A bridge detection system and method balancing image quality and drone energy consumption

Abstract

The application discloses a bridge detection system and method balancing image quality and unmanned aerial vehicle energy consumption, and belongs to the technical field of bridge structure detection. In order to solve the problem that there is no coordination mechanism for image quality and energy consumption of the unmanned aerial vehicle in the bridge detection task, the application establishes an image quality evaluation model, analyzes the flight energy consumption of the unmanned aerial vehicle, establishes an image frame density and area weighting function model, constructs an image sampling sufficiency coefficient model, analyzes image change sensitivity, centrally represents the image quality evaluation model, the image sampling sufficiency coefficient model and the image change sensitivity, obtains a global image information total value, considers the global image information total value and the flight energy consumption of the unmanned aerial vehicle, establishes a multi-target task benefit index and obtains the optimal flight speed, the optimal flight height and the optimal camera frame rate of the unmanned aerial vehicle in the bridge detection. The application can reasonably consider the perception quality and the endurance efficiency of the unmanned aerial vehicle in the bridge detection.

Description

Technical Field

[0001] This invention belongs to the field of bridge structure inspection technology, specifically relating to a bridge inspection system and method that balances image quality and UAV energy consumption. Background Technology

[0002] In practical scenarios of bridge structure inspection, using drones to acquire high-quality images has gradually become fundamental for identifying bridge defects. However, to improve image quality, drones generally need to maintain a relatively stable flight state, a slow flight speed, and suitable attitude and flight path parameters. This extends the mission time and significantly increases energy consumption. In adverse weather conditions or complex terrain, the upward trend in energy consumption becomes even more pronounced as the frequency of dynamic adjustments to flight speed and path increases.

[0003] In current bridge inspection processes, most drones focus solely on the clarity or coverage of image acquisition, typically relying on fixed path planning and uniform frame rate settings. They fail to adequately consider the balance between image value and unit energy consumption. Furthermore, traditional flight control methods often prioritize minimizing energy consumption or minimizing the path, neglecting the varying contributions of different task areas, image acquisition densities, and the different contributions of sensitive areas to image value. Consequently, during actual inspection tasks, issues such as coverage redundancy, image blurring, and energy waste arise, significantly impacting the efficiency of bridge inspection. Summary of the Invention

[0004] This invention addresses the problem of the lack of a coordination mechanism between image quality and energy consumption in bridge inspection tasks using unmanned aerial vehicles (UAVs). It proposes a bridge inspection system and method that balances image quality and UAV energy consumption.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A bridge detection method that balances image quality and UAV power consumption includes the following steps:

[0007] S1. Establish an image quality evaluation model for bridge inspection images collected by UAVs;

[0008] S2. Analyze the energy consumption of UAVs during bridge inspection;

[0009] S3. Based on the importance of different bridge components in bridge structure detection, establish an image frame density and region weighting function model;

[0010] S4. Considering the impact of image frame drops or displacement on bridge detection defects, use the image frame density and region weighting function obtained in step S3 to construct an image sampling sufficiency coefficient model;

[0011] S5. When inspecting concrete bridges and steel structures, consider the impact of abrupt cracks and corrosion areas on image gradients, and analyze the sensitivity to image changes.

[0012] S6. The image quality evaluation model, the image sampling sufficiency coefficient model, and the image change sensitivity are centrally characterized to obtain the total value of global image information;

[0013] S7. Considering the total value of global image information and the energy consumption of UAV flight, establish a multi-objective task benefit index and obtain the optimal flight speed, optimal flight altitude, and optimal camera frame rate of UAV when detecting bridges; Step S1 considers the impact of camera tilt angle on resolution and the exponential effect of flight speed and altitude on image blur, and establishes an image quality evaluation model, the expression of which is:

[0014]

[0015] in, Rate the image quality; These are the camera configuration coefficients, obtained from the camera calibration file, with dimensions similar to... Maintain consistency; The camera's tilt angle relative to the ground ranges from 0 to 90 degrees and is obtained from the UAV monitoring platform; cos is the cosine function. This is the current flight altitude of the drone; To prevent the denominator from being zero by a minimal constant for flight altitude, and Dimensions are consistent; The resolution control index is determined by historical sample calibration or expert experience. The motion fuzzing intensity adjustment coefficient, with the dimension of time, is determined by simulation; For the drone's flight speed, The arc length coordinates of the drone along the preset inspection trajectory are obtained by the drone monitoring platform.

