Building wall hollowing defect detection system and method

Abstract

The invention belongs to the technical field of building detection, and discloses a building wall hollowing defect detection system and method.The system comprises an unmanned aerial vehicle, a thermal infrared imager, a high-definition camera, a labeling mechanism, a hollowing recognition mechanism, an anomaly detection mechanism and a route planning mechanism; an infrared thermal imager is fixedly arranged at the front end of the unmanned aerial vehicle; the hollowing identification mechanism is used for analyzing the collected infrared images and identifying hollowing defects on the wall body; the abnormity detection mechanism is used for carrying out abnormity checking on the hollowing defect position and checking whether an abnormal condition exists at the hollowing defect position of the wall body or not; the labeling mechanism comprises a coating storage box and a suction pump; the flight route planning mechanism reasonably plans the flight route of the unmanned aerial vehicle according to the position coordinates of all the walls; and the total flight distance of the unmanned aerial vehicle is shortest. According to the invention, the collected infrared image is analyzed through the hollowing identification mechanism, and hollowing defects and position coordinate information on the wall are identified in time; and a mark is sprayed on the wall body at the position through the marking mechanism.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of building detection, in particular to a building wall hollow defect detection system and method. BACKGROUND

[0002] The hollow of the building outer wall is easy to cause the outer wall facing brick to fall, form high-altitude falling objects, and possibly injure vehicles and pedestrians; after the wall skin falls off, the wall body loses a layer of protection and directly bears the sun and rain, which may slowly cause leakage and affect the normal use and service life of the building: the falling of the facing brick or wall skin also affects the beauty of the building surface, as shown in the following picture, and people can see a "big face" from a distance, and the original beautiful appearance of the building no longer exists, and people also doubt the overall quality and safety of the building, causing panic. The outer wall facing is mainly divided into facing brick, mosaic and mortar according to the material type, and is widely used in the facing engineering of the building outer wall due to its beauty and economy. However, with the passage of time, the bonding quality of the facing will gradually decrease due to temperature stress, weathering and other factors, and the facing is easy to hollow and fall off, causing safety accidents such as personnel injury and property loss. Therefore, the bonding defect detection of the building outer wall (especially the high-rise building with a service life of more than 10 years) is becoming an important work that cannot be delayed. The single visual inspection and knocking detection method has great subjectivity, low work efficiency and poor detection effect.

[0003] The prior art document with the publication number CN108445078A provides a building outer wall hollow detection device and method. The building outer wall hollow detection device comprises a launching mechanism for launching light small balls, a cartridge for storing the light small balls, a cloud platform and a sound pickup device for listening to the sound emitted by the light small balls launched by the launching mechanism when hitting the building outer wall. By launching the light small balls through the launching mechanism, and listening to the sound emitted by the light small balls hitting the building outer wall through the sound pickup device, it can be determined whether the building outer wall is hollow according to the sound listened to, and the detection is convenient; the cloud platform is provided to connect the unmanned aerial vehicle, so that the detection of any point can be facilitated, and the efficiency is high; after the light small balls are launched, they will leave the unmanned aerial vehicle and completely separate from the unmanned aerial vehicle, so that after the light small balls are bounced back by the building outer wall, they will not affect the unmanned aerial vehicle, and the balance of the unmanned aerial vehicle in the air can be maintained well, so that the unmanned aerial vehicle can fly smoothly.

[0004] The prior technical solutions in the above have the following defects: 1. The position of the wall for the hollow detection by the unmanned aerial vehicle and the light small ball is random, cannot cover the entire wall surface, and cannot timely and accurately detect the defects of the hollow of the wall. 2. There are more noises outdoors, the hollow defect is detected by the existing technology through the pickup to listen to the sound of the light small ball hitting the outer wall of the building, and the detection result may be inaccurate, and the detection process is easily disturbed by the external noise.

[0005] In view of this, the building wall hollow defect detection system and method are provided. SUMMARY

[0006] 1. Technical problem to be solved

[0007] The building wall hollow defect detection system and method are provided, which solves the technical problems in the above background art, realizes the hollow detection of the wall by the unmanned aerial vehicle and the thermal imager, analyzes the collected infrared image by the hollow recognition mechanism, timely recognizes the hollow defects on the wall, records the position coordinate information, size information and height information of the hollow defects, and sprays the marks on the wall at the position by the marking mechanism, so that the defects position can be quickly found in the subsequent process.

[0008] 2. Technical solution

[0009] The technical solution of the present application provides a building wall hollow defect detection system, which comprises an unmanned aerial vehicle, an infrared thermal imager, a high-definition camera, an unmanned aerial vehicle altimeter, a marking mechanism, a hollow recognition mechanism, an abnormality detection mechanism and a flight path planning mechanism.

[0010] The front end of the unmanned aerial vehicle is fixedly provided with the infrared thermal imager and the high-definition camera, the infrared image of the wall is collected by the infrared thermal imager, and the image of the wall is collected by the high-definition camera.

[0011] The unmanned aerial vehicle altimeter is fixedly arranged on the unmanned aerial vehicle and is used for monitoring and recording the flight height of the unmanned aerial vehicle.

[0012] The hollow recognition mechanism analyzes the collected infrared image and recognizes the hollow defects on the wall, records the position coordinate information, size information and height information of the hollow defects when the hollow defects on the wall are detected.

