Method for carrying out disease detection on building facade by carrying out infrared thermal imaging on unmanned aerial vehicle

By combining UAV flight path planning and image deep learning, the background interference problem in UAV infrared thermal imaging for building facade inspection is solved, enabling rapid and accurate identification and quantitative assessment of defects, and is suitable for intelligent detection of building facade defects.

CN118914285AInactive Publication Date: 2024-11-08CHINA MCC5 GROUP CORP LTD

Patent Information

Application Number
CN202410948498.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-16
Publication Date
2024-11-08
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

In existing technologies, the use of drones equipped with infrared thermal imaging for detecting defects on building facades suffers from significant background image interference, low efficiency in manual processing, and difficulty in achieving rapid identification and quantitative assessment.

Method used

By combining UAV flight path planning with deep learning image detection, background interference is removed through image stitching and intelligent detection. Parametric flight path shooting, dual-channel image preprocessing, and deep learning models are used to identify building facade defects.

Benefits of technology

It enables rapid identification, location, and quantitative assessment of defects in infrared thermal imaging images of building facades, improving detection efficiency and accuracy, reducing false identification rates, and is suitable for long-term, regular inspections.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118914285A_ABST
    Figure CN118914285A_ABST
Patent Text Reader

Abstract

The invention discloses a method for carrying out disease detection on a building facade through unmanned aerial vehicle carried infrared thermal imaging, and the method comprises the steps: S1, obtaining the information of the building facade, and determining the shooting distance of an unmanned aerial vehicle; s2, determining a flight photographing route of the unmanned aerial vehicle; s3, controlling the unmanned aerial vehicle to complete flight shooting; s4, the images shot by the unmanned aerial vehicle are numbered and spliced; s5, visible light image rejection is carried out on the spliced image; s6, performing parameter adjustment on the visible light image; and S7, carrying out disease detection on the real temperature image through a detection module. According to the invention, two aspects of unmanned aerial vehicle flight path planning and image deep learning detection are combined, influences of background interference and external facade other factor interference in the infrared thermal imaging picture are eliminated through image splicing and intelligent detection, and rapid identification, positioning and quantitative evaluation of a disease part in the building external facade infrared thermal imaging picture are realized. The detection efficiency is improved; and the detection quality is ensured.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of building quality inspection, and in particular relates to a method for detecting defects on building facades using an unmanned aerial vehicle equipped with infrared thermal imaging. Background Art

[0002] With the advent of the era of building stock, more and more building structures are facing a lot of inspection and maintenance work as their service life increases. Damage and disease on building facades is a typical problem that is the most problematic and poses a great threat to personal safety.

[0003] Take exterior wall tiles as an example. Due to their long-term exposure to the outdoors and the coupling of multiple physical fields such as rain, temperature differences, and wind loads, a large number of exterior wall tiles have problems such as hollowing, cracks, and water seepage.

[0004] For example, in the case of an anchored exterior wall structure, cement mortar is used to bond the insulation board to the base wall, which is then fixed with rivets and finally painted on the outer layer. However, this can easily fall off over time.

[0005] Commonly used insulation materials, such as polyurethane, are prone to aging and deformation over time. Rock wool materials are prone to collapse after absorbing water due to the loose internal fiber structure. External coatings such as real stone paint are also prone to aging after years of sun and rain, and the deformation of real stone paint and insulation board is inconsistent;

[0006] At present, the standard specifications for the use of infrared thermal imaging in the field of building facade quality inspection have clear provisions on the equipment parameters of drones and infrared thermal imagers, but there are no detailed provisions on the specific implementation methods during the inspection process. As a result, in practice, the applicant has found that the shooting method has a great influence on the imaging results. If there are many interference factors in the background image (for example, the sky, windows, different depths of field, different colors, etc.), it will have a great impact on the detection and discrimination of the thermal image. At the same time, after the infrared thermal imaging image is taken, the temperature of the infrared thermal image is currently only identified one by one through manual processing. Due to the large number of test images taken, the manual processing method will lead to a serious lag in recognition efficiency and a significant increase in the difficulty of recognition (judgment of the temperature points of the image pixels). As a result, the effect of infrared thermal imaging in the detection of building facade defects is not significant, which is not conducive to its practical application and promotion. Summary of the Invention

[0007] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a method for detecting defects on building facades using an unmanned aerial vehicle (UAV) equipped with infrared thermal imaging. The method combines UAV flight path planning with image deep learning detection. Through image stitching and intelligent detection, background interference in the infrared thermal imaging image and the influence of interference from other factors on the facade are eliminated, thereby achieving rapid identification, positioning and quantitative evaluation of defective parts in infrared thermal imaging images of building facades, improving detection efficiency and ensuring detection quality.