[0016] Furthermore, step S2 integrates three types of flight energy consumption, including speed-related thrust power, control compensation due to acceleration, and energy loss caused by external wind disturbances, to calculate the UAV's flight energy consumption, expressed as:

[0017]

[0018]

[0019] in, Let x be the energy cost per unit path at point x; This is the equivalent coefficient for the basic hovering power of a drone, with the dimension being power. To prevent the denominator from being zero, a minimal constant is used to calculate the flight speed of drones. Dimensions are consistent; The acceleration of the UAV is determined implicitly by the rate of change of velocity, based on the chain rule of spatial domain. The disturbance wind speed is obtained from the wind speed sensor; To control smoothness penalty items; , , , These are the kinetic energy consumption weighting coefficient, the operating load energy consumption coefficient, the wind disturbance loss coefficient, and the acceleration penalty coefficient, respectively, and are also used for matching dimensions; This is the equivalent transformation expression of the time derivative of acceleration at x.

[0020] Furthermore, the specific implementation method of step S3 includes the following steps:

[0021] S3.1. Establish a regional weighting function to evaluate the priority of task regions. The expression is:

[0022]

[0023] in, For regional weighting functions; The adjustment coefficient for the kth critical region is determined in conjunction with the task configuration parameters; The attenuation width of the kth critical region is obtained from historical samples and its dimension is consistent with x; K is the total number of critical regions. This represents the center position of the kth critical region in the arc-length coordinate system.

[0024] S3.2. Based on the region weighting function, construct the coupling relationship between image frame density and structural priority, expressed as:

[0025]

[0026] in, Image frame density; The density adjustment coefficient is determined by expert experience; This represents the camera's actual frame rate.

[0027] Furthermore, step S4 uses exponential decay to quantify the image sampling sufficiency or coverage approximation at each location, constructing an image sampling sufficiency coefficient model, the expression of which is:

[0028]

[0029] in, To ensure sufficient sampling; The adjustment coefficient corresponding to sufficiency is determined by historical sample calibration or expert experience, with the dimension being length, to ensure that the exponential term is dimensionless.

[0030] Furthermore, step S5 analyzes the sensitivity to image changes based on the second derivative of the prior structural complexity continuous scoring function, expressed as:

[0031]

[0032] in, Sensitivity to image changes; The adjustment coefficient, which is related to the sensitivity to image changes, is determined by expert experience and its dimension is area. This is a continuous scoring function for prior structural complexity.

[0033] Furthermore, step S6 performs a comprehensive value assessment on the images acquired at different path points to obtain the total value of the global image information, expressed as:

[0034]

[0035]

[0036] in, Rate the value of the image at the current location; The total value of global image information; The total length of the detection path is determined by the detection personnel. The sensitivity weighting coefficient is obtained from historical task data.

[0037] Furthermore, the specific implementation method of step S7 includes the following steps:

[0038] S7.1. Establish a multi-objective task benefit index, expressed as follows:

[0039]

[0040] in, For multi-objective task benefit indicators, To prevent energy consumption from having a denominator of zero, a minimum constant is preset by technicians, with dimensions consistent with... Maintain consistency;

[0041] S7.2. Establish constraints and solve for the optimal flight speed, optimal flight altitude, and optimal camera frame rate of the UAV during bridge inspection. The expressions are as follows:

[0042]

[0043]

[0044] in, , These are the minimum and maximum flight speeds, respectively, set by staff in conjunction with safety flight parameters for critical and non-critical areas. The wind speed coupling safety factor is determined by simulation calibration or expert experience. , These are the minimum and maximum flight altitudes, respectively, set by staff in conjunction with safety flight parameters for critical and non-critical areas. , These are the minimum and maximum values ​​of the camera frame rate, which are set by staff in conjunction with the safety flight parameters for critical and non-critical areas, respectively. The maximum permissible acceleration is set by staff in conjunction with the safe flight parameters for both critical and non-critical areas. Let be the optimal flight speed, and be the solution value of the equation; is the optimal flight altitude, and is the solution value of the equation; is the optimal camera frame rate, and is the solution value of the equation.

