Concrete quality defect detection method, computer equipment and storage medium

Abstract

The present invention provides a method, computer device, and storage medium for detecting concrete quality defects. The detection method comprises: obtaining each frame of the original image of the inspection video and its position and posture information; the position and posture information includes camera coordinate data and camera angle and posture data; performing defect target detection on each frame of the original image, retaining the original image with quality defects; pre-classifying each quality defect in the original image with quality defects to obtain a target defect set; identifying all quality defects in the target defect set to obtain multiple target defects of the target defect type; determining and marking the coordinate information of each target defect based on the position and posture information of the original image corresponding to the target defect; aggregating the coordinate information of all target defects and performing structured processing to output structured data. The method solves the technical problem in the prior art of inadequate inspection of concrete quality defects at the construction site, leading to untimely rectification of on-site defects.

Description

Technical Field

[0001] The present invention relates to the technical field of concrete quality detection, and in particular to a concrete quality defect detection method, computer equipment and storage medium. Background Art

[0002] During construction of the main structure at a construction site, after the cast-in-place concrete structure is completed and the formwork support system is removed, but before the concrete surface is corrected and decorated, quality defects such as exposed reinforcement, honeycombs, voids, slag inclusions, and cracks often occur in the concrete structure. If these quality issues are not corrected, they not only affect the usability and visual function of the structure, but also seriously impact the normal construction of subsequent decoration and renovation, leading to a series of risks such as rework, increased costs, and delayed construction schedules. Construction site quality management is generally controlled through on-site inspections and acceptance by quality management personnel. Key control items and general items are fully inspected and technical solutions are proposed. Due to the complex construction site environment, it is difficult for on-site management personnel to conduct comprehensive inspections, which increases the risk of project quality management. Summary of the Invention

[0003] In response to the deficiencies in the prior art, the present invention provides a concrete quality defect detection method, computer equipment and storage medium, which solve the technical problem in the prior art of inadequate inspection of concrete quality defects at construction sites leading to untimely rectification of on-site defects.

[0004] In a first aspect, the present invention provides a method for detecting concrete quality defects, the method comprising:

[0005] Obtain each frame of the original image of the inspection video and its posture information; the posture information includes camera coordinate data and camera angle posture data;

[0006] Perform defect target detection on each frame of original image and retain the original images with quality defects;

[0007] Pre-classify each quality defect in the original image with quality defects to obtain a target defect set;

[0008] Identify all quality defects in the target defect set to obtain multiple target defects of the target defect type;

[0009] Based on the pose information of the target defect corresponding to the original image, the coordinate information of each target defect is determined and marked;

[0010] Collect the coordinate information of all target defects, perform structured processing, and output structured data.

[0011] Furthermore, defect target detection is performed on each frame of the original image, including:

[0012] Each frame of the original image is input into the target detection model, which marks and identifies each quality defect in the original image and marks the defect box.

[0013] Furthermore, each quality defect in the original image with quality defects is pre-classified to obtain a target defect set, including:

[0014] Based on the pose information corresponding to the original image where the quality defect is located and the position of the quality defect in the original image, determine the depression angle of the center of the quality defect relative to the camera;

[0015] Obtain all quality defects with a depression angle greater than a set degree to generate a target defect set.

[0016] Furthermore, the target defect types include exposed rib defects, honeycomb defects, void defects, slag inclusion defects, and crack defects; all quality defects in the target defect set are identified to obtain multiple target defects of the target defect type, including:

[0017] Identify each quality defect in the target defect set through a preset target recognition model to obtain the target defect;

[0018] Each target defect is marked with the corresponding target defect type.

[0019] Furthermore, based on the pose information of the target defect corresponding to the original image, the coordinate information of each target defect is determined and marked, including:

[0020] Construct a spatial rectangular coordinate system for the construction site;

[0021] Based on the pose information of the target defect corresponding to the original image and the position of the center of the target defect in the original image, the defect equation of the straight line where the center of the target defect is located is determined;

[0022] According to the defect equation and the surface equation of the inspection floor, the coordinates of the center of the quality defect are confirmed.

