A non-destructive testing method for disease of cable of cable-stayed bridge based on image sensor

Abstract

The application discloses a nondestructive testing method for cable diseases of a cable-stayed bridge based on an image sensor and particularly relates to the technical field of cable disease detection. The application realizes the technical leap from the superficial surface observation to the deep internal flaw detection by comprehensively detecting through the appearance screening and the active infrared thermal excitation technology path, firstly accurately positioning the surface cracking and the rust area, then performing the fixed-point hovering and the active heating on each defect, quantifying the temperature rise, the peak temperature and the cooling coefficient by analyzing the unique thermal response, accurately judging the internal extension degree and the damage depth of the defect, solving the problems of the missed detection and the misjudgment of the traditional method, providing the reliable data support for the disease evaluation, and realizing the technical leap from the superficial surface observation to the deep internal flaw detection.

Description

Technical Field

[0001] This invention relates to the field of cable defect detection technology, and more specifically, to a non-destructive testing method for cable defects in cable-stayed bridges based on image sensors. Background Technology

[0002] As one of the mainstream structural forms of long-span bridges, cable-stayed bridges rely on cables as core load-bearing components, which play a crucial role in transferring the load of the main girder to the bridge towers. The health of the cables directly determines the structural safety and service life of the cable-stayed bridge. However, during long-term service, cables are susceptible to environmental corrosion, fatigue loads, stress concentration, and human damage, resulting in various defects such as rust and cracking. If these defects are not detected and addressed in a timely manner, they may lead to a decrease in the cable's load-bearing capacity or even cause serious safety accidents such as bridge collapse.

[0003] However, existing methods for detecting cable defects in cable-stayed bridges still have the following shortcomings: Conventional nondestructive testing techniques often rely on a single testing method, which can only identify surface defects and cannot accurately determine the extent of internal extension and damage depth of defects. This results in test conclusions remaining on the surface, failing to achieve accurate diagnosis "from the surface to the inside", and making it difficult to scientifically assess the true severity and potential risks of defects. Meanwhile, existing methods lack quantitative analysis of the thermal response characteristics of defective areas, making it difficult to scientifically assess the severity of damage and potential risks, and failing to provide comprehensive and reliable data support for bridge maintenance.

[0004] To address this, a non-destructive testing method based on image sensors for cable defects in cable-stayed bridges is proposed. Summary of the Invention

[0005] To overcome the above-mentioned deficiencies of the prior art, embodiments of the present invention provide a non-destructive testing method for cable defects in cable-stayed bridges based on image sensors.

[0006] To achieve the above objectives, the present invention provides the following technical solution: A non-destructive testing method for cable defects in cable-stayed bridges based on image sensors includes: S1: Based on cable image data, identify and locate the appearance defect areas on the cable surface, construct a set of appearance defect areas, generate thermal excitation commands, and send them to the UAV; S2: After receiving the thermal excitation command, the UAV will control the UAV to arrive and hover according to the spatial coordinates of each appearance defect area in the appearance defect set. The onboard infrared heating lamp will be activated to actively excite the appearance defect area according to the preset heating time. During the heating process, the onboard infrared thermal imager will continuously take pictures at a preset frequency to obtain a time series thermal image set and transmit it to the server. S3: Based on thermal image atlas, extract the temperature-time curve of each pixel in the target area divided by the appearance defect area during the entire thermal excitation cycle. By comparing the temperature-time curve of the target area with the adjacent healthy area, and combining the cooling stage data after the thermal excitation cycle, output the thermal response difference coefficient of the current appearance defect area. S4: For each appearance defect area within the set of appearance defect areas, define the division area type of the cable to which each appearance defect area belongs; after comprehensive processing by combining the division area type, defect type, defect area area and thermal response difference coefficient, generate a defect detection report.

[0007] Specifically, step S1 involves identifying areas of visual defects; Using a pre-trained deep learning image recognition model, cable image data is analyzed in real time to automatically identify and locate suspected defect areas. The suspected defect areas are classified according to the types of cracking and corrosion, and the cracking defect areas and corrosion defect areas are output and a set of appearance defect areas is constructed.