[0045] A bridge detection system that balances image quality and UAV power consumption includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed, it implements the steps of the bridge detection method that balances image quality and UAV power consumption.

[0046] The beneficial effects of this invention are:

[0047] This invention discloses a bridge detection method that balances image quality and UAV energy consumption. The method designs a control approach based on the correlation between image quality and flight energy efficiency. First, an image quality evaluation model is established, comprehensively considering the combined effects of camera attitude, flight altitude, and motion blur on image sharpness. Second, an energy consumption analysis model is introduced, quantifying energy consumption during flight from dimensions such as velocity, acceleration, and disturbed wind field. Next, a coupling mechanism between image frame density and the criticality of structural regions is established, guiding detection resources towards key components. Furthermore, relying on image coverage coefficients and an information mutation sensitivity model based on prior structural features, the completeness of image acquisition and structural change characteristics at various locations are dynamically evaluated. Using image information value as the core, an optimization index for the information-energy ratio is constructed to guide the acquisition of optimal flight control and acquisition parameter outputs.

[0048] This invention discloses a bridge detection method that balances image quality and UAV energy consumption. It constructs a coupled model of image quality and energy consumption, dynamically optimizing flight parameters based on regional criticality and abrupt image changes. This significantly improves image effectiveness and detection energy efficiency, offering high image acquisition quality, strong task adaptability, and high system energy utilization efficiency. It enhances the rationality and efficiency of UAV bridge detection. This invention optimizes flight speed, sampling frequency, and flight altitude, reasonably balancing perception quality and endurance efficiency during UAV bridge detection. Attached Figure Description

[0049] Figure 1 This is a flowchart of a bridge detection method that balances image quality and UAV power consumption according to the present invention;

[0050] Figure 2 This is a comparison chart of the drone's flight speed, high-speed flight, and camera frame rate before and after the adjustments made in this invention.

[0051] Figure 3 This is a comparison chart of image quality scores and task completion times before and after the adjustments made in this invention. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the invention and are not intended to limit the invention; that is, the described specific embodiments are merely a part of the embodiments of the invention, and not all of them. The components of the specific embodiments of the invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations, and the invention may also have other embodiments.

[0053] Therefore, the following detailed description of specific embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected specific embodiments of the invention. All other specific embodiments obtained by those skilled in the art based on these specific embodiments without inventive effort are within the scope of protection of this invention.

[0054] To further understand the invention's content, features, and effects, the following specific embodiments are provided, along with accompanying drawings. Figure 1 - Appendix Figure 3 Detailed explanation is as follows:

[0055] Example 1:

[0056] A bridge detection method that balances image quality and UAV power consumption includes the following steps:

[0057] S1. Establish an image quality evaluation model for bridge inspection images collected by UAVs;

[0058] Furthermore, step S1 considers the impact of camera tilt angle on resolution and the exponential effect of flight speed and altitude on image blur, establishing an image quality evaluation model, expressed as:

[0059]

[0060] in, Rate the image quality; These are the camera configuration coefficients, obtained from the camera calibration file, with dimensions similar to... Maintain consistency; The camera's tilt angle relative to the ground ranges from 0 to 90 degrees and is obtained from the UAV monitoring platform; cos is the cosine function. This is the current flight altitude of the drone; To prevent the denominator from being zero by a minimal constant for flight altitude, and Dimensions are consistent; The resolution control index is determined by historical sample calibration or expert experience. The motion fuzzing intensity adjustment coefficient, with the dimension of time, is determined by simulation; For the drone's flight speed, The arc length coordinates of the drone along the preset inspection trajectory are obtained by the drone monitoring platform.

[0061] Furthermore, the quality of images acquired by UAVs is a crucial foundation for disease identification. Multiple factors influence image quality, such as camera attitude, flight altitude, and motion blur. Typical models often neglect device attitude or linearly simplify blur effects, failing to accurately represent image degradation in highly dynamic scenes. The formula proposed in this step incorporates the impact of camera tilt on resolution, as well as the exponential effects of flight speed and altitude on image blur, constructing a quality function that simultaneously includes geometric degradation and dynamic blur mechanisms. This processing method allows for advance adjustments to flight strategies when performing disease identification tasks, enabling the acquisition of crucial structural images.

