Method for detecting quality of functional layer of outer wall of building by unmanned aerial vehicle

Abstract

The invention discloses a method for detecting the quality of a functional layer of a building outer wall by an unmanned aerial vehicle, and particularly relates to the technical field of building outer wall detection.The method comprises the steps that firstly, an infrared image of a building outer wall facing object is collected through the unmanned aerial vehicle, pixel points and remaining pixel points of a serious hollowing area are determined, and the abnormal degree of the remaining pixel points is calculated; clustering is carried out by adopting a defect probability-based region growing method; based on seed point screening of morphological preprocessing, structural elements of different sizes are comprehensively determined to be used through weighted summation calculation according to the jitter frequency of the unmanned aerial vehicle and the image resolution, and noise is filtered step by step. Generating a multi-scale pyramid for the original image, and detecting candidate seed points at each level; sorting and preferentially selecting the candidate seed points based on the morphological closure degree and the shape regularity of each candidate seed point; pixel points in the neighborhood of the growth seed points are merged to obtain all hollow defect connected domains; and judging the quality condition of the building outer wall functional layer according to the area of the hollowing defect connected domain.

Description

Technical Field

[0001] The present invention relates to the technical field of building exterior wall detection, and more specifically, to a method for detecting the quality of a functional layer of a building exterior wall using an unmanned aerial vehicle (UAV). Background Art

[0002] Facing bricks are used for exterior building decoration because they are more stain-resistant than paint. Traditionally, exterior wall facing bricks are bonded directly to the wall surface using adhesive mortar. However, due to environmental factors and poor construction quality, the tiles often become detached from the exterior wall. Over time, under the influence of rain, strong winds, and earthquakes, the detachment area gradually increases, eventually causing the tiles to peel off and fall from the main structure, resulting in safety accidents. Therefore, hollowing and leak detection of building exterior walls is essential.

[0003] Currently, the main methods for testing the quality of exterior wall finishing layers include visual inspection, hammering, pull-off, and infrared technology. Visual inspection relies on the subjective experience of the inspector to determine the quality of the finish; hammering uses the frequency emitted by the struck surface to determine internal defects in the finish; the pull-off method uses sampling to inspect exterior walls for damage; and infrared technology analyzes the internal condition of an object by detecting the surface temperature distribution.

[0004] In the prior art, if the seed point is located in a noisy area, it may cause the spread of erroneous connected domains. The selection of the initial seed point directly affects the generation result of the connected domain and the generation of the final result. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a method for detecting the quality of the functional layer of the exterior wall of a building using a drone to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions: The method for detecting the quality of the functional layer of the building exterior wall by using a drone includes the following steps: First, the drone collects infrared images of building exterior wall finishes, identifies pixels in areas with severe hollowing and the remaining pixels, and calculates the degree of abnormality of the remaining pixels. Clustering is performed using a region growing method based on defect probability. Based on morphological preprocessing, seed points are screened and noise is filtered step by step using structural elements of different sizes to select stable candidate regions. The specific size of the structural element is determined by weighted summation based on the drone's vibration frequency and image resolution. Generate a multi-scale pyramid for the original image and detect candidate seed points at each level; retain only pixels that are determined to be high-probability hollow areas at multiple scales as seed points; sort and prioritize candidate seed points based on their morphological closure and shape regularity; and select the optimal seed point; The pixels in the neighborhood of the growth seed point are merged to obtain the connected domain of all hollowing defects.

[0007] In a preferred embodiment, the seed point screening based on morphological preprocessing eliminates noise interference through morphological operations. First, an opening operation is performed to remove isolated noise points and small interference areas, retaining the main hollow area; then a closing operation is performed to fill the holes inside the hollow area to enhance connectivity.

[0008] In a preferred embodiment, the structure elements of different sizes are used to filter noise step by step to screen out stable candidate areas, which include candidate seed points. The specific size of the structure elements is comprehensively determined by weighted summation calculation based on the drone jitter frequency and image resolution.

[0009] In a preferred embodiment, a multi-scale pyramid is generated for the original image, and candidate seed points are detected at each level; only pixel points that are determined to be high-probability hollow areas at multiple scales are retained as seed points; and the candidate seed points are sorted and optimized based on the morphological closure and shape regularity of each candidate seed point.