[0008] Specifically, the rules for defining the target area and adjacent healthy areas in step S3; Obtain the contour information of the appearance defect area, calculate the arithmetic mean of the coordinates of all pixels within the contour, use it as the construction point, draw a circle with the construction point as the center and set the radius, and extract the part of the appearance defect area within the circle as the target area. Identify the entire heating area affected by the infrared heating lamp. After segmenting and removing the current appearance defect area within the heating area, mark the remaining area as the healthy heating area. Randomly select two position points on the segmentation line, and draw circles with preset radii using the two position points as the centers. After segmenting and extracting the part belonging to the heating area within the circle, obtain two sub-regions. Merge these two sub-regions as the adjacent healthy area.

[0009] Specifically, the comparison logic between the target area and adjacent healthy areas in step S3; S3-1: Extract the temperature data of each pixel at each time point within the thermal excitation cycle from the temperature-time curve of the target area. After averaging the temperature data of each pixel at the same time point, obtain the internal temperature performance data of the current appearance defect area at each time point within the thermal excitation cycle. Construct the defect temperature sequence based on the internal temperature performance data at each time point. S3-2: Similarly to step S3-1, extract the temperature data of each pixel at each time point within the thermal excitation cycle from the temperature-time curve of the adjacent healthy area, and construct a healthy temperature sequence. S3-3: Label each set of data in the defect temperature sequence as... The data sets of the healthy temperature sequence are labeled as follows: ; Number the time points; use the formula Calculate the temperature rise coefficient of the target area compared to the adjacent healthy area; S3-4: Identify the highest temperature values ​​in the defect temperature sequence and the healthy temperature sequence as the defect peak and the healthy peak. Calculate the ratio with the defect peak as the numerator and the healthy peak as the denominator to obtain the peak temperature coefficient of the target region relative to the adjacent healthy region.

[0010] Specifically, the data for the cooling phase after the thermal excitation cycle in step S3; After the thermal excitation cycle ends, the cooling phase begins. Temperature data of the target area and adjacent healthy areas are recorded after entering the cooling phase. The internal temperature performance data and health temperature performance data of the target area and adjacent healthy areas before the thermal excitation cycle are used as the baseline values. Temperature data recording ends when the internal temperature performance data and health temperature performance data of the target area and adjacent healthy areas reach the baseline values. For both the target area and the adjacent healthy area, the time difference between the end recording time and the start recording time is calculated to obtain the cooling time required for the target area and the cooling time required for the adjacent healthy area. The cooling coefficient of the target area compared to the neighboring healthy area is calculated by using the cooling time required for the target area as the numerator and the cooling time required for the neighboring healthy area as the denominator.

[0011] Specifically, the output thermal response difference coefficient is generated in step S3; The heating coefficient, peak temperature coefficient, and cooling coefficient of the current appearance defect area are comprehensively processed using weighted logic, and the thermal response difference coefficient of the current appearance defect area is output.

[0012] Specifically, the S4 step incorporates the specific logic of comprehensive processing; The regions can be categorized into key regions, regions of interest, and general regions. Set a set of regional additional coefficients corresponding to different regional division types; Set a set of additional coefficients for each type of cracking and corrosion disease; After determining the regional additional coefficient and the disease additional coefficient of each group of appearance defect areas, the disease severity coefficient of each group of appearance defect areas is obtained by combining the defect area and the thermal response difference coefficient and using weighted logic for comprehensive processing.

[0013] Specifically, the process of generating a disease detection report in step S4; After sorting the appearance defect areas in each group from largest to smallest according to the magnitude of the disease severity coefficient, the image information of the appearance defect areas is combined with the output of a pre-constructed report template to generate a disease detection report.