[0013] The abnormality detection mechanism performs abnormality investigation on the position of the detected hollow defects, investigates whether there is an abnormal situation affecting the detection result at the position of the wall hollow defects, and if there is an abnormal situation such as steel bars, pipelines, windows, air conditioners and the like interfering with the hollow detection at the position, it is considered that there is no hollow defect at the position, and the position is ignored.

[0014] The marking mechanism comprises a paint storage tank and a suction pump; the paint storage tank and the suction pump are fixedly arranged on the unmanned aerial vehicle, the input end of the suction pump is in communication with the paint storage tank, and the output end of the suction pump is connected with a spray head through a pipeline. A flow control valve is arranged on the pipeline. The paint storage tank is filled with red or yellow paint, which is easy to be found and the defect position is quickly found after the red or yellow paint is sprayed on the wall.

[0015] The flight route planning mechanism: the wall to be detected is numbered, the position coordinate information of each wall is collected, the height and width information of the wall is collected, the flight route of the unmanned aerial vehicle is reasonably planned according to the position coordinates of each wall, the total flight distance of the unmanned aerial vehicle is shortest, the flight route and height of the unmanned aerial vehicle are planned for each wall according to the monitoring range of the infrared thermal imager, the flight height of the unmanned aerial vehicle is from low to high, the unmanned aerial vehicle flies according to an S-shaped route, and there is a repeated detection area in the adjacent two return processes; so that the monitoring range of the infrared thermal imager can cover the entire wall without omission.

[0016] The application provides a building wall hollow defect detection method, which comprises the following steps:

[0017] S1, the wall to be detected is numbered, the position coordinate information of each wall is collected, the height and width information of the wall is collected;

[0018] S2, the flight route of the unmanned aerial vehicle is reasonably planned according to the position coordinates of each wall by the flight route planning mechanism; the total flight distance of the unmanned aerial vehicle is shortest;

[0019] S3, the flight route and height of the unmanned aerial vehicle are planned for each wall according to the monitoring range of the infrared thermal imager, the flight height of the unmanned aerial vehicle is from low to high, the unmanned aerial vehicle flies according to an S-shaped route, and there is a repeated detection area in the adjacent two return processes; so that the monitoring range of the infrared thermal imager can cover the entire wall without omission.

[0020] S4, an air environment data sensor is arranged at each wall for detecting real-time environment data of each sub-area;

[0021] S5, whether the wall hollow detection can be performed by the unmanned aerial vehicle and the thermal imager is judged according to the real-time environment data, if the detection can be performed, the flight environment can be used unmanned aerial vehicle type and camera parameters are determined; according to the size and terrain of the detection area, the appropriate unmanned aerial vehicle model and configuration are selected. Ensure that the battery of the unmanned aerial vehicle is fully charged;

[0022] S6, the wall is detected for hollow according to the planned flight route by the unmanned aerial vehicle and the thermal imager, and the infrared image of the wall surface is acquired;

[0023] S7, the hollow identification mechanism analyzes the collected infrared image, and identifies the hollow defect on the wall; when detecting that the wall has a hollow defect, record the position coordinate information, size information and height information of the hollow defect;

[0024] S8, the abnormality detection mechanism carries out abnormality investigation on the position of the detected hollow defect, and investigates whether there is an abnormal situation affecting the detection result at the position of the wall hollow defect; if there is abnormal situation such as reinforcing steel bar, pipeline, window, air conditioner and the like interfering with the hollow detection at the position, it is considered that there is no hollow defect at the position, and the position is ignored;

[0025] S9, confirming that there is no abnormal situation at the position of the wall hollow defect in the detection position; through the marking mechanism, mark is sprayed on the wall at the position, so that the defect position can be quickly found in the subsequent process.

[0026] As an optional scheme of the application, in step S2, three-dimensional A* algorithm is used to plan the flight route of the unmanned aerial vehicle, so that the total flight distance of the unmanned aerial vehicle is the shortest;

[0027] The to-be-expanded point is stored in the open table, and the found minimum cost waypoint is stored in the close table, and the steps of the three-dimensional A* algorithm for path planning include:

[0028] S21, set the planning starting point as S and the ending point as E, the starting point coordinates as (S x ,S y ,S z ), the ending point coordinates as (E x ,E y ,E z ), and the current waypoint coordinates as (N x ,N y ,N z ); set the initial coordinates of the current waypoint as the coordinates of the starting point S, i.e. (N x ,N y ,N z ) = (S x ,S y ,S z );

[0029] S22, judge whether |S x -E x | ≥ |S y -E y |, this condition is used to determine which axis (x axis or y axis) to expand the node first;

[0030] If the condition is not met, go to step S25, and expand the node along the y axis direction first;

[0031] If it is met, continue; expand the node along the x axis direction first;

[0032] S23. Expand the nodes along the x-axis, calculate the coordinates of each point, and check whether they are within the threat range (such as obstacles, no-fly zones, etc.). If there are points within the threat range, remove them. If the z-coordinate of the waypoint to be expanded is less than the actual elevation value corresponding to the point, these points should also be removed.

[0033] For the remaining candidate nodes, calculate their g value (the actual cost from the starting point to the node) and h value (the estimated cost from the node to the destination, usually expressed as distance). Calculate the f value based on the g and h values, f = (g + h), and add these nodes to the open list (a list of nodes to be expanded); f is the total cost (distance).

[0034] S24. Select the point with the smallest f value from the open table and set it as the current waypoint, i.e., update (N). x N y N z Let S be the coordinates of that point; x =N x S y =N y This expansion is now complete.