[0010] In a preferred embodiment, the morphological closure and shape regularity of each candidate seed point are determined, the priority coefficient of each candidate seed point is calculated by weighted summation, and the seed point is selected according to the priority coefficient value of the seed point, and the one with the largest value is selected as the seed point.

[0011] In a preferred embodiment, the morphological closure is the core indicator for quantifying the effect of the closing operation, and is used to evaluate the integrity of the target area after preprocessing and the quality of internal structure repair; its calculation needs to be combined with the changes in image features before and after the closing operation, and mainly includes the following two key dimensions: area closure rate and hole filling completeness index; area closure rate: determined by the area of the region after the opening operation and the area of the region after the closing operation; a positive value indicates that the closing operation fills the boundary loss caused by the opening operation, and a negative value indicates the risk of over-corrosion; the hole filling completeness index is determined by the number of holes inside the region before the closing operation and the number of remaining holes after the closing operation.

[0012] In a preferred embodiment, the shape regularity converts the geometric characteristics of the shape into comparable values through mathematical indicators.

[0013] In a preferred embodiment, a seed point is determined, and a search is performed in the neighborhood of the selected seed point. Pixels in the neighborhood that have a probability of defect are retained and merged into one region. This region is then used as a new growth seed point, and a search is performed again in its neighborhood. Pixels in the neighborhood that have a probability of defect are retained, and a new region is updated to obtain a new seed point region. This process is repeated multiple times until the neighborhood contains no pixels with a probability of defect; at this point, a connected region of hollowing defects is obtained. By screening out the marked pixels, the defect probability of the non-standard pixels is calculated with the marked pixels as a reference, and the hollowing defect connected domain is obtained; then the area of the hollowing defect connected domain is obtained.

[0014] In a preferred embodiment, the area of the connected domain of the hollowing defect is determined. For the hollowing defect, the more serious the hollowing defect is, the larger the area occupied by the hollowing is, and the worse the quality of the corresponding building exterior wall is.

[0015] In a preferred embodiment, the drone is an industrial-grade drone with strong stability, equipped with an obstacle avoidance system and an RTK positioning module to ensure hovering accuracy.

[0016] Technical effects and advantages of the present invention: The present invention first uses a drone to collect infrared images of building exterior wall finishes, determines the pixels in the area with severe hollowing and the remaining pixels, and calculates the abnormality degree of the remaining pixels; clustering is performed using a region growing method based on defect probability; seed point screening based on morphological preprocessing uses structural elements of different sizes to filter noise step by step to screen out stable candidate areas. The specific size of the structural element is comprehensively determined by weighted summation calculation based on the drone jitter frequency and image resolution.

[0017] A multi-scale pyramid is generated for the original image, and candidate seed points are detected at each level; only pixel points that are judged to be high-probability hollowing areas at multiple scales are retained as seed points; the candidate seed points are sorted and optimized based on their morphological closure and shape regularity; the optimal seed points are selected; the selected seed points are more in line with actual requirements, and at the same time, the seed points are avoided from being located in noise areas, avoiding the errors in the generation of connected domains caused by the selection of seed points in noise areas in subsequent operations; then the pixels in the neighborhood of the growing seed points are merged to obtain the connected domains of all hollowing defects; the quality of the functional layer of the building's exterior wall is judged according to the area of the connected domain of the hollowing defects. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings; Figure 1 This is a flow chart of a method for detecting the quality of a functional layer of a building's exterior wall using a drone according to the present invention; Figure 2 Schematic diagram of the flow of the region growing algorithm of the present invention. DETAILED DESCRIPTION

[0019] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. 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 making creative efforts are within the scope of protection of the present invention.

[0020] The present invention first uses a drone to collect infrared images of building exterior wall finishes, determines the pixels in the area with severe hollowing and the remaining pixels, and calculates the abnormality degree of the remaining pixels; clustering is performed using a region growing method based on defect probability; seed point screening based on morphological preprocessing uses structural elements of different sizes to filter noise step by step to screen out stable candidate areas. The specific size of the structural element is comprehensively determined by weighted summation calculation based on the drone jitter frequency and image resolution.

