formula:

[0102]

[0103]

[0104] in is the original image quality coefficient, is the weight coefficient of channel e, is the number of pixels in the original image with channel e as the maximum channel item, is the total number of pixels of the original image, is the focusing constant at the coordinate (i, j) in the focusing characteristic diagram, representing the focusing degree of the position at the coordinate (i, j) in the original image. is the element value at coordinate (i, j) in the original image and on channel e, It is the number of channels of a single pixel in the original image. By comparing multiple original images, the influence of a single original photo on the calculation results is reduced, and the reliability of the detection results is further improved.

[0105] It should be noted that for ease of understanding, the simulation data is set as follows:

[0106] Set to 5, which means a total of 5 original images are obtained;

[0107] In this embodiment, they are , , , , , whose values ​​are set to 200, 210, 220, 230, and 190 respectively. According to the formula , can be obtained The value of is 1.05, which represents the grayscale value correction multiple of the first original image, reducing the impact of external factors such as the shooting angle of the first original image and the ambient brightness (for example, the external brightness at noon and in the evening is different) on the detection.

[0108] Set to 3. The pixel channels of ordinary mobile phone pictures are 3, namely red, green and blue. In this embodiment, they are marked as channel 0, channel 1 and channel 2 respectively, and the value of each channel is usually between 0 and 225.

[0109] Set to 25 million; In this embodiment, they are , , , and their values ​​are 16 million, 4.5 million, and 4.5 million respectively. In this embodiment, they are , , , according to the above data and formula , , , The calculated values ​​are 0.64, 0.18, and 0.18;

[0110] Take any pixel on the original image, its position is (i, j), the value of its channel 0, channel 1, channel 2 , , They are 200, 50, and 50 respectively.

[0111] Substitute into the formula , we can calculate The value of is 323.4, and the calculation method for other positions on the original image is the same. The value reflects the importance of the position in the original image, and is adjusted by the original image quality coefficient. The correction can further reduce the impact of a single original photo on the calculation results and improve the reliability of the detection results.

[0112] Embodiment three:

[0113] like Figure 5 As shown, the estimated calculation method includes:

[0114] G1: Based on the multiple real-time project progresses obtained, record the last real-time project progress obtained and compare it with the estimated time of the project progress;

[0115] G2: Calculate the estimated detection time according to the formula:

[0116]

[0117] in Estimated time for the current test item. Plan time for the current test project. The estimated time of the project progress corresponding to the last acquired real-time project progress in the engineering project collection. The time when the real-time project progress was last obtained. is the time fluctuation coefficient of the rth process in the engineering project set, indicating the degree of time fluctuation that may occur in this process in actual production. It is set by the technical staff. is the total number of processes before the current inspection item, In order to obtain the number of processes before the last real-time project progress, different time fluctuation coefficients are assigned to each project in the engineering project set, so as to determine the sensitivity of each project to the completion time and more accurately calculate the estimated inspection time of the current inspection project.

[0118] It should be noted that for ease of understanding, the simulation data is set as follows:

[0119] Set it to 120 days after the project starts. Set it to the 50th day after the project starts. Set it to the 40th day after the project starts. Set to 5, Set to 2, In this embodiment, the corresponding , , , , They are 2, 3, 1, 4, and 2 respectively. The values ​​are set by the staff. The above values ​​represent The difference between the actual construction time and the planned construction time is the largest, and the construction progress is most easily disturbed by external factors. The error between the actual construction duration and the planned construction duration is minimal, and the construction progress is not easily disturbed by external factors.

[0120] According to the above data, put it into the formula , we can calculate It is the 106th day after the start of the project, so it is necessary to remind the inspection project 106 days after the start of the project, and assign different time fluctuation coefficients to each project in the project set, so as to judge the sensitivity of each project to the completion time and more accurately calculate the estimated inspection time of the current inspection project.