[0014] The technical effects and advantages of this invention are as follows: (1) By combining visual screening with active infrared thermal excitation technology, the surface cracks and corrosion areas are first accurately located. Then, each defect is hovered and actively heated. By analyzing its unique thermal response, the temperature rise, peak temperature and cooling coefficient are quantified, and the internal extension degree and damage depth of the defect are accurately judged. This solves the problem of missed detection and false detection in traditional methods, provides reliable data support for disease assessment, and realizes the technological leap from superficial surface observation to profound internal flaw detection. (2) By comprehensively processing the location importance, type, scale of defects and thermal response difference coefficients that reveal the degree of internal damage, a quantitative degree coefficient of defects is finally calculated for each defect area. Based on this, a detailed report containing risk ranking and maintenance recommendations is automatically generated. Managers can clearly understand the degree ranking and status of cable defects without interpreting complex data, thereby enabling the scientific allocation of maintenance resources and realizing the intelligent management and maintenance upgrade from passive response to proactive prevention and precise measures. Attached Figure Description

[0015] Figure 1 This is a flowchart of a non-destructive testing method for cable defects in cable-stayed bridges based on an image sensor, according to the present invention. Detailed Implementation

[0016] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] like Figure 1 As shown, a non-destructive testing method for cable defects in cable-stayed bridges based on image sensors includes: Path planning: Control the UAV to move at a constant speed along the cable along a pre-planned detection path, collect image data using the onboard image sensor, and transmit the image data to the server; Implementation instructions Based on the BIM model or 3D point cloud model of the cable-stayed bridge, a flight path is planned for the UAV that can cover all the cables to be inspected, and the path ensures that the UAV maintains a preset safe distance from the cables. And ensure that the sensor has the optimal imaging angle for the cable surface; The drone is controlled to fly at a constant speed along a preset path, and during the flight, the high-resolution visible light camera on the drone continuously collects high-definition images of the cable surface.

[0018] Appearance defect localization: Based on image data, the server identifies and locates appearance defect areas on the cable surface, constructs a set of appearance defect areas, generates thermal excitation commands, and sends them to the UAV. Specifically: Using a pre-trained deep learning image recognition model, cable image data is analyzed in real time, and suspected defect areas are automatically identified and located. The suspected defect areas are classified according to the types of cracking and corrosion, and the cracking defect areas and corrosion defect areas are output and a set of appearance defect areas is constructed. Implementation instructions After receiving the cable image data, the server first performs standardized preprocessing: Enhanced contrast: Makes disease features clearer; Geometric correction: Unfolds the image of the bent cable into a planar view for easier analysis; Noise reduction: Reduce image noise interference; After preprocessing, a pre-trained semantic segmentation model (such as U-Net or DeepLab) is called. The preprocessed image is input into the model. The model will classify each pixel in the image, determining whether it belongs to "background," "cracks," or "rust." The model will generate two corresponding "pixel-level masks": Crack mask: A binary image in which all pixels that the model identifies as "cracks" are white (value 1) and the rest are black (value 0). Rust mask: It is also a binary image, in which all pixels that the model judges as "rust" are white (value 1), and the rest are black (value 0); The original binary mask may contain some isolated noise points or discontinuous regions, requiring post-processing: Connectivity analysis: Connectivity analysis is performed on both the "cracked mask" and the "rusted mask". This process groups interconnected white pixels into a single "region". Filtering: Set an area threshold (for example, areas smaller than 10 pixels are considered noise) to filter out these small, unimportant areas; Ultimately, several independent cracking and corrosion defect areas were obtained for the cable at the current time point.

[0019] Regional heating excitation: After receiving the thermal excitation command, the UAV controls the UAV to hover after arriving at each appearance defect area according to the spatial coordinates of each appearance defect area in the appearance defect set; the spatial coordinates are calculated by fusing the UAV's GNSS / IMU system with the airborne visual odometry; the onboard infrared heating lamp is activated to actively excite the appearance defect area for a preset heating time (10-60 seconds); during the heating process, the onboard infrared thermal imager continuously takes pictures at a preset frequency (such as 1-10 frames per second) to obtain a time series thermal image set and transmit it to the server; Cracking: In addition to the crack area, its internal thermal response is crucial. If a long, thin crack is accompanied by a strong internal thermal anomaly, it indicates that the crack has penetrated the sheath, and moisture and air have entered, causing internal steel wire corrosion. Its risk level should be increased. Rust spots: If the rust spot area also shows internal thermal anomalies, it means that this is not just surface rust, but the "tip of the iceberg" of internal corrosion spreading outward, and its severity is far greater than that of simple surface corrosion without thermal anomalies.