[0021] A multi-scale pyramid is generated for the original image, and candidate seed points are detected at each level. Only pixels that are judged to be in high-probability hollowing areas at multiple scales are retained as seed points. The candidate seed points are sorted and optimized based on their morphological closure and shape regularity. The optimal seed point is selected. The pixels in the neighborhood of the growing seed point are merged to obtain the connected domain of all hollowing defects. The quality of the functional layer of the building's exterior wall is judged based on the area of the connected domain of hollowing defects.

[0022] In the embodiment, the method for detecting the quality of the functional layer of the exterior wall of a building by using a drone of the present invention is as follows: Figure 1 As shown, the following steps are included: First, infrared images of building exterior wall finishes were collected using drones to identify pixels in areas with severe hollowing and the remaining pixels, and the degree of abnormality of the remaining pixels was calculated. Clustering was performed using a region growing method based on defect probability. Based on morphological preprocessing, seed points are screened and noise is filtered step by step using structural elements of different sizes to select stable candidate regions. The specific size of the structural element is determined by weighted summation based on the drone's vibration frequency and image resolution. Generate a multi-scale pyramid for the original image and detect candidate seed points at each level; retain only pixels that are determined to be high-probability hollow areas at multiple scales as seed points; sort and prioritize candidate seed points based on their morphological closure and shape regularity; and select the optimal seed point; Merge the pixels in the neighborhood of the growth seed point to obtain the connected domain of all hollow defects; The stable candidate region includes candidate seed points.

[0023] specific; Use the camera carried by the drone to collect infrared images of the building's exterior wall finishes; the drone camera uses a high-resolution thermal imaging camera (such as the FLIR series) to ensure that the thermal sensitivity (NETD ≤ 50mk) and spatial resolution meet the requirements of building defect detection; at the same time, calibrate the infrared sensor to eliminate errors caused by ambient temperature fluctuations, and use a blackbody calibration tool if necessary.

[0024] The drone should be an industrial-grade drone with strong stability, equipped with an obstacle avoidance system and RTK positioning module to ensure hovering accuracy; check the gimbal compatibility to avoid image blur due to shaking of the infrared camera.

[0025] Furthermore, before the drone takes off, use 3D modeling software (such as Pix4Dcapture) to preset the route and set the image overlap rate to more than 70% to facilitate later stitching; keep the drone 5-15 meters away from the wall, and shoot at a vertical or tilted angle (30°-60°) to cover the details of the entire facade; the drone must comply with local airspace regulations when taking off and avoid no-fly zones; set the automatic return power threshold and equip a parachute device to prevent sudden failures; avoid rainy days, strong winds or extreme temperatures (such as high temperatures at noon), as excessive temperature differences will interfere with heat conduction characteristics.

[0026] The best time to shoot is before sunrise or within 2 hours after sunset. At this time, the surface temperature distribution of the building is stable, making it easy to capture abnormal areas such as water seepage and hollowing. If conditions permit, artificial heat sources such as air conditioners and lighting equipment around the building can be turned off. At the same time, mark and avoid surfaces directly exposed to the sun or reflective materials (such as glass curtain walls) to reduce reflected heat noise.

[0027] During the construction and use of exterior wall finishes, defects such as hollowing (in mortar finishes) or debonding (in brick finishes) can easily form between the wall structure and the finish. These defects can lead to the formation of air pockets, which reduces the thermal conductivity of the wall. Under solar radiation, the surface temperature of defective areas differs from that of intact areas. Infrared thermal imaging technology can effectively detect these defects.

[0028] Infrared thermal imagers don't directly measure an object's temperature. Instead, they use infrared detectors to receive radiation from the object being inspected. This radiation includes the object's own radiation, radiation reflected from surrounding objects, and radiation from the atmosphere. As infrared radiation propagates through the atmosphere, it is absorbed and scattered by certain gases and particles, causing the radiation energy to attenuate. Consequently, the captured infrared images often suffer from low contrast, poor signal-to-noise ratio, and image quality, which affects the accuracy of defect identification and leads to inaccurate defect location and detection. Therefore, by processing the image's grayscale information and extracting target pixels, and combining the distribution patterns of target pixels, the relationships between neighboring pixels, and the gradient direction, we can more accurately locate defect areas and thus obtain a more precise estimate of the defect area and severity.