[0008] The object of the present invention is achieved through the following technical solutions:

[0009] A method for detecting building facade defects using an unmanned aerial vehicle (UAV) equipped with infrared thermal imaging, comprising:

[0010] Step S1: Obtain building facade information and determine the shooting distance of the drone;

[0011] Step S2: Determine the flight route of the drone based on the horizontal width of the building facade, the imaging width of a single image, and the overlap ratio of two adjacent images when stitched together.

[0012] Step S3: Based on the GPS positioning system of the drone, control the drone to complete flight shooting;

[0013] Step S4: number the images taken by the drone and complete the stitching;

[0014] Step S5: removing visible light images from the stitched image;

[0015] Step S6: Segment the visible light image after step S5 again, dividing the visible light image into multiple regions based on the different color types of the building facade as the segmentation criteria, and adjust the parameters of the visible light images of different regions to obtain a true temperature distribution map;

[0016] Step S7: Segment the real temperature distribution map according to the image size in the detection module to obtain multiple real temperature images, and perform disease detection on the real temperature images through the detection module.

[0017] In one embodiment, step S1 includes:

[0018] Obtain the horizontal width W, vertical height H, and single-story height h1 of the building facade. Based on the zoom factor z of visible light and infrared thermal imaging, make the imaging height h2 of a single image satisfy H = h2 * n1, where n1 is an integer, to determine the drone's shooting distance l1.

[0019] In one embodiment, step S2 includes:

[0020] Divide the horizontal width of the building facade into n w A single image of the same width;

[0021] W=w1*n w -(n w -1)*w1*O;

[0022] O = 20% to 30%;

[0023] Where W is the horizontal width of the building facade, w1 is the imaging width of a single image, and n w is the number of images divided on the horizontal width of the building facade, O is the overlap ratio of two adjacent images;

[0024] The first shooting coordinates of the drone are established as The coordinate origin is located at the leftmost end of the positive and negative zero positions on the building facade.

[0025] In one embodiment, in step S2, determining the flight route of the drone for taking photos includes: starting from the first shooting coordinate and taking photos along a serpentine reciprocating path from top to bottom;

[0026] The drone starts shooting from the starting point and shoots n times in the horizontal direction according to the flight distance w1. w After taking the image, take the height h2 as the descent height, and return to take n images in the horizontal direction according to the flight distance w1. w images until the entire building facade is captured.

[0027] In one embodiment, step S4 includes:

[0028] Step S401: Name the captured visible light images in the order of 1 to n;

[0029] Step S402: stitching the visible light images on the same shooting path to obtain multiple rows of stitched images;

[0030] Step S403: splicing two adjacent rows of visible light images.

[0031] In one embodiment, in step S402, the method further includes:

[0032] Perform grayscale processing on visible light images and use feature detection algorithms to obtain at least 20 feature points in each image;

[0033] Taking the first image on the same shooting path as the reference origin, all feature points are paired, and the first 10 feature points with the smallest pairing error are selected. The transformation matching matrix of the left and right adjacent images is calculated, and the stitching of multiple images on the same shooting path is completed in sequence.

[0034] In one embodiment, in step S403, the method further includes:

[0035] Feature extraction is performed on each row of spliced ​​images, and they are spliced ​​from top to bottom based on the pairing effect to form a complete image of the building facade.

[0036] In one embodiment, step S5 includes:

[0037] Scale the full image of the building facade to match the image detection size of the deep learning model;

[0038] Use deep learning models to select and remove interference factors from the complete image of the building facade;

[0039] The deep learning model is trained using at least 2,000 infrared thermal imaging images of building facade defects.