[0035] • S25. Expand nodes along the y-axis, using the same method as step S23; however, this time candidate nodes are generated along the y-axis. Similarly, check the threat range and actual elevation value, remove points that do not meet the conditions, calculate the g, h, and f values ​​of the remaining candidate nodes, and add them to the open table.

[0036] S26. Select the point with the smallest f value from the open table and set it as the current waypoint. Let S x =N x S y =N y This expansion is now complete.

[0037] S27. Repeat steps S22 to S26 until the destination E is found or the open table is empty (meaning there is no path to the destination). During each expansion, the close table (a list of waypoints with the lowest cost) also needs to be updated, and the processed waypoints are added to it to avoid duplicate processing.

[0038] As an optional embodiment of the present invention, real-time environmental data includes: temperature, humidity, wind speed, particulate matter concentration, light intensity, and weather (sunny, cloudy, rainy, and snowy, etc.);

[0039] The drone information includes model, weight, flight time, payload, flight speed, and operating environment range. Camera parameter information includes camera model, resolution, focal length, and aperture. Meteorological report data includes weather, air pressure, visibility, and precipitation.

[0040] As an optional solution of the present application, the infrared obstacle avoidance sensor is fixedly arranged on the unmanned aerial vehicle, and the infrared obstacle avoidance sensor can be used to monitor the obstacles on the flight route during the flight of the unmanned aerial vehicle, so that the unmanned aerial vehicle can timely avoid the obstacles and avoid collision accidents.

[0041] As an optional solution of the present application, the abnormality detection mechanism is fixedly arranged on the unmanned aerial vehicle; in the building wall hollow defect detection, the steel bars, pipelines, windows, air conditioners and the like on the wall may interfere with the hollow detection, so special attention should be paid to these abnormal conditions when analyzing the images.

[0042] The abnormality detection mechanism comprises:

[0043] The data collection module: a large number of images of the steel bars, pipelines, windows, air conditioners and the like on the wall are collected, and the images are labeled as reference samples;

[0044] The image acquisition module: comprising a high-definition camera and an LED, the image of the wall for hollow detection is acquired;

[0045] The image preprocessing module: the acquired wall image is preprocessed, including denoising, contrast enhancement and color correction and the like;

[0046] The feature extraction module: the preprocessed image is subjected to feature extraction, and the features related to the steel bars, pipelines, windows, air conditioners and the like are extracted by an edge detection method; including color, texture and shape features;

[0047] The abnormality recognition module: the image after feature extraction is compared and analyzed with the reference sample, and whether the wall hollow detection position exists the abnormal conditions such as the steel bars, pipelines, windows, air conditioners and the like is timely recognized.

[0048] As an optional solution of the present application, the infrared thermal imager preferably adopts the FLUKE TiR32 infrared thermal imager.

[0049] As an optional solution of the present application, the hollow recognition mechanism comprises:

[0050] The infrared image acquisition unit: comprising an infrared thermal imager, the infrared thermal imager is used to scan the building wall surface and acquire the infrared image of the wall surface; the infrared thermal imager can capture the temperature difference of different parts of the wall surface, which is the key to detect the hollow;

[0051] The infrared image preprocessing unit: the collected infrared image is preprocessed to improve the image quality and the accuracy of feature extraction. The preprocessing steps may include image denoising, contrast and clarity enhancement and the like;

[0052] Infrared image feature extraction unit: extract useful information and features from the pre-processed infrared image, which may include temperature anomaly areas, hot spots or cold spots, etc.

[0053] Hollow identification unit: based on the extracted features, identify the area of the hollow. Generally, the hollow area will appear as an area different from the surrounding wall temperature in the infrared image, forming a clear hot spot or cold spot;

[0054] Hollow profile size identification unit: according to the infrared image extracted by feature extraction, the edge detection method is used to extract the edge profile of the wall hollow defect; the size of the wall hollow defect is calculated and corresponds to the hollow defect position height coordinate one by one, and is stored in the information storage unit;

[0055] Information storage unit: used to store information in the process of wall hollow defect detection, including unmanned aerial vehicle flight route, real-time environmental data, collected images and infrared images, and detection results.

[0056] As an optional solution of the present application, the hollow profile size identification unit specifically includes the following contents:

[0057] A, edge detection:

[0058] A1, gradient calculation: first, gradient calculation is performed on the pre-processed infrared image; gradient represents the rate of change of pixel value in the image, which helps to highlight the edge information. By calculating the gradient intensity and direction of each pixel point in the image, preliminary information of the edge can be obtained.

[0059] A2, edge enhancement: in order to display the edge more clearly, filter (such as Sobel, Canny, etc.) can be used to filter the gradient image to enhance the edge features and reduce noise interference.

[0060] A3, threshold processing: set appropriate threshold to binarize the enhanced edge image, so that the edge profile of the hollow defect will be presented in the form of obvious black and white contrast.

[0061] B, profile extraction:

[0062] B1, contour tracking: in the binarized image, the edge profile of the hollow defect is extracted by contour tracking algorithm (such as chain code representation method, Freeman chain code, etc.). These algorithms can traverse the pixel points in the image and identify the continuous boundary point sequence;

[0063] B2, contour optimization: smooth the extracted contour to eliminate the contour jitter or breakage phenomenon caused by noise or poor image quality. This can be achieved by filtering, interpolation, etc.