[0029] Obtain the target pixel point; for the building's exterior wall finishing layer, hollowing defects will cause an air layer to appear inside the finishing layer. Since the thermal resistance of air is greater than the thermal resistance of the wall material, the heat transfer at the location of the hollowing defect is reduced, and heat accumulates on the exterior wall surface at the location of the hollowing defect, causing the exterior wall surface temperature at the hollowing location to be higher than the temperature of the exterior wall surface in the normal area, that is, the brightness corresponding to the hollowing area in the infrared thermal image grayscale is relatively large, that is, the grayscale value corresponding to the hollowing area is relatively large. Since the internal medium of other normal areas of the finishing layer is the same, the temperature of the normal areas is similar or the same, that is, the corresponding grayscale levels are similar or the same. For the finishing layer, most areas are usually normal areas. When the area of the hollowing area exceeds a certain critical point, the finishing layer will peel off and fall off from the main structure. Therefore, the normal area accounts for a large proportion in the corresponding infrared thermal image grayscale image, that is, the number of pixels with grayscale levels corresponding to the normal area is the largest.

[0030] The degree of abnormality of each remaining pixel is calculated based on the grayscale mean of each remaining pixel and its corresponding neighboring pixels. The greater the difference between a remaining pixel and the background grayscale and the greater the grayscale fluctuation in the area where the pixel is located, the greater the degree of abnormality of the pixel.

[0031] The larger the grayscale difference between the marked pixel and the background, the more likely this type of pixel is to be a pixel in a serious hollowing area. Under the influence of rain, strong winds, earthquakes, etc., the detachment area of the hollowing defect will gradually increase, that is, the serious hollowing defect will gradually spread, and eventually the slight hollowing defect will become a serious hollowing defect. The normal area at the edge of the slight hollowing area will gradually become a slight hollowing defect, until the hollowing defect reaches the critical point, causing the finishing layer to fall off. Due to the extension of the hollowing defect, the grayscale level of the hollowing defect area fluctuates, and the degree of fluctuation is larger than that of the normal area. The degree of abnormality of each remaining pixel is calculated based on the grayscale mean of each remaining pixel and the corresponding neighboring pixel. Among them, the greater the difference between a remaining pixel and the background grayscale, the greater the grayscale fluctuation of the area where it is located, and the greater the degree of abnormality of the pixel. A sliding window is performed with each remaining pixel as the center point. Each sliding window contains pixels in the severe hollowing area. The probability of each remaining pixel being a pixel in the hollowing area is calculated using the abnormality of each sliding window center point, the gradient direction of the center point, and the gradient direction of the pixels in the sliding window containing the severe hollowing area. To avoid randomness, the sliding window should contain at least two pixels in the severe hollowing area.

[0032] Since hollow defects usually exist in clusters, the region growing method based on defect probability is used for clustering. Figure 2 As shown, the steps are as follows: The selection of the initial seed point directly affects the generation result of the connected domain; if the seed point is located in a noise area (such as a local thermal radiation anomaly), it may cause the spread of erroneous connected domains and affect the generation of the final result.

[0033] Based on morphological preprocessing, seed points are screened and noise interference is eliminated through morphological operations to improve seed point reliability. First, an opening operation (erosion followed by dilation) is performed to remove isolated noise points and small interference areas, while retaining the main hollow areas. A closing operation (dilation followed by erosion) is then performed to fill holes within the hollow areas and enhance connectivity.

[0034] Using structural elements of different sizes, noise is filtered step by step to select stable candidate regions, which contain candidate seed points. The specific size of the structural element is determined by weighted summation based on the drone's vibration frequency and image resolution; the specific formula is as follows: ; Where C represents the size of the structural element; PL represents the UAV jitter frequency. The larger the UAV jitter frequency, the larger the size of the structural element should be, and vice versa; FB represents the image resolution. The larger the image resolution, the smaller the size of the structural element, and vice versa; α and β are the weight coefficients of the UAV jitter frequency and image resolution, respectively, which are set by the staff.