[0121] Embodiment 4:

[0122] like Figure 6 As shown, the estimation calculation method includes:

[0123] Record the real-time project progress obtained multiple times, and calculate the estimated detection time through the formula:

[0124]

[0125]

[0126] in is the number of times the implementation project progress is obtained, and ≥2, To calculate from the back to the front, The time to obtain the real-time project progress. To calculate from the back to the front, The estimated time of the project progress corresponding to the real-time project progress obtained in the project collection, To calculate from the back to the front, The impact operator of the real-time project progress obtained, Estimated time for the current test item. Plan the time for the inspection of the current inspection project. According to the number of times the real-time project progress is obtained, the later the real-time project progress is, the greater the impact on the estimated inspection time of the current inspection project. According to the size of the influence, fully consider its influence on the estimated inspection time of the current inspection project, so that the estimated inspection time of the current inspection project is more accurate.

[0127] It should be noted that this calculation method does not require staff to set the time fluctuation coefficient of the process in advance, and the operation is simpler.

[0128] For ease of understanding, the simulation data is set as follows:

[0129] Set to 3, In this embodiment, the corresponding , , They are the 60th, 40th and 20th days after the start of the project. In this embodiment, the corresponding , , The 70th, 60th and 30th days after the start of the project respectively. Set to 120 days after the project starts.

[0130] Substitute into the formula , we can calculate , , They are -9, -0.18, and -0.1 respectively, which are rounded to integers to -9, -2, and 0. According to the formula , we can calculate It is the 109th day after the start of the project, so the inspection project needs to be reminded before the 109th day after the start of construction. Although there was a process of completing the indicators 20 days ahead of schedule, the reference value of obtaining the real-time project progress for the last time is higher, so the estimated inspection time is set to 11 days before the inspection planning time.

[0131] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is limited by the attached embodiments and their equivalents.

Claims

1. A supervision method based on construction project quality inspection, characterized by: The method comprises: Obtain a collection of engineering projects, which includes the estimated time of project progress, the inspection planning time, and the inspection data storage set; Obtain real-time project progress, re-plan the inspection time through the estimated calculation method, and obtain the estimated inspection time; According to the re-planned estimated inspection time, prompt inspection is carried out in advance and the original inspection photos are uploaded to the inspection data storage set; The original photos are inspected according to the image focus allocation detection method, the detection results are obtained, and the data is recorded. The focus constant is calculated according to the actual situation of each original photo, so that the feature values ​​of the positions with high importance in the original feature map will be enhanced, while the feature values ​​of the positions with low importance will be suppressed, so as to achieve different degrees of focus detection, thereby improving the accuracy of detection and the detection ability of small-scale defects; When the test fails, make corrections and record the correction process. Similar problems can be matched and referenced in the future. If the test fails after correction, the last correction process can be queried to facilitate the re-formulation of the correction plan. When the test passes, the above test steps can be cycled and the next test can be carried out until the project is completed. The picture focus allocation detection method comprises: S1: According to the number of pixels in the image and the number of channels of each pixel, the focusing constant of each pixel is obtained through the focusing constant calculation method to reflect the importance of the target pixel. All the focusing constants constitute a focusing characteristic map; S2: Obtain a grayscale image based on the color image, and perform edge sharpening through a convolution algorithm to obtain an original feature map, thereby improving the clarity of the grayscale image and obtaining more accurate edge information; S3: Multiply the original feature map and the focus characteristic map element by element to obtain the feature magnification map, calculate the number of pixels in the feature magnification map whose feature values ​​reach the threshold, and when the number of pixels exceeds the set standard value, the detection is judged to be unqualified, and these pixels are marked on the original image, and the marked image is output to facilitate the determination of the defect location.

2. A supervision method based on construction project quality detection according to claim 1, characterized in that: The focusing constant calculation method comprises: S1: Compare the values ​​of several channels of the same pixel and record the channel item with the largest value; S2: Get the number of pixels with the same maximum channel item and the total number of pixels of the original image, and then calculate according to the formula: ； ； in is the weight coefficient of channel e, is the number of pixels in the original image with channel e as the maximum channel item, is the total number of pixels of the original image, is the focusing constant at the coordinate (i, j) in the focusing characteristic diagram, representing the focusing degree of the position at the coordinate (i, j) in the original image. is the element value at coordinate (i, j) in the original image and on channel e, is the number of channels for a single pixel in the original image.