[0020] Implementation instructions a. Areas of visual defects identified by drone hovering; b. The infrared heating lamp array integrated on the starter provides short-term, uniform active thermal excitation to the specific area (heating time is usually 10-60 seconds).

[0021] c. Before heating begins, during heating, and during cooling after heating stops, the airborne infrared thermal imager continuously captures images of the target area at a preset frame rate (e.g., 1-10 frames per second), obtaining a time-series thermal image set. Each thermal image is essentially a two-dimensional temperature matrix.

[0022] Time-series data analysis: Based on thermal image atlases, the server extracts the temperature-time curve of each pixel within the target area divided by the appearance defect area during the entire thermal excitation cycle. By comparing the temperature-time curves of the target area with those of the adjacent healthy areas, and combining the cooling stage data after the thermal excitation cycle, the server outputs the thermal response difference coefficient of the current appearance defect area. On the thermal image, a standard analytical region (target region) is drawn for the defect to be detected, and an equally standard analytical region (health region) is drawn next to it for a healthy control sample, thus enabling scientific comparison. The specific region division rules are as follows: Obtain the outline information of the appearance defect area from the server; Calculate the arithmetic mean of the coordinates of all pixels within the outline, and use it as the construction point; The instructions issued by the server contain a pixel-level mask of the defective area, and the location of the construction point is calculated based on the mask.

[0023] Using the construction point as the center, draw a circle with a set radius, and extract the part of the area that belongs to the appearance defect area inside the circle as the target area; Target area radius setting rules: For areas with cracked appearance defects, the radius can be set to 2-3 times the average width of the crack to ensure that the entire crack and its heat-affected zone are covered. For areas with rust-like defects, the radius can be set to a value that can cover most of its area (e.g., 70%-80%). Alternatively, a common default value can be used, which can be 20-50 pixels (converted to the actual distance based on the image resolution, for example, 5-10 centimeters).

[0024] In the healthy cable section near the defect and similarly heated, a fair control area is selected to identify the entire heating area affected by the infrared heating lamp; the entire range illuminated by the infrared heating lamp; after segmenting and removing the current appearance defect area in the heating area, the remaining area is marked as the healthy heating area; after removing the identified appearance defect area, the remaining area is the healthy area; two points are randomly selected on the dividing line, and circles are drawn with preset radii using the two points as centers. The portion belonging to the heating area within the circle is segmented and extracted to obtain two sub-regions. These two sub-regions are merged as the adjacent healthy area; Output the thermal response difference coefficient of the current appearance defect area, specifically: Extract the temperature data of each pixel at each time point within the thermal excitation cycle from the temperature-time curve of the target area; After averaging the temperature data of each pixel at the same time point, the internal temperature performance data of the current appearance defect area at each time point within the thermal excitation cycle is obtained. A defect temperature sequence was constructed using internal temperature data at various time points. Extract the temperature data of each pixel at each time point within the thermal excitation cycle from the temperature-time curve of the adjacent healthy area; After averaging the temperature data of each pixel at the same time point, we obtain the health temperature performance data of the current neighboring healthy area at each time point within the thermal excitation cycle. A health temperature series was constructed using health temperature performance data at various time points. A comprehensive calculation is performed between the defect temperature sequence and the healthy temperature sequence to output the temperature rise coefficient of the current appearance defect area; The data sets in the defect temperature sequence are labeled as follows: The data sets of the healthy temperature sequence are labeled as follows: ; Number the time points; Using formula Calculate the temperature rise coefficient of the target area compared to the adjacent healthy area; This reflects the difference in the overall thermal behavior of the two regions throughout the heating and cooling process; when When the value is approximately 1, it indicates that the temperature of the defective area is basically the same as that of the healthy area throughout the process; this suggests that the thermophysical properties (heat capacity, thermal conductivity) inside the defective area are not significantly different from those of the healthy area; the defect is likely limited to the surface and has not formed serious damage inside that would cause a significant change in thermal resistance, thus the risk level is low.

[0025] when A value >1 indicates that the average temperature of the defective region remains higher than that of the healthy region throughout the entire thermal excitation cycle; this suggests an anomaly exists within the defective area. The higher the value, the higher the likelihood of internal defects, and the more severe the defects may be.