[0062] S2. Analyze the energy consumption of UAVs during bridge inspection;

[0063] Furthermore, step S2 integrates three types of flight energy consumption, including speed-related thrust power, control compensation due to acceleration, and energy loss caused by external wind disturbances, to calculate the UAV's flight energy consumption, expressed as:

[0064]

[0065]

[0066] in, Let x be the energy cost per unit path at point x; This is the equivalent coefficient for the basic hovering power of a drone, with the dimension being power. To prevent the denominator from being zero, a minimal constant is used to calculate the flight speed of drones. Dimensions are consistent; The acceleration of the UAV is determined implicitly by the rate of change of velocity, based on the chain rule of spatial domain. The disturbance wind speed is obtained from the wind speed sensor; To control smoothness penalty items; , , , These are the kinetic energy consumption weighting coefficient, the operating load energy consumption coefficient, the wind disturbance loss coefficient, and the acceleration penalty coefficient, respectively, and are also used for matching dimensions; This is the equivalent transformation expression of the derivative of acceleration with respect to time at x; The term represents the base energy consumption per unit path that increases dramatically due to the extended loiter time when the UAV is cruising at low speed. This results in a physical penalty for extremely low-speed flight during optimization, preventing the optimization from getting stuck in a deadlock where the speed approaches 0.

[0067] Furthermore, in actual bridge inspections, drones typically need to hover at low altitudes with high precision for extended periods, or cruise slowly along the interface. The rapid increase in energy consumption becomes a bottleneck for long-duration drone operations, especially when covering long-span bridges like suspension bridges. They often have to withstand sudden crosswinds and adjust their attitude multiple times. In such cases, simply considering the relationship between speed and wind resistance makes it difficult to accurately estimate the cost per unit path. The formula in step S2 integrates these three main types of consumption: thrust power related to speed, control compensation due to acceleration, and energy loss caused by external wind disturbances, providing a fine-grained assessment of energy consumption costs along the path. This model combines the spatial domain acceleration derivative with speed, improving the ability to suppress excessive speed fluctuations and helping to ensure the stability of the flight path.

[0068] S3. Based on the importance of different bridge components in bridge structure detection, establish an image frame density and region weighting function model;

[0069] Furthermore, the specific implementation method of step S3 includes the following steps:

[0070] S3.1. Establish a regional weighting function to evaluate the priority of task regions. The expression is:

[0071]

[0072] in, For regional weighting functions; The adjustment coefficient for the kth critical region is determined in conjunction with the task configuration parameters; The attenuation width of the kth critical region is obtained from historical samples and its dimension is consistent with x; K is the total number of critical regions. This represents the center position of the kth critical region in the arc-length coordinate system.

[0073] S3.2. Based on the region weighting function, construct the coupling relationship between image frame density and structural priority, expressed as:

[0074]

[0075] in, Image frame density; The density adjustment coefficient is determined by expert experience; This represents the camera's actual frame rate.

[0076] Furthermore, in bridge structure inspection, different components have varying degrees of importance. For example, critical areas such as the main beam base plate, supports, and hanger nodes require higher image resolution. To ensure that these sensitive areas have enough image frames for analysis, it is necessary to establish a coupling relationship between the frame density and structural priority. The formula in step S3 combines the image transmission frequency and the current position velocity, reflecting the density distribution of frame acquisition per unit length. It also introduces a structural partitioning weight function to evaluate the priority of the task area, so that the sampling density of the UAV at different locations will adapt to the inspection needs of structures with different levels of criticality.

[0077] S4. Considering the impact of image frame drops or displacement on bridge detection defects, use the image frame density and region weighting function obtained in step S3 to construct an image sampling sufficiency coefficient model;

[0078] During bridge inspection, defects and blemishes are often difficult to capture continuously due to dropped frames or displacement of the image. However, accurately determining the detection coverage of a certain area is crucial for image continuity and blemish evolution tracking systems. Therefore, this step establishes a function based on frame density distribution and region priority to quantify the sufficiency or approximation of image sampling at each location. This expression uses an exponential decay form, which can accurately simulate the degree of coverage loss caused by image capture in different regions.