[0064] C, size calculation:

[0065] C1, Contour fitting: fitting the extracted hollowing contour with a suitable geometric shape (such as rectangle, ellipse, etc.).

[0066] C2, Size calculation: by calculating the parameters (such as length, width, area, etc.) of the fitted shape, the size of the hollowing defect can be obtained, and the wall hollowing and its hollowing range can be objectively and accurately evaluated according to the position and size of the hollowing, serving as a strong basis for the maintenance of the exterior wall finish.

[0067] 3. Beneficial effects

[0068] One or more technical solutions provided in the technical solution of the present application have at least the following technical effects or advantages:

[0069] 1. The flight route of the unmanned aerial vehicle is reasonably planned by the flight route planning mechanism according to the position coordinates of each wall, so that the total flight distance of the unmanned aerial vehicle is the shortest.

[0070] 2. According to the monitoring range of the infrared thermal imager, the flight route and height of the unmanned aerial vehicle for each wall are planned, and the unmanned aerial vehicle flies according to the S-shaped route, so that the monitoring range of the infrared thermal imager can cover the entire wall without omission.

[0071] 3. Whether the wall hollowing detection can be performed by the unmanned aerial vehicle and the thermal imager can be determined according to real-time environmental data, and if the detection can be performed, the unmanned aerial vehicle model and camera parameters that can be used in the flight environment are determined, so as to ensure that the wall hollowing detection is performed under appropriate flight detection conditions.

[0072] 4. The wall hollowing is detected by the unmanned aerial vehicle and the thermal imager, the infrared image collected is analyzed by the hollowing recognition mechanism, and the hollowing defect on the wall is identified in a timely manner.

[0073] 5. When the hollowing defect on the wall is detected, the position coordinate information, size information and height information of the hollowing defect are recorded, and the marking mechanism is used to spray marks on the wall at the position, so as to facilitate the subsequent rapid finding of the defect position.

[0074] 6. The infrared thermal imaging method uses a non-contact method, does not need a scaffold, and can quickly and large-area scan the exterior wall finish of a building, so as to objectively and accurately evaluate whether the finish is hollowed and the hollowing range thereof, serving as a strong basis for the maintenance of the exterior wall finish. BRIEF DESCRIPTION OF DRAWINGS

[0075] Figure 1 The flowchart of the building wall hollowing defect detection method disclosed in a preferred embodiment of the present application;

[0076] Figure 2A schematic view of a hollow identification mechanism of a building wall hollow defect detection system disclosed in a preferred embodiment of the present application;

[0077] Figure 3 A schematic view of an abnormality detection mechanism of a building wall hollow defect detection system disclosed in a preferred embodiment of the present application;

[0078] Figure 4 A schematic view of the overall structure of a building wall hollow defect detection system disclosed in a preferred embodiment of the present application. DETAILED DESCRIPTION

[0079] The present application will be further described in detail below in conjunction with the accompanying drawings.

[0080] Reference Figure 1 The embodiment of the present application provides a building wall hollow defect detection method, which comprises the following steps:

[0081] S1, numbering the wall to be detected, collecting the position coordinate information of each wall; collecting the height and width information of the wall;

[0082] S2, the flight route of the unmanned aerial vehicle is reasonably planned by the flight route planning mechanism according to the position coordinates of each wall; so that the total flight distance of the unmanned aerial vehicle is the shortest;

[0083] S3, according to the monitoring range of the infrared thermal imager, the flight route and height of the unmanned aerial vehicle are planned for each wall, the flight height of the unmanned aerial vehicle is from low to high, and the unmanned aerial vehicle flies according to the S-shaped route, and there is a repeated part of the detection area in the adjacent two round trips; so that the monitoring range of the infrared thermal imager can cover the entire wall without omission;

[0084] S4, an air environment data sensor is arranged at each wall, which is used for detecting the real-time environment data of each sub-area;

[0085] S5, whether the wall hollow detection can be carried out by the unmanned aerial vehicle and the thermal imager is judged according to the real-time environment data, if the detection can be carried out, the unmanned aerial vehicle model and camera parameters that can be used in the flight environment are determined; according to the size and terrain of the detection area, the appropriate unmanned aerial vehicle model and configuration are selected. Ensure that the battery of the unmanned aerial vehicle is fully charged;

[0086] S6, according to the planned flight route, the wall is detected by the unmanned aerial vehicle and the thermal imager, and the infrared image of the wall surface is obtained;

[0087] S7, the hollow identification mechanism analyzes the collected infrared image, and identifies the hollow defect on the wall; when the hollow defect on the wall is detected; the position coordinate information, size information and height information of the hollow defect are recorded;

[0088] S8, the abnormality detection mechanism performs abnormality investigation on the detected hollow defect position, investigates whether the wall hollow defect position exists abnormal situation affecting the detection result; if there is abnormal situation such as reinforcing steel bar, pipeline, window, air conditioner and the like interfering with the hollow detection at the position, it is considered that there is no hollow defect at the position, and the position is ignored;

[0089] S9, confirming that there is no abnormal situation at the wall hollow defect position; through the marking mechanism, marking is sprayed on the wall at the position, so as to facilitate subsequent rapid finding of the defect position.