[0035] The jitter frequency of the drone is obtained by the drone's built-in jitter sensor; in a strong wind environment, the size magnification factor is 1.5-1.8 times based on the drone's jitter frequency; high-resolution images allow the use of small-sized structural elements; low-resolution images need to be enlarged to avoid valid features being mistakenly filtered.

[0036] Then, the seed points for multi-scale feature fusion are optimized; image features at different resolutions are combined to avoid single-scale misjudgment.

[0037] Generate a multi-scale pyramid (such as original resolution, 1 / 2, 1 / 4) for the original image, and detect candidate seed points at each level; only retain pixels that are judged to be high-probability hollow areas at multiple scales as seed points; sort and select the candidate seed points based on their morphological closure and shape regularity.

[0038] Specifically, the morphological closure and shape regularity of each candidate seed point are determined, and the priority coefficient of each candidate seed point is calculated by weighted summation. The formula is as follows: Where Y represents the priority coefficient of the candidate seed point; BH represents the morphological closure of the candidate seed point. The greater the morphological closure of the candidate seed point, the greater the priority coefficient of the candidate seed point, and vice versa. XZ represents the shape regularity of the candidate seed point. The greater the shape regularity of the candidate seed point, the greater the priority coefficient of the candidate seed point, and vice versa. γ and δ represent the weight coefficients of the morphological closure and shape regularity of the candidate seed point, respectively, and are set by the staff as needed. The seed point is selected according to the value of the priority coefficient of the seed point, and the largest value is selected as the seed point.

[0039] The morphological closure is the core indicator for quantifying the effect of the closing operation, and is used to evaluate the integrity of the target area after preprocessing and the quality of internal structure repair. Its calculation needs to be combined with the changes in image features before and after the closing operation, and mainly includes the following two key dimensions: area closure rate and hole filling completeness index; area closure rate: determined by the area of the region after the opening operation and the area of the region after the closing operation; reflects the ability of the closing operation to compensate for the expansion of the target area. A positive value indicates that the closing operation fills the boundary loss caused by the opening operation, and a negative value indicates the risk of over-corrosion. The hole filling completeness index is determined by the number of holes in the region before the closing operation and the number of remaining holes after the closing operation. The shape regularity converts the geometric features of the shape into comparable values through mathematical indicators.

[0040] A search is performed within the neighborhood of the selected seed point. Pixels within the neighborhood that have a high probability of defect are retained and merged into one region. This region is then used as a new growth seed point. A search is performed within its neighborhood again. Pixels within the neighborhood that have a high probability of defect are retained and a new region is updated to obtain a new seed point region. This process is repeated multiple times until the neighborhood contains no pixels with a high probability of defect. The connected domain of the hollowing defect is then obtained.

[0041] By screening out the marked pixels, the defect probability of the non-standard pixels is calculated with the marked pixels as a reference, and the hollowing defect connected domain is obtained; then the area of the hollowing defect connected domain is obtained.

[0042] For hollowing defects, the more severe the hollowing defect and the larger the area occupied by the hollowing, the worse the quality of the corresponding building exterior wall, and the more likely it is to cause the facing tiles to peel off or fall off from the main structure, thereby causing a safety accident. For hollowing defects of the same area, the greater the proportion of severe hollowing defects, the worse the corresponding quality.

[0043] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0044] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0045] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0046] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0047] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A method for detecting the quality of a functional layer of a building's exterior wall using a drone, characterized in that: The following steps are involved: First, the drone collects infrared images of the building's exterior wall finishes, identifies the pixels in the area with severe hollowing and the remaining pixels, and calculates the degree of abnormality of the remaining pixels. Clustering is performed using the region growing method based on defect probability; Based on morphological preprocessing, seed points are screened and noise is filtered step by step using structural elements of different sizes to select stable candidate regions. The specific size of the structural element is determined by weighted summation based on the drone's vibration frequency and image resolution. Generate a multi-scale pyramid for the original image and detect candidate seed points at each level; Only pixels that are determined to be high-probability hollow areas at multiple scales are retained as seed points; candidate seed points are sorted and selected based on their morphological closure and shape regularity; and the optimal seed point is selected; The pixels in the neighborhood of the growth seed point are merged to obtain the connected domain of all hollowing defects.