[0079] Furthermore, step S4 uses exponential decay to quantify the image sampling sufficiency or coverage approximation at each location, constructing an image sampling sufficiency coefficient model, the expression of which is:

[0080]

[0081] in, To ensure sufficient sampling; The adjustment coefficient corresponding to sufficiency is determined by historical sample calibration or expert experience, with the dimension being length, to ensure that the exponential term is dimensionless.

[0082] S5. When inspecting concrete bridges and steel structures, consider the impact of abrupt cracks and corrosion areas on image gradients, and analyze the sensitivity to image changes.

[0083] When inspecting concrete bridges and steel structures, abrupt cracks and corrosion areas typically cause abrupt changes in the image gradient, revealing prominent second-order structural features. To predict these high-frequency defect areas in advance, this step no longer relies on a single flight state parameter but introduces a priori structural complexity score for the bridge. Based on the spatial second derivative of the prior score, this step constructs a sensitivity response model, which can identify areas of drastic change in the structural image. This function is used to trigger local deceleration, increase the frame rate, or reduce altitude, controlling resource allocation towards potentially defect-sensitive areas and improving the efficiency of defect detection.

[0084] Furthermore, step S5 analyzes the sensitivity to image changes based on the second derivative of the prior structural complexity continuous scoring function, expressed as:

[0085]

[0086] in, Sensitivity to image changes; The adjustment coefficient, which is related to the sensitivity to image changes, is determined by expert experience and its dimension is area. For the prior structural complexity continuous scoring function; It represents the second-order infinitesimal change in the prior structural complexity score.

[0087] Furthermore, since the feature scores extracted from historical inspection images are discrete spatial point data, this step employs a Gaussian kernel smoothing algorithm to reconstruct the spatial continuity of the discrete scores in order to achieve spatial differentiation. The specific expression of the prior structural complexity continuous scoring function is as follows:

[0088]

[0089] in, This represents the total number of sampling points in the historical inspection images; Let i be the spatial location of the i-th historical sampling point; For in position The structural complexity feature values ​​extracted are preferably texture contrast, local entropy, edge density, or gray-level co-occurrence matrix features. The smoothing scale parameter is obtained by calibration based on the typical spatial scale of historical disease areas or determined by expert experience.

[0090] S6. The image quality evaluation model, the image sampling sufficiency coefficient model, and the image change sensitivity are centrally characterized to obtain the total value of global image information;

[0091] For bridge inspection tasks involving long spans or high complexity, it is usually necessary to evaluate the comprehensive value of image acquisition at different path points in order to adjust the flight control rhythm and equipment energy consumption. This step introduces an information value function to centrally characterize the effects of image quality, frame density integrity, and sudden changes, providing a unified evaluation metric for optimization objectives.

[0092] Furthermore, step S6 performs a comprehensive value assessment on the images acquired at different path points to obtain the total value of the global image information, expressed as:

[0093]

[0094]

[0095] in, Rate the value of the image at the current location; The total value of global image information; The total length of the detection path is determined by the detection personnel. The sensitivity weighting coefficient is obtained from historical task data.

[0096] S7. Considering the total value of global image information and the energy consumption of UAV flight, establish a multi-objective task benefit index and obtain the optimal flight speed, optimal flight altitude, and optimal camera frame rate of the UAV when detecting bridges.

[0097] When conducting bridge inspection missions using drones, it is often necessary to balance information gain and energy consumption. For drones with limited battery life, it is crucial to capture clear and continuous images of defects while controlling mission costs. In this step, an information-energy ratio optimization model is constructed to balance information value and energy cost. This provides a comprehensive and optimal decision-making basis for controlling the quality and speed of image acquisition.