3. A supervision method based on construction project quality inspection according to claim 1, characterized in that: The focusing constant calculation method comprises: L1: Sequentially number the multiple original images obtained, obtain the grayscale image of each image, calculate the average grayscale value, and then calculate according to the formula: ； in is the quality coefficient of the original image, numbered according to the order of the original images, The average gray value of the first original image, is the average gray value of the pth original image, For the total number of original images obtained, the original image quality coefficient of the first original image is calculated by comparing multiple original images to reduce the impact of external factors such as the shooting angle and overall brightness of a single image on the detection; L2: According to the sequence number of the original images, compare the values ​​of several channels of each pixel in the first image, and record the channel item with the largest value; L3: Get the number of pixels with the same maximum channel item and the total number of pixels of the original image, and then calculate according to the formula: ； ； in is the original image quality coefficient, is the weight coefficient of channel e, is the number of pixels in the original image with channel e as the maximum channel item, is the total number of pixels of the original image, is the focusing constant at the coordinate (i, j) in the focusing characteristic diagram, representing the focusing degree of the position at the coordinate (i, j) in the original image. is the element value at coordinate (i, j) in the original image and on channel e, It is the number of channels of a single pixel in the original image. By comparing multiple original images, the influence of a single original photo on the calculation results is reduced, and the reliability of the detection results is further improved.

4. A supervision method based on construction project quality inspection according to claim 1, characterized in that: The estimation calculation method comprises: G1: Based on the multiple real-time project progresses obtained, record the last real-time project progress obtained and compare it with the estimated time of the project progress; G2: Calculate the estimated detection time according to the formula: ； in Estimated time for the current test item. Plan time for the current test project. The estimated time of the project progress corresponding to the last acquired real-time project progress in the engineering project collection. The time when the real-time project progress was last obtained. is the time fluctuation coefficient of the rth process in the engineering project set, indicating the degree of time fluctuation that may occur in this process in actual production. It is set by the technical staff. is the total number of processes before the current inspection item, In order to obtain the number of processes before the last real-time project progress, different time fluctuation coefficients are assigned to each project in the engineering project set, so as to determine the sensitivity of each project to the completion time and more accurately calculate the estimated inspection time of the current inspection project.

5. The supervision method based on construction project quality detection according to claim 1 is characterized by: The estimation calculation method comprises: Record the real-time project progress obtained multiple times, and calculate the estimated detection time through the formula: ； ； in is the number of times the implementation project progress is obtained, and ≥2, To calculate from the back to the front, The time to obtain the real-time project progress. To calculate from the back to the front, The estimated time of the project progress corresponding to the real-time project progress obtained in the project collection, To calculate from the back to the front, The impact operator of the real-time project progress obtained, Estimated time for the current test item. Plan the time for the inspection of the current inspection project. According to the number of times the real-time project progress is obtained, the later the real-time project progress is, the greater the impact on the estimated inspection time of the current inspection project. According to the size of the influence, fully consider its influence on the estimated inspection time of the current inspection project, so that the estimated inspection time of the current inspection project is more accurate.

6. A supervision method based on construction project quality inspection according to claim 1, characterized in that: There are several thresholds, each of which has a corresponding standard value. When the number of pixels reaching the threshold reaches the standard value of the target threshold, the detection is judged to be unqualified. The positions where the feature values ​​in the feature magnification image exceed different thresholds are marked with different colors at the corresponding positions on the original image. Different defect forms can be marked differently, which is convenient for detection personnel to make rectifications.

7. A supervision method based on construction project quality inspection according to claim 1, characterized in that: The convolution operation algorithm includes: B1: The grayscale image is convolved using the Sobel operator, and a greater weight is given to the central pixel when calculating the gradient, which can better highlight the edge information and has relatively good anti-noise performance; B2: Detect the grayscale value on the grayscale image and compare the grayscale value of each pixel with the four grayscale values ​​of adjacent pixels. When the ratio of the grayscale value of a pixel to the grayscale value of its adjacent pixels is above a set multiple, the grayscale value of the target pixel is set to the average of the grayscale values ​​of the four adjacent pixels, which can effectively remove noise.

8. A supervision system based on construction project quality inspection, characterized by: A supervision method based on construction project quality inspection as described in any one of claims 1 to 7 is used.

Images