[0040] In one embodiment, in step S7, after the real temperature distribution image is segmented according to the image size in the detection module, the disease area in each real temperature image is calculated as follows:

[0041]

[0042] Among them, A 病害像素 is the pixel area of ​​the disease marked by the detection module, f is the focal length of the UAV camera, and l1 is the shooting distance of the UAV.

[0043] In one embodiment, step S3 includes:

[0044] The initial position coordinates of the drone and the coordinates of each shooting position during the movement along the shooting path are converted into GPS data and transmitted to the GPS positioning system of the drone to control the drone to complete flight shooting.

[0045] The beneficial effects of the present invention are:

[0046] (1) By integrating the three factors of drone path planning, infrared thermal imaging characteristics, and disease identification, detection and evaluation, the application of infrared thermal imaging to detect building facade defects has become more practical by introducing parameterized route shooting instructions, dual-channel image preprocessing methods, and new technologies of artificial intelligence detection, thereby improving the accuracy and intelligence of building facade defect identification.

[0047] (2) The determined drone path can effectively avoid the problem of other background images entering the building facade during the shooting process, which leads to a large temperature difference range of infrared thermal imaging and is not convenient for the later disease identification. The digital route planning design method is universal for all building shooting. The designed route plan can be saved as an electronic file to facilitate the comparative analysis of building facade diseases during long-term regular inspections;

[0048] (3) The image stitching and elimination processing method solves the problem that interference factors in traditional single photos are difficult to identify due to local appearance. That is, by adjusting the conversion matrix to deduct the presence of interference factors in infrared thermal imaging, the problem of difficulty in distinguishing interference factors from defects in infrared thermal imaging is made up, and the problem of misidentification caused by the presence of interference factors when the deep learning model detects defects on the building facade is reduced.

[0049] (4) The method of using the secondary image segmentation processing method and adjusting the parameters later to obtain the real infrared temperature solves the inconvenience of repeatedly adjusting the material reflectivity parameters during the infrared thermal imaging image shooting process and the problem that the infrared thermal imaging temperature does not match the actual temperature due to the difference in material reflectivity. Under the principle of judging the disease based on the temperature difference, the disease can be identified more accurately.

[0050] (5) A deep learning model is used to remove interference from the image, and the infrared thermal imaging defects are identified, located and classified in the real temperature image, and all pixels at the defect location are marked to obtain the defect information of the building facade. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The present invention will be described in more detail below based on embodiments and with reference to the accompanying drawings, wherein:

[0052] Figure 1 A schematic diagram showing the shooting path of the drone of the present invention is shown;

[0053] Figure 2 A schematic diagram of stitching visible light images on the same shooting path of the present invention is shown;

[0054] In the drawings, like reference numerals are used for like parts, but the drawings are not necessarily true to scale. DETAILED DESCRIPTION

[0055] The present invention will be further described below with reference to the accompanying drawings.

[0056] The present invention provides a method for detecting defects on building facades using an unmanned aerial vehicle (UAV) equipped with infrared thermal imaging, comprising:

[0057] Step S1: Obtain building facade information and determine the shooting distance of the drone;

[0058] Obtain the horizontal width W, vertical height H, and single-story height h1 of the building facade. Based on the zoom factor z of visible light and infrared thermal imaging, make the imaging height h2 of a single image satisfy H = h2 * n1, where n1 is an integer, to determine the drone's shooting distance l1.

[0059] It should be noted that in step S1, the imaging height h2 of a single image is determined by the zoom factor z of visible light and infrared thermal imaging and the shooting distance l1 of the drone. At the same time, when H=h2*n1 is satisfied, n1 is preferably set to an integer multiple of the total number of floors of the building, so as to adjust the zoom factor z of visible light and infrared thermal imaging and determine the shooting distance l1 of the drone.

[0060] Step S2: Determine the flight route of the drone based on the horizontal width of the building facade, the imaging width of a single image, and the overlap ratio of two adjacent images when stitched together.