2. The method for detecting the quality of the functional layer of the exterior wall of a building by using an unmanned aerial vehicle according to claim 1, characterized in that: The seed point screening based on morphological preprocessing eliminates noise interference through morphological operations. First, an opening operation is performed to remove isolated noise points and small interference areas, and retain the main hollow area; Then a closing operation is performed to fill the holes inside the hollow area and enhance connectivity.

3. The method for detecting the quality of the functional layer of the exterior wall of a building by using an unmanned aerial vehicle according to claim 1, characterized in that: The method uses structural elements of different sizes to filter noise step by step and select stable candidate areas, which include candidate seed points. The specific size of the structural elements is comprehensively determined by weighted summation calculation based on the drone's jitter frequency and image resolution.

4. The method for detecting the quality of the functional layer of the building exterior wall by using an unmanned aerial vehicle according to claim 1, characterized in that: The method generates a multi-scale pyramid for the original image and detects candidate seed points at each level; only retains pixel points that are determined to be high-probability hollow areas at multiple scales as seed points; and sorts and selects the candidate seed points based on their morphological closure and shape regularity.

5. The method for detecting the quality of the functional layer of the building exterior wall by using an unmanned aerial vehicle according to claim 4, characterized in that: Determine the morphological closure and shape regularity of each candidate seed point, calculate the priority coefficient of each candidate seed point by weighted summation, select the seed point according to the priority coefficient value of the seed point, and select the seed point with the largest value as the seed point.

6. The method for detecting the quality of the functional layer of the building exterior wall by using an unmanned aerial vehicle according to claim 5, characterized in that: The morphological closure degree is a core indicator for quantifying the effect of the closing operation, and is used to evaluate the integrity of the target area after preprocessing and the quality of internal structure restoration. Its calculation requires consideration of the changes in image features before and after the closing operation, and mainly includes the following two key dimensions: area closure rate and hole filling completeness index. Area closure rate is determined by the area of the region after the opening operation and the area of the region after the closing operation. A positive value indicates that the closing operation has filled the boundary loss caused by the opening operation, while a negative value indicates the risk of over-corrosion. The hole filling completeness index is determined by the number of holes in the region before the closing operation and the number of holes remaining after the closing operation.

7. The method for detecting the quality of the functional layer of the building exterior wall by using an unmanned aerial vehicle according to claim 5, characterized in that: The shape regularity converts the geometric characteristics of the shape into comparable numerical values through mathematical indicators.

8. The method for detecting the quality of the functional layer of the building exterior wall by using an unmanned aerial vehicle according to claim 5, characterized in that: Determine the seed point and search within the neighborhood of the selected seed point. Pixels in the neighborhood with a high probability of defect are retained and merged into one region. This region is used as a new growth seed point and search is performed again within its neighborhood. Pixels in the neighborhood with a high probability of defect are retained and a new region is updated to obtain a new seed point region. Iterate multiple times until the neighborhood contains no pixels with a high probability of defect. At this point, the connected domain of the hollowing defect is obtained. By screening out the marked pixels, the defect probability of the non-standard pixels is calculated with the marked pixels as a reference, and the hollowing defect connected domain is obtained; then the area of the hollowing defect connected domain is obtained.

9. The method for detecting the quality of the functional layer of the building exterior wall by using an unmanned aerial vehicle according to claim 8, characterized in that: Determine the area of the connected domain of the hollowing defect. For the hollowing defect, the more serious the hollowing defect is, the larger the area occupied by the hollowing is, and the worse the quality of the corresponding building exterior wall is.

10. The method for detecting the quality of the functional layer of the building exterior wall by using an unmanned aerial vehicle according to claim 1, characterized in that: The drone is an industrial-grade drone with strong stability, equipped with an obstacle avoidance system and an RTK positioning module to ensure hovering accuracy.

Images