[0026] The highest temperature values ​​in the defect temperature sequence and the healthy temperature sequence are identified as the defect peak and the healthy peak, respectively. The ratio of the defect peak as the numerator and the healthy peak as the denominator is calculated to obtain the peak temperature coefficient of the target area compared with the adjacent healthy area. After the thermal excitation cycle ends, the cooling phase begins. Temperature data of the target area and adjacent healthy areas are recorded after entering the cooling phase. The internal temperature performance data and health temperature performance data of the target area and adjacent healthy areas before the thermal excitation cycle are used as the baseline values. Temperature data recording ends when the internal temperature performance data and health temperature performance data of the target area and adjacent healthy areas reach the baseline values. For both the target area and the adjacent healthy area, the time difference between the end recording time and the start recording time is calculated to obtain the cooling time required for the target area and the cooling time required for the adjacent healthy area. The cooling coefficient of the target area compared to the neighboring healthy area is calculated by using the cooling time required for the target area as the numerator and the cooling time required for the neighboring healthy area as the denominator. When the cooling coefficient is approximately 1, the time required for the target area to cool to the reference temperature is almost the same as that for the healthy area. This indicates that the thermal conductivity characteristics inside the defective area are not significantly different from those in the healthy area, and heat can be dissipated from the area with the same efficiency. The appearance defect is likely not accompanied by serious internal damage, and the internal steel wires and filling materials maintain good thermal conductivity. When the cooling coefficient is greater than 1, the time required for the target area to cool to the reference temperature is longer than that for the healthy area. The larger the coefficient, the longer the time. This is a clear signal that there is a "thermal insulation layer" or "thermal reservoir". Internal defects (such as air gaps formed by broken steel wires or loose and porous rust products) act as a thermal insulation layer. There is a high probability of defects inside the area with the external defects. The larger the cooling coefficient, the thicker and more complete the internal thermal resistance layer, which corresponds to more serious steel wire breakage or more extensive internal corrosion.

[0027] The heating coefficient, peak temperature coefficient, and cooling coefficient of the current appearance defect area are comprehensively processed using weighted logic, and the thermal response difference coefficient of the current appearance defect area is output. The specific process of weighted logic synthesis is as follows: Using formula Calculate the thermal response difference coefficient K of the current appearance defect area; where and These represent the peak temperature coefficient and the cooling coefficient, respectively. , as well as These are the preset temperature rise reference coefficient, peak temperature reference coefficient, and cooling reference coefficient, respectively; where... , as well as These are preset weighting coefficients; The reference coefficient represents the ratio that the target area and the healthy area should exhibit in various dimensions under an acceptable state of slight abnormality. For example, under a healthy state, it is possible for one area to cool 10% slower than another due to various random factors. However, if the cooling coefficient of an area exceeds 1.10, significantly exceeding the normal range of health fluctuations, it is highly likely to be caused by internal diseases.

[0028] Detection result generation: For each appearance defect area within the set of appearance defect areas, define the division area type of the cable to which each appearance defect area belongs; after comprehensive processing by combining the division area type, defect type, defect area area and thermal response difference coefficient, generate a defect detection report and send it to the management personnel. Specifically: The regions can be categorized into key regions, regions of interest, and general regions. Based on bridge mechanics knowledge and maintenance standards, individual cables are divided into different importance levels. Critical areas are defined as follows: near the anchorage end (such as within 0-2 meters of the upper and lower anchor heads), where stress is concentrated and even minor damage can lead to catastrophic anchorage failure; and at the locations of cable clamps / shock absorbers, where constraints are complex and fatigue damage is likely to occur, and these areas need to be designated as critical areas. Area of ​​concern: The area 2 to 10 meters above the bridge deck and the middle area, which are exposed to wind and rain and are the main areas of vibration and wind-induced fatigue, should be set as the area of ​​concern. General area: The middle section of the cable and other non-critical parts have relatively uniform stress and are less affected by environmental loads, so they should be set as general areas.

[0029] Different regional division types are assigned a set of regional additional coefficients; the range of regional additional coefficients is limited to 1.03-1.18. The regional additional coefficient for key areas > the regional additional coefficient for areas of interest > the regional additional coefficient for general areas, for example, can be set to 1.16, 1.10, and 1.04 respectively.