[0090] Further, in step S2, the flight route of the unmanned aerial vehicle is planned by using the three-dimensional A* algorithm, so that the total flight distance of the unmanned aerial vehicle is the shortest;

[0091] The to-be-expanded point is stored in the open table, and the found minimum cost waypoint is put into the close table. The steps of the three-dimensional A* algorithm for path planning include:

[0092] S21, setting the planning starting point as S and the ending point as E, the starting point coordinates as (S x ,S y ,S z ), the ending point coordinates as (E x ,E y ,E z ), and the current waypoint coordinates as (N x ,N y ,N z ); setting the initial coordinates of the current waypoint as the coordinates of the starting point S, i.e. (N x ,N y ,N z ) = (S x ,S y ,S z );

[0093] S22, judging whether |S x -E x | ≥ |S y -E y | is met or not. This condition is used to determine which axis (x axis or y axis) is firstly expanded;

[0094] If the condition is not met, jump to step S25, and firstly expand the node along the y axis direction;

[0095] If the condition is met, continue; firstly expand the node along the x axis direction;

[0096] S23, expand the node along the x-axis direction, respectively calculate the coordinate value of each point, check whether it is located in the threat range (such as obstacles, no-fly zones, etc.); if there is a point located in the threat range, it is removed, and if the z coordinate value of the to-be-expanded waypoint is less than the actual elevation value corresponding to the point, these points are also removed;

[0097] For the remaining candidate nodes, calculate the g value (the actual cost from the starting point to the point) and the h value (the estimated cost from the point to the terminal point, usually using distance). According to the g value and the h value, calculate the f value, f = (g + h), and add these nodes to the open table (the list of nodes to be expanded); f is the total cost (distance).

[0098] S24, select the point with the minimum f value from the open table, and set it as the current waypoint, that is, update (N x ,N y ,N z ) as the coordinates of the point; let S x =N x , S y =N y , and the current expansion ends.

[0099] S25, expand the node along the y-axis direction, and the specific method is the same as step S23; but this time, generate candidate nodes along the y-axis direction, and also check the threat range and the actual elevation value, remove the points that do not meet the conditions, calculate the g value, h value and f value of the remaining candidate nodes, and add them to the open table.

[0100] S26, select the point with the minimum f value from the open table, and set it as the current waypoint, let S x =N x , S y =N y , and the current expansion ends.

[0101] S27, repeat steps S22 to S26 until the terminal point E is found or the open table is empty (indicating that there is no path to reach the terminal point), and at each expansion, the close table (the list of waypoints with the minimum cost found) also needs to be updated, and the processed waypoints are added to it to avoid repeated processing.

[0102] Further, the real-time environmental data includes: temperature, humidity, wind speed, particulate matter concentration, light intensity, weather (sunny, cloudy, rain and snow, etc.);

[0103] The unmanned aerial vehicle model information includes model, weight, endurance time, load, flight speed, working environment range, camera parameter information includes camera model, resolution, focal length and aperture, and meteorological report data includes weather, air pressure, visibility and precipitation.

[0104] The suitable use environment of the infrared thermal imager is: the ambient temperature is 0-40 DEG C; there is no rain, the relative humidity RH should be no more than 75%, and there is no dew; the average wind speed is no more than 3 m / s outdoors; when measuring, the non-measured object should be avoided to enter the imaging range.

[0105] The infrared obstacle avoidance sensor is fixedly arranged on the unmanned aerial vehicle, and the obstacle on the flight route can be monitored through the infrared obstacle avoidance sensor during the flight of the unmanned aerial vehicle, so that the unmanned aerial vehicle can timely avoid the obstacle and avoid collision accidents.

[0106] Reference Figure 3 Further, the abnormality detection mechanism is fixedly arranged on the unmanned aerial vehicle; in the building wall hollow defect detection, the steel bars, pipelines, windows, air conditioners and the like on the wall may interfere with the hollow detection, and therefore special attention should be paid to these abnormal conditions when analyzing the image.

[0107] The abnormality detection mechanism comprises:

[0108] The data collection module: a large number of images of the steel bars, pipelines, windows, air conditioners and the like on the wall are collected, and the images are labeled as reference samples;

[0109] The image acquisition module: comprising a high-definition camera and an LED, the image of the wall for hollow detection is acquired;

[0110] The image preprocessing module: the acquired wall image is preprocessed, including denoising, contrast enhancement and color correction and the like;

[0111] The feature extraction module: the features related to the steel bars, pipelines, windows, air conditioners and the like are extracted from the preprocessed image through the edge detection method; including color, texture and shape features;

[0112] The abnormality recognition module: the image after feature extraction is compared and analyzed with the reference sample, and whether the steel bars, pipelines, windows, air conditioners and the like exist in the wall hollow detection position is recognized in time.

[0113] Further, the infrared thermal imager is preferably adopted, and the FLUKE TiR32 infrared thermal imager is adopted;

[0114] The detection principle of the infrared thermal imaging method: the infrared thermal imager is used to reflect the energy distribution pattern of the measured target to the photosensitive element of the infrared detector through the infrared detector and the optical imaging objective, so as to obtain the infrared thermal image. The thermal image corresponds to the thermal distribution field of the object surface. The infrared thermal imager converts the invisible infrared energy emitted by the object into a visible thermal image. Different colors on the thermal image represent different temperatures of the measured object, and warm and cold colors represent temperature high and low, or bright white represents high temperature and dark black represents low temperature.