[0023] Furthermore, based on the pose information of the target defect corresponding to the original image and the position of the center of the target defect in the original image, the defect equation of the straight line where the center of the target defect is located is determined, including:

[0024] Based on the camera coordinate data, confirm the camera position equation (x0, y0, z0);

[0025] Based on the camera angle pose data and the position of the center of the target defect in the original image, confirm the direction vector (a, b, c) of the quality defect center relative to the camera;

[0026] Based on the camera position equation (x0, y0, z0) and direction vector (a, b, c), the defect equation is confirmed to be:

[0027]

[0028] Furthermore, based on the defect equation and the surface equation of the inspection floor, the coordinates of the quality defect center are determined, including:

[0029] Get the equation of the surface where the inspection floor is located: Ax+By+Cz+D=0;

[0030] Combine the defect equations to obtain the coordinates of the mass defect center:

[0031]

[0032] Furthermore, the detection method also includes: before performing defect target detection on each frame of original image, each frame of original image needs to be preprocessed; the preprocessing includes image compression, noise filtering, edge extraction and image sharpening.

[0033] In a second aspect, the present invention provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements a detection method when executing the computer program: obtaining each frame of the original image of the inspection video and its posture information; the posture information comprises camera coordinate data and camera angle posture data; performing defect target detection on each frame of the original image, and retaining the original image with quality defects; pre-classifying each quality defect in the original image with quality defects to obtain a target defect set; identifying all quality defects in the target defect set to obtain multiple target defects of the target defect type; determining and marking the coordinate information of each target defect based on the posture information of the original image corresponding to the target defect; collecting the coordinate information of all target defects and performing structured processing to output structured data.

[0034] In a third aspect, the present invention provides a readable storage medium having a computer program stored thereon, which implements a detection method when executed by a processor: obtaining each frame of the original image of the inspection video and its posture information; the posture information includes camera coordinate data and camera angle posture data; performing defect target detection on each frame of the original image, and retaining the original image with quality defects; pre-classifying each quality defect in the original image with quality defects to obtain a target defect set; identifying all quality defects in the target defect set to obtain multiple target defects of the target defect type; determining and marking the coordinate information of each target defect based on the posture information of the original image corresponding to the target defect; collecting the coordinate information of all target defects and performing structured processing to output structured data.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] The present invention obtains original images with quality defects by performing defect target detection on each frame of the original image of the inspection video, so as to delete the original images without quality defects in the inspection video; pre-classifies each defect in the original image with quality defects to obtain a target defect set, so as to avoid the need for a large amount of algorithm training and calculation in the subsequent defect identification process, making the algorithm lightweight. By identifying all quality defects in the target defect set, multiple target defects of the target defect type and the coordinate information corresponding to each target defect are obtained and structured processing is performed, and structured data is output to manage defects of the target defect type. The technical problem existing in the prior art of inadequate inspection of concrete quality defects at the construction site leading to untimely rectification of on-site defects is solved. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A diagram of method steps according to an embodiment of the present invention. DETAILED DESCRIPTION

[0038] The technical solution of the present invention is further described below with reference to the accompanying drawings and embodiments.

[0039] Example 1:

[0040] like Figure 1 As shown, the present invention provides a method for detecting concrete quality defects, the detection method comprising:

[0041] S101: Acquire each frame of the original image of the inspection video and its posture information; the posture information includes camera coordinate data and camera angle posture data;

[0042] Inspection videos are captured using a 360° panoramic camera mounted on a person's body. Each time a raw image is captured, the camera acquires the camera's coordinate data and angle and attitude data for the current state, which are then bound to the original image. In this embodiment, a spatial coordinate system is constructed based on the building. When the panoramic camera enters a building and is powered on, the camera's built-in IMU (Infrared Measurement Unit) is used to acquire the camera's coordinate data and angle and attitude data.

[0043] After the inspection personnel or inspection robot enters the construction site with a panoramic camera, they turn on the panoramic camera to shoot the construction site. When the inspection is completed, the panoramic camera is turned off to obtain the inspection video.

[0044] S102: performing defect target detection on each frame of the original image, and retaining the original image with quality defects;

[0045] The deep learning target detection model is used to detect whether there are quality defects in each frame of the original image. If so, the original image is retained; otherwise, the original image is deleted.

[0046] S103: Pre-classifying each quality defect in the original image with quality defects to obtain a target defect set;

[0047] In order to obtain the defects at the target position and make the subsequent defect recognition algorithm more lightweight, each defect in the original image is pre-classified to obtain the target defect set.

[0048] S104: Identify all quality defects in the target defect set to obtain multiple target defects of the target defect type;

[0049] The target defects in the target defect set are identified using a target recognition model to obtain target defects. Target defects include several types of defects that have a significant impact on concrete quality, including exposed reinforcement, honeycombs, voids, slag inclusions, cracks, etc.

[0050] S105: Based on the pose information of the target defect corresponding to the original image, determine the coordinate information of each target defect and mark it;

[0051] When defects are detected on the concrete surface, the coordinate information of each target defect needs to be determined so that workers can quickly know the location of the defect when arranging the subsequent defect elimination work.