[0030] Set a set of additional coefficients for each type of cracking and corrosion; the range of the additional coefficients is limited to 1.12-1.19; cracking has a higher priority than corrosion. For example, the additional coefficients for the two types of corrosion can be set to 1.16 and 1.12 respectively. After determining the regional additional coefficient and the disease additional coefficient of each group of appearance defect areas, the disease severity coefficient of each group of appearance defect areas is obtained by combining the defect area and the thermal response difference coefficient and using weighted logic for comprehensive processing; the defect area is obtained by counting the number of pixels in the appearance defect area and converting it into the actual area using image resolution. The process of obtaining the disease severity coefficient through weighted logical synthesis is as follows: The regional additional coefficient and the defect additional coefficient determined by each group of appearance defect areas are respectively marked as follows: and ; After normalizing the defect area and thermal response difference coefficient of the appearance defect area, they are respectively labeled as and ; Using formula The disease severity coefficient was calculated; where and These are the preset weighting coefficients.

[0031] After sorting the appearance defect areas in each group from largest to smallest according to the magnitude of the disease severity coefficient, the image information of the appearance defect areas is combined with the output of a pre-constructed report template to generate a disease detection report.

[0032] The above formulas are all dimensionless calculations. Dimensionless calculations can be performed using various methods such as standardization, which will not be elaborated here. The formulas are derived from software simulations based on a large amount of collected data, and the preset parameters in the formulas can be set by those skilled in the art according to the actual situation.

[0033] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, ATA hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. The semiconductor medium can be a solid-state ATA hard disk.

[0034] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0035] Those skilled in the art will recognize that the units and algorithm steps of the various examples 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 implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art 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.

[0036] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0037] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0038] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0039] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable ATA hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0040] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A non-destructive testing method for cable defects in cable-stayed bridges based on image sensors, characterized in that, include: S1: Based on cable image data, identify and locate the appearance defect areas on the cable surface, construct a set of appearance defect areas, generate thermal excitation commands, and send them to the UAV; S2: After receiving the thermal excitation command, the UAV will control the UAV to arrive and hover according to the spatial coordinates of each appearance defect area in the appearance defect set. The onboard infrared heating lamp will be activated to actively excite the appearance defect area according to the preset heating time. During the heating process, the onboard infrared thermal imager will continuously take pictures at a preset frequency to obtain a time series thermal image set and transmit it to the server. S3: Based on thermal image atlases, extract the temperature-time curve of each pixel within the target area defined by the appearance defect region throughout the entire thermal excitation cycle. By comparing the temperature-time curves of the target area with those of adjacent healthy areas, and combining the cooling stage data after the thermal excitation cycle: Extract the temperature data of each pixel at each time point within the thermal excitation cycle from the temperature-time curve of the target area, and calculate the average value of the temperature data of each pixel at the same time point to construct the defect temperature sequence. Similarly, a healthy temperature sequence is constructed from neighboring healthy regions; The temperature rise coefficient is obtained by calculating the ratio of the temperature data at each time point in the defect temperature sequence and the healthy temperature sequence, and then calculating the average value of the ratio. The highest temperature values ​​in the defect temperature sequence and the healthy temperature sequence are identified, and the peak temperature coefficient is obtained by calculating the ratio with the defect peak as the numerator and the healthy peak as the denominator. After the thermal excitation cycle ends, the cooling phase begins. The cooling coefficient is calculated by taking the cooling time required for the target area as the numerator and the cooling time required for the adjacent healthy area as the denominator. The heating coefficient, peak temperature coefficient, and cooling coefficient of the current appearance defect area are comprehensively processed using weighted logic, and the thermal response difference coefficient of the current appearance defect area is output. S4: For each appearance defect area in the appearance defect area set, define the division area type of the cable to which each appearance defect area belongs; After comprehensively processing the classification of the region type, the type of disease, the area of ​​the defective region, and the thermal response difference coefficient, a disease detection report is generated.

2. The non-destructive testing method for cable defects in cable-stayed bridges based on image sensors according to claim 1, characterized in that: Identify areas of visual defects in step S1; Using a pre-trained deep learning image recognition model, cable image data is analyzed in real time to automatically identify and locate suspected defect areas. The suspected defect areas are classified according to the types of cracking and corrosion, and the cracking defect areas and corrosion defect areas are output and a set of appearance defect areas is constructed.