[0061] like Figure 1 As shown, the horizontal width of the building facade is divided into n w A single image of the same width;

[0062] W=w1*n w -(n w -1)*w1*O;

[0063] O = 20% to 30%;

[0064] Where W is the horizontal width of the building facade, w1 is the imaging width of a single image, and n w is the number of images divided on the horizontal width of the building facade, O is the overlap ratio of two adjacent images;

[0065] The first shooting coordinates of the drone are established as The coordinate origin is located at the leftmost end of the positive and negative zero positions on the building facade;

[0066] It should be noted that in step S2, the horizontal width W of the building facade is used as a reference and divided into n w images, the width of a single image is w1, in n w Take an integer and satisfy W=w1*n w -(n w -1)*w1*O, so as to get the width of a single image as w1 and n w , in order to obtain the distance the drone moves each time, such as Figure 1 As shown, after determining the starting point of the drone and the unique distance for each shooting, the flight route of the drone can be determined. First, the first shooting coordinate of the drone is established as The coordinate origin is located at the leftmost end of the positive and negative zero position of the building facade. Shoot along the serpentine reciprocating path from top to bottom, and shoot n times in the horizontal direction according to the flight interval w1. w After taking the image, take the height h2 as the descent height, and return to take n images in the horizontal direction according to the flight distance w1. w images until the entire building facade is captured;

[0067] The height of each descent of the drone is h2, that is, the number of rows of shooting completed by the drone is n. h =H / h2, when n h If it is not an integer, it will be rounded up, that is, when the drone moves to the bottom row, it will still maintain the same descent height h2 for shooting;

[0068] Step S3: Based on the GPS positioning system of the drone, control the drone to complete flight shooting;

[0069] That is, the initial position coordinates of the drone obtained in step S2 and the coordinates of each shooting position during the movement along the shooting path are converted into GPS data and transmitted to the GPS positioning system of the drone to control the drone to complete the flight shooting;

[0070] Step S4: number the images taken by the drone and complete the stitching;

[0071] Specifically, it includes:

[0072] Step S401: Name the captured visible light images in the order of 1 to n;

[0073] Step S402: stitching the visible light images on the same shooting path to obtain multiple rows of stitched images;

[0074] Perform grayscale processing on visible light images and use feature detection algorithms to obtain at least 20 feature points in each image;

[0075] Taking the first image on the same shooting path as the reference origin, all feature points are paired, and the first 10 feature points with the smallest pairing error are selected. The transformation matching matrix of the left and right adjacent images is calculated. During the stitching process, the visible light images on the same shooting path are stitched sequentially.

[0076] Step S403: splicing two adjacent rows of visible light images;

[0077] Perform feature extraction on each row of stitched images, and stitch them together from top to bottom based on the matching results to form a complete image of the building facade.

[0078] It should be noted that in step S4, the visible light images taken by the drone are processed. First, the visible light images on each line of the shooting path are spliced ​​according to the shooting order. Based on the shooting method in step S2, the visible light images taken are overlapped in pairs to obtain a complete visible light image. Figure 2 As shown, after the visible light images with serial numbers 1 and 2 are stitched together, the visible light images with serial numbers 3, ..., and n are stitched together in sequence. During the stitching process, at least 20 feature points are obtained from each image, all matching points are paired, and the first 10 feature points with the smallest pairing errors are selected. The conversion matching matrix of the left and right adjacent images is calculated. After each row of visible light images is stitched together, the adjacent two rows of visible light images are stitched together in the vertical direction. During the stitching process, the feature points of each row of visible light images are also obtained, and all matching points are paired. This improves the stitching efficiency and accuracy of the visible light images during automatic stitching, so as to obtain a complete visible light image of the building facade. In this process, the existence of interference factors in infrared thermal imaging is deducted by adjusting the conversion matrix, thereby compensating for the problem that interference factors and defects in infrared thermal imaging are difficult to distinguish.

[0079] Step S5: removing visible light images from the stitched image;

[0080] Scale the full image of the building facade to match the image detection size of the deep learning model;

[0081] Use deep learning models to select and remove interference factors from the complete image of the building facade;

[0082] The deep learning model is trained using at least 2,000 infrared thermal imaging images of building facade defects.

[0083] It should be noted that in step S5, the deep learning model is used to select and remove interference factors such as windows, pipes, and blinds from the complete visible light image of the building facade obtained in step S4. Since interference factors may appear locally in a single photo during the shooting process, after the stitching in step S4 and the compression in step S5, the interference factors are presented in a complete form. The use of the deep learning model greatly improves the recognition accuracy of the interference factors and avoids the misjudgment of building facade defects due to interference factors.