[0052] S106: Collecting the coordinate information of all target defects and performing structured processing to output structured data.

[0053] The collection includes but is not limited to: 1) the original image including the target defect; 2) the type of the target defect; 3) the coordinates of the target defect;

[0054] Structural processing includes defining and saving the collected information as structured information. For example, the video frame corresponding to the detected target defect image is saved as image data of a specific size (such as a 640*640 png format image), the target defect type is saved as honeycomb, void, slag inclusion, or crack, and the coordinates of the defect are saved as coordinates represented by three numerical values ​​in the world coordinate system.

[0055] Beneficial effects: 1) It facilitates data storage and transmission; 2) It facilitates data utilization. For example, a simple algorithm can be used to set the edge detection equipment to only save and transmit a certain type of target defect data, and discard or not transmit other types of defect data with less obvious characteristics, thereby saving computing resources and network bandwidth; engineers can write simple algorithms to determine the location of the target defect in the project based on the spatial coordinates of the target defect, and then rectify possible hazardous sources through management actions; 3) It facilitates standardization and promotion and application.

[0056] The specific implementation of this embodiment includes:

[0057] This embodiment performs defect target detection on each frame of the original image of the inspection video to obtain the original image with quality defects, so as to delete the original image without quality defects in the inspection video; each defect in the original image with quality defects is pre-classified to obtain the target defect set, so as to avoid the need for a large amount of algorithm training and calculation in the subsequent defect identification process, making the algorithm lightweight. By identifying all quality defects in the target defect set, multiple target defects of the target defect type and the coordinate information corresponding to each target defect are obtained and structured processing is performed, and structured data is output to manage the defects of the target defect type. The technical problem existing in the prior art of inadequate inspection of concrete quality defects at the construction site leading to untimely rectification of on-site defects is solved.

[0058] In this embodiment, defect target detection is performed on each frame of the original image, including:

[0059] Each frame of the original image is input into the object detection model, which identifies each quality defect in the original image and marks the defect box. The object detection model is trained on the target dataset using deep learning methods.

[0060] It should be noted that the size of the defect box is related to the size of the quality defect. The larger the quality defect, the larger the defect box; the smaller the quality defect, the smaller the defect box; the center of the defect box is the center of the quality defect.

[0061] Pre-classify each quality defect in the original image with quality defects to obtain the target defect set, including:

[0062] Based on the pose information corresponding to the original image where the quality defect is located and the position of the quality defect in the original image, determine the depression angle of the center of the quality defect relative to the camera;

[0063] All quality defects with a depression angle greater than a set number of degrees are obtained to generate a target defect set. In this embodiment, the set number of degrees is 17 degrees.

[0064] In this embodiment, the target defect types include exposed rib defects, honeycomb defects, void defects, slag inclusion defects, and crack defects; all quality defects in the target defect set are identified to obtain multiple target defects of the target defect type, including: identifying each quality defect in the target defect set through a preset target recognition model to obtain the target defect; marking each target defect with the corresponding target defect type.

[0065] In this embodiment, exposed reinforcement defects, honeycomb defects, void defects, slag inclusion defects, and crack defects are selected as target defect types because they have a great impact on concrete quality and occur frequently.

[0066] In this embodiment, based on the pose information of the target defect corresponding to the original image, the coordinate information of each target defect is determined and marked, including:

[0067] Construct a spatial rectangular coordinate system for the construction site;

[0068] Based on the pose information of the target defect corresponding to the original image and the position of the center of the target defect in the original image, the defect equation of the line where the center of the target defect is located is confirmed; including: confirming the camera position equation (x0, y0, z0) based on the camera coordinate data; confirming the direction vector (a, b, c) of the quality defect center relative to the camera based on the camera angle pose data and the position of the center of the target defect in the original image; based on the camera position equation (x0, y0, z0) and the direction vector (a, b, c), confirming the defect equation as:

[0069] According to the defect equation and the surface equation of the inspection floor, the coordinates of the quality defect center are determined, including: obtaining the surface equation of the inspection floor Ax+By+Cz+D=0; and simultaneously solving the defect equations to obtain the coordinates of the quality defect center:

[0070]

[0071] It should be noted that the detection method also includes: before performing defect target detection on each frame of original image, each frame of original image needs to be preprocessed; the preprocessing includes image compression, noise filtering, edge extraction and image sharpening.