[0115] For a certain outer wall surface, under normal circumstances, the infrared thermal radiation occurring on the entire surface layer should be uniform, and the color of the corresponding area on the infrared thermal image is single and uniform, without obvious color difference. When it is abnormal, such as the existence of defects such as hollowing and leakage in the surface layer, the thermal conductivity of the wall surface layer and the wall body is directly changed at these positions, and therefore there will be infrared radiation abnormalities relative to the normal positions. In this case, the color of the abnormal area on the infrared thermal energy picture is not single and uniform. By comparing the color of the abnormal area on the infrared thermal energy picture with the ordinary picture, excluding the interference factors such as stains and shadows on the outer wall surface, the position of the abnormal area in the actual wall body can be directly and clearly seen.

[0116] Reference Figure 2 Further, the hollow recognition mechanism comprises:

[0117] The infrared image acquisition unit comprises an infrared thermal imager, which is used to scan the building wall surface and acquire the infrared image of the wall surface. The infrared thermal imager can capture the temperature difference of different parts of the wall surface, which is the key to detecting hollowing;

[0118] The infrared image preprocessing unit pre-processes the collected infrared image to improve the image quality and the accuracy of feature extraction. The preprocessing steps may include image denoising, contrast enhancement and clarity enhancement, etc.

[0119] The infrared image feature extraction unit extracts useful information and features from the pre-processed infrared image, which may include temperature abnormal areas, hot spots or cold spots, etc.

[0120] The hollow recognition unit recognizes the hollow area based on the extracted features. Generally, the hollow area will appear as an area with different temperature from the surrounding wall surface in the infrared image, forming obvious hot spots or cold spots.

[0121] The hollow contour size identification unit extracts the edge contour of the wall hollow defect by using edge detection method according to the infrared image extracted by the feature extraction unit, calculates the size of the wall hollow defect, and stores the size corresponding to the position and height coordinates of the wall hollow defect in the information storage unit.

[0122] The information storage unit is used to store the information in the process of detecting the wall hollow defect, including the flight route of the unmanned aerial vehicle, real-time environmental data, collected images and infrared images, and detection results.

[0123] Further, the hollow contour size identification unit specifically comprises the following contents:

[0124] A. Edge detection:

[0125] A1, Gradient calculation: First, the preprocessed infrared image is gradient calculated; the gradient represents the rate of change of pixel value in the image, which helps to highlight the edge information. By calculating the gradient intensity and direction of each pixel point in the image, the preliminary information of the edge can be obtained.

[0126] A2, Edge enhancement: In order to display the edge more clearly, a filter (such as Sobel, Canny, etc.) can be used to filter the gradient image to enhance the edge features and reduce noise interference.

[0127] A3, Threshold processing: Set a suitable threshold to binarize the enhanced edge image, so that the edge profile of the hollow defect will be presented in the form of obvious black and white contrast.

[0128] B, Contour extraction:

[0129] B1, Contour tracking: In the binarized image, the edge profile of the hollow defect is extracted by a contour tracking algorithm (such as chain code representation, Freeman chain code, etc.). These algorithms can traverse the pixel points in the image and identify the continuous boundary point sequence;

[0130] B2, Contour optimization: The extracted contour is smoothed to eliminate the contour jitter or break phenomenon caused by noise or poor image quality. This can be achieved by filtering, interpolation, etc.

[0131] C, Size calculation:

[0132] C1, Contour fitting: The extracted hollow contour is fitted using appropriate geometric shapes (such as rectangle, ellipse, etc.).

[0133] C2, Size calculation: By calculating the parameters (such as length, width, area, etc.) of the fitted shape, the size of the hollow defect can be obtained. The position and size of the hollow can be objectively and accurately evaluated to assess the wall hollow and its hollow range, which can serve as a powerful basis for exterior wall decoration maintenance.

[0134] Referring Figure 4 , the present application provides a kind of building wall hollow defect detection system, comprising: unmanned aerial vehicle, infrared thermal imager, high-definition camera, unmanned aerial vehicle altimeter, marking mechanism, hollow identification mechanism, anomaly detection mechanism and route planning mechanism;

[0135] The front end of the unmanned aerial vehicle is fixedly provided with an infrared thermal imager and a high-definition camera, which collects infrared images of the wall by the infrared thermal imager; the high-definition camera collects images of the wall;

[0136] The unmanned aerial vehicle is fixedly provided with an unmanned aerial vehicle altimeter for monitoring and recording the flight height of the unmanned aerial vehicle;

[0137] The air-dry identification mechanism: analyzes the collected infrared image, identifies the air-dry defect on the wall, and records the position coordinate information, size information and height information of the air-dry defect when the air-dry defect on the wall is detected.

[0138] The abnormality detection mechanism: carries out abnormality investigation on the position of the detected air-dry defect, investigates whether there is an abnormal situation affecting the detection result at the position of the air-dry defect of the wall, and if there is an abnormal situation such as reinforcing steel bars, pipelines, windows, air conditioners and the like interfering with the air-dry detection at the position, it is considered that there is no air-dry defect at the position and the position is ignored.

[0139] The marking mechanism includes a paint storage tank and a suction pump.

[0140] The paint storage tank and the suction pump are fixedly arranged on the unmanned aerial vehicle, the input end of the suction pump is in communication with the paint storage tank, and the output end of the suction pump is connected with a spray head through a pipeline. A flow control valve is arranged on the pipeline. The paint storage tank is filled with red or yellow paint, which is easy to find after being sprayed on the wall, and the defect position can be quickly found.