[0084] Step S6: Segment the visible light image after step S5 again, dividing the visible light image into multiple regions based on the different color types of the building facade as the segmentation criteria, and adjust the parameters of the visible light images of different regions to obtain a true temperature distribution map;

[0085] It should be noted that in step S6, the visible light image after eliminating interference factors is segmented again. Since the shooting process is always performed with a single material reflectance λ, but the actual shooting object has a variety of objects with different material reflectances λ, after the visible light image is divided into multiple regions using color type as the segmentation standard, the image is adjusted with the corresponding material reflectance λ to obtain a true temperature distribution map. This solves the inconvenience of repeatedly adjusting the material reflectance parameters during the infrared thermal imaging process and the problem that the infrared thermal imaging temperature does not match the actual temperature due to the difference in material reflectance. That is, under the principle of determining the disease based on the temperature difference, the disease can be identified more accurately.

[0086] Step S7: Segment the real temperature distribution map according to the image size in the detection module to obtain multiple real temperature images, and perform disease detection on the real temperature images through the detection module;

[0087] Specifically, in step S7, the disease area in each real temperature image is calculated as follows:

[0088]

[0089] Among them, A 病害像素 is the pixel area of ​​the disease marked by the detection module, f is the focal length of the UAV camera, and l1 is the shooting distance of the UAV;

[0090] That is, the deep learning model is used again to identify the true temperature distribution map after adjusting the material reflectivity λ, and then identify, locate and classify it. During the recognition process, the true temperature distribution map is first divided into multiple visible light images according to the recognition size of the deep learning model. All pixels at the defect location in each visible light image are labeled based on semantic segmentation. Combined with the focal length and shooting distance of the drone camera, the defect area of ​​the visible light image can be obtained, and then accurate defect information of the building facade can be obtained;

[0091] That is, by comprehensively utilizing the three factors of drone flight path planning, the advantages and disadvantages of infrared thermal imaging, and disease identification, detection and evaluation, and by introducing parameterized flight route shooting instructions, dual-channel image preprocessing methods, and artificial intelligence detection technology, the application of infrared thermal imaging to detect building facade defects is more practical, which is conducive to its practical application and promotion, and effectively improves the accuracy and intelligence of building facade defect identification;

[0092] In the description of the present invention, it should be understood that the terms "upper", "lower", "bottom", "top", "front", "back", "inside", "outside", "left", "right", etc., indicating directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as limiting the present invention.

[0093] Although the present invention is described herein with reference to specific embodiments, it should be understood that these embodiments are merely illustrative of the principles and applications of the invention. It should be understood that many modifications may be made to the illustrative embodiments, and that other arrangements may be devised, without departing from the spirit and scope of the invention as defined by the appended claims. It should be understood that the various dependent claims and features described herein may be combined in ways other than those described in the original claims. It should also be understood that features described in conjunction with individual embodiments may be employed in conjunction with other described embodiments.

Claims

1. A method for detecting defects on building facades using an unmanned aerial vehicle equipped with infrared thermal imaging, characterized in that: include: Step S1, obtaining building facade information and determining the shooting distance of the drone; Step S2, determining the flight route of the drone according to the horizontal width of the building facade, the imaging width of a single image, and the overlap rate when two adjacent images are stitched together; Step S3: Based on the GPS positioning system of the drone, control the drone to complete flight shooting; Step S4: number the images taken by the drone and complete stitching; Step S5, removing visible light images from the stitched images; Step S6, dividing the visible light image after completing step S5 again, dividing the visible light image into multiple regions based on different color types of the building facade as the segmentation standard, adjusting parameters of the visible light images of different regions, and obtaining a true temperature distribution map; Step S7: segment the real temperature distribution map according to the image size in the detection module to obtain multiple real temperature images, and perform disease detection on the real temperature images through the detection module.

2. The method for detecting defects on building facades by using an unmanned aerial vehicle equipped with infrared thermal imaging according to claim 1, characterized in that: In step S1, it includes: The horizontal width W, vertical height H and height h1 of the building facade are obtained. According to the zoom factor z of visible light and infrared thermal imaging, the imaging height h2 of a single image satisfies H=h2*n1, where n1 is an integer, to determine the shooting distance l1 of the drone.