[0072] Before performing defect target detection on each frame of original image, preprocessing each frame of original image can make the image features of quality defects more prominent, which is convenient for the target detection model to detect.

[0073] Example 2:

[0074] The present invention provides a computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements a detection method when executing the computer program: obtaining each frame of an original image of an inspection video and its posture information; the posture information comprises camera coordinate data and camera angle posture data; performing defect target detection on each frame of the original image, and retaining the original image with quality defects; pre-classifying each quality defect in the original image with quality defects to obtain a target defect set; identifying all quality defects in the target defect set to obtain multiple target defects of the target defect type; determining and marking the coordinate information of each target defect based on the posture information of the original image corresponding to the target defect; aggregating the coordinate information of all target defects and performing structured processing to output structured data.

[0075] Example 3:

[0076] The present invention provides a readable storage medium having a computer program stored thereon, which implements a detection method when executed by a processor: obtaining each frame of an original image of an inspection video and its posture information; the posture information includes camera coordinate data and camera angle posture data; performing defect target detection on each frame of the original image, and retaining the original image with quality defects; pre-classifying each quality defect in the original image with quality defects to obtain a target defect set; identifying all quality defects in the target defect set to obtain multiple target defects of the target defect type; determining and marking the coordinate information of each target defect based on the posture information of the original image corresponding to the target defect; collecting the coordinate information of all target defects and performing structured processing to output structured data.

[0077] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0078] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device that includes a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a..." does not exclude the presence of additional identical elements in the process, method, article, or device that includes the element.

Claims

1. A method for detecting concrete quality defects, characterized in that: The detection method comprises: Obtain each frame of the original image of the inspection video and its posture information; the posture information includes camera coordinate data and camera angle posture data; Perform defect target detection on each frame of original image and retain the original images with quality defects; Pre-classify each quality defect in the original image with quality defects to obtain a target defect set; Identify all quality defects in the target defect set to obtain multiple target defects of the target defect type; Based on the pose information of the target defect corresponding to the original image, the coordinate information of each target defect is determined and marked; Collect the coordinate information of all target defects, perform structured processing, and output structured data; Among them, each quality defect in the original image with quality defects is pre-classified to obtain the target defect set, including: Based on the pose information corresponding to the original image where the quality defect is located and the position of the quality defect in the original image, determine the depression angle of the center of the quality defect relative to the camera; Obtain all quality defects with depression angles greater than a set number of degrees to generate a target defect set; The target defect types include exposed rib defects, honeycomb defects, void defects, slag inclusion defects, and crack defects. All quality defects in the target defect set are identified to obtain multiple target defects of the target defect type, including: Identify each quality defect in the target defect set through a preset target recognition model to obtain the target defect; Each target defect is marked with the corresponding target defect type.

2. A method for detecting concrete quality defects according to claim 1, characterized in that: Perform defect target detection on each frame of the original image, including: Each frame of the original image is input into the target detection model, which marks and identifies each quality defect in the original image and marks the defect box.

3. A method for detecting concrete quality defects according to claim 2, characterized in that: Based on the pose information of the target defect corresponding to the original image, the coordinate information of each target defect is determined and marked, including: Construct a spatial rectangular coordinate system for the construction site; Based on the pose information of the target defect corresponding to the original image and the position of the center of the target defect in the original image, the defect equation of the straight line where the center of the target defect is located is determined; According to the defect equation and the surface equation of the inspection floor, the coordinates of the center of the quality defect are confirmed.

4. A method for detecting concrete quality defects according to claim 3, characterized in that: Based on the pose information of the target defect corresponding to the original image and the position of the center of the target defect in the original image, the defect equation of the line where the center of the target defect is located is determined, including: Based on the camera coordinate data, confirm the camera position equation (x0, y0, z0); Based on the camera angle pose data and the position of the center of the target defect in the original image, determine the direction vector (a, b, c) of the quality defect center relative to the camera; Based on the camera position equation (x0, y0, z0) and direction vector (a, b, c), the defect equation is confirmed to be:

5. A method for detecting concrete quality defects according to claim 4, characterized in that: According to the defect equation and the surface equation of the inspection floor, the coordinates of the quality defect center are determined, including: Get the equation of the surface where the inspection floor is located: Ax+By+Cz+D=0; Combine the defect equations to obtain the coordinates of the mass defect center:

6. A method for detecting concrete quality defects according to claim 1, characterized in that: The detection method also includes: before performing defect target detection on each frame of original image, each frame of original image needs to be preprocessed; the preprocessing includes image compression, noise filtering, edge extraction and image sharpening.

7. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

8. A readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images