[0141] The flight route planning mechanism: numbers the walls to be detected, collects the position coordinate information of each wall, collects the height and width information of the wall, reasonably plans the flight route of the unmanned aerial vehicle according to the position coordinates of each wall, makes the total flight distance of the unmanned aerial vehicle shortest, plans the flight route and height of the unmanned aerial vehicle for each wall according to the monitoring range of the infrared thermal imager, the flight height of the unmanned aerial vehicle is from low to high, and the unmanned aerial vehicle flies in an S-shaped route, there is a repeated part of the detection area in the adjacent two round trips, so that the monitoring range of the infrared thermal imager can cover the entire wall without omission.

[0142] The working principle of the building wall hollow defect detection system is as follows: first, the wall to be detected is numbered, and the position coordinate information of each wall is collected; the height and width information of the wall is collected; the flight route of the unmanned aerial vehicle is reasonably planned by the route planning mechanism according to the position coordinates of each wall; the total flight distance of the unmanned aerial vehicle is the shortest; according to the monitoring range of the infrared thermal imager, the flight route and height of the unmanned aerial vehicle for each wall are planned, the flight height of the unmanned aerial vehicle is from low to high, and the unmanned aerial vehicle flies along the S-shaped route, and there is a repeated part in the detection area in the adjacent two round trips; so that the monitoring range of the infrared thermal imager can cover the entire wall without omission; an air environment data sensor is arranged at each wall for detecting real-time environmental data of each sub-area; whether the wall hollow detection can be performed by the unmanned aerial vehicle and the thermal imager is determined according to the real-time environmental data, and if the detection can be performed, the unmanned aerial vehicle model and camera parameters that can be used in the flight environment are determined; the wall is detected for hollow by the unmanned aerial vehicle and the thermal imager according to the planned flight route, and the infrared image of the wall surface is obtained; the collected infrared image is analyzed by the hollow identification mechanism to identify the hollow defect on the wall; when the hollow defect on the wall is detected, the position coordinate information, size information and height information of the hollow defect are recorded; the abnormal detection mechanism checks the abnormality at the position of the detected hollow defect, and checks whether there is an abnormal situation affecting the detection result at the position of the wall hollow defect; if there is an abnormal situation such as steel bars, pipelines, windows, air conditioners and the like interfering with the hollow detection at the position, it is considered that there is no hollow defect at the position, and the position is ignored; it is confirmed that there is no abnormal situation at the position of the wall hollow defect to be detected; the marking mechanism sprays marks on the wall at the position, so that the defect position can be quickly found subsequently.

[0143] The route planning mechanism reasonably plans the flight route of the unmanned aerial vehicle according to the position coordinates of each wall; the total flight distance of the unmanned aerial vehicle is the shortest; according to the monitoring range of the infrared thermal imager, the flight route and height of the unmanned aerial vehicle for each wall are planned, the unmanned aerial vehicle flies along the S-shaped route, and the monitoring range of the infrared thermal imager can cover the entire wall without omission. Whether the wall hollow detection can be performed by the unmanned aerial vehicle and the thermal imager can be determined according to the real-time environmental data, and if the detection can be performed, the unmanned aerial vehicle model and camera parameters that can be used in the flight environment are determined, so as to ensure that the wall is detected for hollow under suitable flight detection conditions; the wall is detected for hollow by the unmanned aerial vehicle and the thermal imager, the collected infrared image is analyzed by the hollow identification mechanism, and the hollow defect on the wall is identified in time; when the hollow defect on the wall is detected, the position coordinate information, size information and height information of the hollow defect are recorded; the marking mechanism sprays marks on the wall at the position, so that the defect position can be quickly found subsequently.

[0144] The above-described embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those skilled in the art that the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced by equivalent replacements; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for detecting hollow defects in building walls, characterized in that, Includes the following steps: S1. Number the walls to be inspected and collect the position coordinates of each wall; collect the height and width information of the walls; S2. The flight path planning agency plans the drone's flight path reasonably based on the position coordinates of each wall, so as to minimize the total flight distance of the drone. S3. Based on the monitoring range of the infrared thermal imager, plan the flight path and altitude of the drone for each wall; so that the monitoring range of the infrared thermal imager can cover the entire wall without any omissions. S4. Install air environment data sensors at each wall to detect real-time environmental data for each sub-area; S5. Determine whether wall hollowing can be detected by drones and thermal imagers based on real-time environmental data. If detection is possible, determine the drone model and camera parameters that can be used in the flight environment. S6. Based on the planned flight route, use drones and thermal imagers to detect hollow areas in the wall and obtain infrared images of the wall surface. S7. The hollow wall identification mechanism analyzes the acquired infrared images to identify hollow wall defects; and records the location coordinates and dimensions of the hollow wall defects. S8. The anomaly detection agency will conduct anomaly investigation at the detected hollow defect locations to check whether there are any abnormalities at the hollow defect locations that may affect the detection results; if there are abnormalities at this location that interfere with the hollow defect detection, it will be considered that there are no hollow defects at this location and this location will be ignored. S9. Confirm that there are no abnormalities at the location of the hollow wall defect in the inspection area; use a marking agency to spray a mark on the wall at this location to facilitate the quick location of the defect later.

2. The method for detecting hollow defects in building walls according to claim 1, characterized in that: Step S2 uses the three-dimensional A* algorithm for trajectory planning.