3. The method for detecting defects on building facades by using an unmanned aerial vehicle equipped with infrared thermal imaging according to claim 2 is characterized in that: In step S2, it includes: Divide the horizontal width of the building facade into n w A single image of the same width; W=w1*n w -(n w -1)*w1*O; O=20%~30%; Where W is the horizontal width of the building facade, w1 is the imaging width of a single image, and n w is the number of images divided on the horizontal width of the building facade, O is the overlap rate of two adjacent images; The first shooting coordinates of the drone are established as The coordinate origin is located at the leftmost end of the positive and negative zero positions on the building's facade.

4. The method for detecting defects on building facades by using an unmanned aerial vehicle equipped with infrared thermal imaging according to claim 3 is characterized in that: In step S2, the flight route of the drone is determined, including: starting from the first shooting coordinate, shooting along a serpentine reciprocating path from top to bottom; The drone starts shooting from the starting point and shoots n times in the horizontal direction according to the flight distance w1. w After taking the image, take the height h2 as the descent height, and return to shoot n images in the horizontal direction according to the flight distance w1. w images until the entire building facade is captured.

5. The method for detecting defects on building facades by using an unmanned aerial vehicle equipped with infrared thermal imaging according to claim 4 is characterized in that: In step S4, it includes: Step S401, naming the captured visible light images in the order of 1 to n; Step S402, stitching the visible light images on the same line of shooting path to obtain multiple lines of stitched images; Step S403: splicing two adjacent rows of visible light images.

6. The method for detecting defects on building facades by using an unmanned aerial vehicle equipped with infrared thermal imaging according to claim 5 is characterized in that: In step S402, it also includes: Perform grayscale processing on visible light images and use feature detection algorithms to obtain at least 20 feature points in each image; Taking the first image on the same shooting path as the reference origin, pair all feature points, select the first 10 feature points with the smallest pairing error, calculate the conversion matching matrix of the left and right adjacent images, and complete the stitching of multiple images on the same shooting path in sequence.

7. The method for detecting defects on building facades by using an unmanned aerial vehicle equipped with infrared thermal imaging according to claim 6, characterized in that: In step S403, it also includes: Feature extraction is performed on each row of images obtained after splicing, and they are spliced ​​from top to bottom according to the pairing effect to form a complete image of the building facade.

8. The method for detecting defects on building facades by using an unmanned aerial vehicle equipped with infrared thermal imaging according to claim 7 is characterized in that: In step S5, it includes: Scale the full image of the building facade to match the image detection size of the deep learning model; Use the deep learning model to select and remove interference factors from the complete image of the building facade; The deep learning model is trained using at least 2,000 infrared thermal imaging images of building facade defects.

9. The method for detecting defects on building facades by using an unmanned aerial vehicle equipped with infrared thermal imaging according to claim 8, characterized in that: In step S7, after the real temperature distribution image is segmented according to the image size in the detection module, the disease area in each real temperature image is calculated as follows: Among them, A 病害像素 is the pixel area of ​​the disease marked by the detection module, f is the focal length of the drone camera, and l1 is the shooting distance of the drone.

10. The method for detecting defects on building facades by using an unmanned aerial vehicle equipped with infrared thermal imaging according to claim 4, characterized in that: In step S3, it includes: The initial position coordinates of the drone and the coordinates of each shooting position during the movement along the shooting path are converted into GPS data and transmitted to the GPS positioning system of the drone to control the drone to complete flight shooting.

Citation Information

Patent Citations

  • Vegetation growth monitoring system and method integrating visible light camera and near-infrared camera

    CN115170964A

  • Building outer wall hollowing information estimation method and system based on infrared thermal imaging

    CN116697960A

  • Building facade dismantling device and dismantling method

    CN117403931A

Cited By

  • Urban area building energy-saving reconstruction evaluation method based on thermal imaging technology

    CN120451837A

  • Splicing drawing method for special-shaped shutter segments based on space coordinates

    CN120543768A

  • A splicing and drawing method for special-shaped blinds based on spatial coordinates

    CN120543768B

  • Visible light and infrared image fused building outer wall defect detection method

    CN120726482A

  • Abnormity detection and identification method for exterior facade of urban building

    CN121545045A