3. The non-destructive testing method for cable defects in cable-stayed bridges based on image sensors according to claim 2, characterized in that: Rules for defining the target area and adjacent healthy areas in step S3; Obtain the contour information of the appearance defect area, calculate the arithmetic mean of the coordinates of all pixels within the contour, use it as the construction point, draw a circle with the construction point as the center and set the radius, and extract the part of the appearance defect area within the circle as the target area. Identify the entire heating area affected by the infrared heating lamp. After segmenting and removing the current appearance defect area within the heating area, mark the remaining area as the healthy heating area. Randomly select two position points on the segmentation line, and draw circles with preset radii using the two position points as the centers. After segmenting and extracting the part belonging to the heating area within the circle, obtain two sub-regions. Merge these two sub-regions as the adjacent healthy area.

4. The non-destructive testing method for cable defects in cable-stayed bridges based on image sensors according to claim 1, characterized in that: The comparison logic between the target area and the adjacent healthy area in step S3; S3-1: Extract the temperature data of each pixel at each time point within the thermal excitation cycle from the temperature-time curve of the target area. After averaging the temperature data of each pixel at the same time point, obtain the internal temperature performance data of the current appearance defect area at each time point within the thermal excitation cycle. Construct the defect temperature sequence based on the internal temperature performance data at each time point. S3-2: Similarly to step S3-1, extract the temperature data of each pixel at each time point within the thermal excitation cycle from the temperature-time curve of the adjacent healthy area, and construct a healthy temperature sequence. S3-3: Label each set of data in the defect temperature sequence as... The data sets of the healthy temperature sequence are labeled as follows: ; Number the time points; use the formula Calculate the temperature rise coefficient of the target area compared to the adjacent healthy area; The temperature rise coefficient is obtained by calculating the ratio of the temperature data at each time point in the defect temperature sequence and the healthy temperature sequence, and then calculating the average value of the ratio. S3-4: Identify the highest temperature values ​​in the defect temperature sequence and the healthy temperature sequence as the defect peak and the healthy peak, respectively. Calculate the ratio of the defect peak as the numerator and the healthy peak as the denominator to obtain the peak temperature coefficient.

5. The non-destructive testing method for cable defects in cable-stayed bridges based on image sensors according to claim 4, characterized in that: Data on the cooling phase after the thermal excitation cycle in step S3; After the thermal excitation cycle ends, the cooling phase begins. Temperature data of the target area and adjacent healthy areas are recorded after entering the cooling phase. The internal temperature performance data and health temperature performance data of the target area and adjacent healthy areas before the thermal excitation cycle are used as the baseline values. Temperature data recording ends when the internal temperature performance data and health temperature performance data of the target area and adjacent healthy areas reach the baseline values. For both the target area and the adjacent healthy area, the time difference between the end recording time and the start recording time is calculated to obtain the cooling time required for the target area and the cooling time required for the adjacent healthy area. The cooling coefficient is calculated by dividing the time required to cool the target area by the time required to cool the adjacent healthy area by the ratio of the two values.

6. The non-destructive testing method for cable defects in cable-stayed bridges based on image sensors according to claim 2, characterized in that: The specific logic of the integrated processing is incorporated in step S4; The regions can be categorized into key regions, regions of interest, and general regions. Set a set of regional additional coefficients corresponding to different regional division types; Set a set of additional coefficients for each type of cracking and corrosion disease; After determining the regional additional coefficient and the disease additional coefficient of each group of appearance defect areas, the disease severity coefficient of each group of appearance defect areas is obtained by combining the defect area and the thermal response difference coefficient and using weighted logic for comprehensive processing.

7. The non-destructive testing method for cable defects in cable-stayed bridges based on image sensors according to claim 6, characterized in that: The specific process of generating a disease detection report in step S4; After sorting the appearance defect areas in each group from largest to smallest according to the magnitude of the disease severity coefficient, the image information of the appearance defect areas is combined with the output of a pre-constructed report template to generate a disease detection report.

Images