3. The method for detecting hollow defects in building walls according to claim 1, characterized in that: The drone is equipped with an infrared obstacle avoidance sensor. During the drone's flight, the infrared obstacle avoidance sensor can monitor obstacles along the flight path, enabling the drone to avoid obstacles in time and prevent collisions.

4. The method for detecting hollow defects in building walls according to claim 1, characterized in that: The anomaly detection mechanism includes: Data collection module: Collects a large number of images of steel bars, pipes, windows and air conditioning equipment on the wall, and annotates the images as reference samples; Image acquisition module: includes a high-definition camera and LED, which acquires images of the wall to be inspected for hollowness; Image preprocessing module: preprocesses the acquired wall image, including noise reduction, contrast enhancement, and color correction; Feature extraction module: Extracts features from the preprocessed image, using edge detection methods to extract features related to steel bars, pipes, windows, and air conditioning equipment; including color, texture, and shape features; Anomaly detection module: Compares and analyzes the feature-extracted image with the reference sample to promptly identify whether there are any anomalies at the wall hollow detection point.

5. The method for detecting hollow defects in building walls according to claim 1, characterized in that: The real-time environmental data includes: temperature, humidity, wind speed, particulate matter concentration, light intensity, and whether it is sunny, rainy, or snowy.

6. The method for detecting hollow defects in building walls according to claim 1, characterized in that: The hollow drum identification mechanism includes: Infrared image acquisition unit: includes an infrared thermal imager, which scans the building wall to acquire infrared images of the wall; Infrared image feature extraction unit: performs feature extraction on the preprocessed infrared image, including temperature anomaly areas, hot spots or cold spots; Hollow Area Recognition Unit: Based on the extracted features, it identifies the hollow areas; Hollow Area Outline Size Identification Unit: Based on the feature-extracted infrared image, the edge detection method is used to extract the edge outline of the hollow wall defect; the size of the hollow wall defect is calculated and corresponding one-to-one with the height coordinates of the hollow defect location, and stored in the information storage unit; Information storage unit: Used to store information during the detection process of hollow wall defects, including the drone flight path, real-time environmental data, collected images and infrared images, and detection results.

7. The method for detecting hollow defects in building walls according to claim 6, characterized in that: The hollow bulge contour size marking unit specifically includes the following: A. Edge detection: A1. Gradient Calculation: Gradient calculation is performed on the preprocessed infrared image; by calculating the gradient intensity and direction of each pixel in the image, preliminary edge information can be obtained. A2. Edge Enhancement: To display edges more clearly, the Sobel filter is used to filter the gradient image to enhance edge features and reduce noise interference. A3. Thresholding: Set an appropriate threshold to binarize the enhanced edge image so that the edge contour of the hollow defect will be presented in a clear black and white contrast. B. Contour extraction: B1. Contour Tracking: In the binarized image, the edge contour of the hollow defect is extracted by the contour tracking algorithm. The algorithm can traverse the pixels in the image and identify the continuous sequence of boundary points. B2. Contour Optimization: The extracted contours are smoothed to eliminate contour jitter or breakage caused by noise or poor image quality. C. Dimension Calculation: C1. Contour Fitting: Fitting the extracted hollow contour using appropriate geometry; C2. Size Calculation: By calculating the parameters of the fitted shape, the size of the hollow defect is obtained. The location and size of the hollow are used to objectively and accurately assess the hollow in the wall and its range, which serves as a strong basis for the inspection of the exterior wall finish.

8. The method for detecting hollow defects in building walls according to claim 1, characterized in that: The drone model information includes model, weight, flight time, payload, flight speed, and operating environment range. The camera parameter information includes camera model, resolution, focal length, and aperture.

9. A building wall hollow defect detection system using the method of claim 1, comprising: Unmanned aerial vehicles (UAVs), infrared thermal imagers, high-definition cameras, UAV altimeters, marking mechanisms, void detection mechanisms, anomaly detection mechanisms, and flight path planning mechanisms; characterized by: The drone is equipped with an infrared thermal imager and a high-definition camera at its front end. The infrared thermal imager captures infrared images of the wall, and the high-definition camera captures images of the wall. The drone is equipped with a drone altimeter, which is used to monitor and record the drone's flight altitude; Hollow Area Detection Mechanism: Analyzes the acquired infrared images to identify hollow areas on the wall; when a hollow area defect is detected, it records the location coordinates, dimensions, and height information of the hollow area defect. Anomaly detection agency: Conduct anomaly investigation at the location of detected hollow defects to check whether there are any abnormalities at the location of hollow defects in the wall that may affect the test results; The labeling mechanism includes a paint storage tank and a suction pump; the paint storage tank and suction pump are fixedly installed on the drone. The input end of the suction pump is connected to the paint storage tank, and the output end of the suction pump is connected to the nozzle through a pipe. Flight path planning: Number the walls to be inspected, collect the position coordinates of each wall; collect the height and width information of the walls; plan the flight path of the drones reasonably according to the position coordinates of each wall; minimize the total flight distance of the drones; according to the monitoring range of the infrared thermal imager, plan the flight path and altitude of the drones for each wall, with the drones flying from low to high altitude in an S-shaped route, with overlapping detection areas in two adjacent round trips; thus, the monitoring range of the infrared thermal imager can cover the entire wall.

10. The building wall hollow defect detection system according to claim 9, characterized in that: The pipeline is equipped with a flow control valve, and the paint storage tank contains red or yellow paint.

Images