[0098] Furthermore, the specific implementation method of step S7 includes the following steps:

[0099] S7.1. Establish a multi-objective task benefit index, expressed as follows:

[0100]

[0101] in, For multi-objective task benefit indicators, To prevent energy consumption from having a denominator of zero, a minimum constant is preset by technicians, with dimensions consistent with... Maintain consistency;

[0102] S7.2. Establish constraints and solve for the optimal flight speed, optimal flight altitude, and optimal camera frame rate of the UAV during bridge inspection. The expressions are as follows:

[0103]

[0104]

[0105] in, , These are the minimum and maximum flight speeds, respectively, set by staff in conjunction with safety flight parameters for critical and non-critical areas. The wind speed coupling safety factor is determined by simulation calibration or expert experience. , These are the minimum and maximum flight altitudes, respectively, set by staff in conjunction with safety flight parameters for critical and non-critical areas. , These are the minimum and maximum values ​​of the camera frame rate, which are set by staff in conjunction with the safety flight parameters for critical and non-critical areas, respectively. The maximum permissible acceleration is set by staff in conjunction with the safe flight parameters for both critical and non-critical areas. Let be the optimal flight speed, and be the solution value of the equation; is the optimal flight altitude, and is the solution value of the equation; is the optimal camera frame rate, and is the solution value of the equation. This constitutes a dynamic speed lower limit. In one embodiment, a dynamic speed constraint is set based on the disturbed wind speed to avoid insufficient attitude control margin due to prolonged low-speed cruising under specific wind field conditions; acceleration It is not an independent optimization variable, but rather constrained by the laws of physical kinematics, implicitly determined by the rate of change of velocity, and participates in the optimization solution as the derivative term of velocity. constraint;

[0106] Using the above methods, the optimal flight speed, optimal flight altitude, and optimal camera frame rate of the UAV can be obtained, and the corresponding acceleration can be determined synchronously based on the derivative relationship of the speed. These aspects work together to determine the richness of image resources, uniformity of coverage, and energy-saving stability of bridge detection, enabling the UAV to complete bridge detection tasks with relatively high image quality and relatively low energy consumption.

[0107] The following is an application example of this embodiment:

[0108] In practical implementation, the inspection trajectory is discretized into multiple track points. The parameter values ​​are calculated at each track point, and the speed, altitude, and frame rate control sequences corresponding to each track point are obtained by using dynamic programming, sequential quadratic programming, or model predictive control methods. Based on the laws of physical kinematics, the corresponding acceleration is calculated by velocity sequence difference to meet the constraints.

[0109] A certain drone is equipped with a fixed focal length camera, flies at a speed of 2.0 m / s, flies at an altitude of 5 m, has a camera frame rate of 10 fps, an average power consumption of 180 W, an average image quality score of 0.65, and a mission completion time of 0.58 h.

[0110] Using this method, the UAV flew at a speed of 0.8 m / s and an altitude of 4 m in critical areas such as supports and main beams, with a camera frame rate of 18 fps; in non-critical areas, the flight speed was 3 m / s, the altitude was 6 m, and the camera frame rate was 8 fps. The overall image quality score was 0.74, and the mission completion time was 0.42 hours. The comparison results are as follows... Figure 2 and Figure 3 As shown.

[0111] Example 2:

[0112] A bridge detection system that balances image quality and UAV power consumption includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed, it implements the steps of a bridge detection method that balances image quality and UAV power consumption as described in Embodiment 1.

[0113] It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0114] Although this application has been described above with reference to specific embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of this application. In particular, as long as there is no structural conflict, the features in the specific embodiments disclosed in this application can be combined with each other in any way. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, this application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A bridge detection method that balances image quality and UAV power consumption, characterized in that, Includes the following steps: S1. Establish an image quality evaluation model for bridge inspection images collected by UAVs; S2. Analyze the energy consumption of UAVs during bridge inspection; S3. Based on the importance of different bridge components in bridge structure detection, establish an image frame density and region weighting function model; S4. Considering the impact of image frame drops or displacement on bridge detection defects, use the image frame density and region weighting function obtained in step S3 to construct an image sampling sufficiency coefficient model; S5. When inspecting concrete bridges and steel structures, consider the impact of abrupt cracks and corrosion areas on image gradients, and analyze the sensitivity to image changes. S6. The image quality evaluation model, the image sampling sufficiency coefficient model, and the image change sensitivity are centrally characterized to obtain the total value of global image information; S7. Considering the total value of global image information and the energy consumption of UAV flight, establish a multi-objective task benefit index and obtain the optimal flight speed, optimal flight altitude, and optimal camera frame rate of UAV when detecting bridges; Step S1 considers the impact of camera tilt angle on resolution and the exponential effect of flight speed and altitude on image blur, and establishes an image quality evaluation model, the expression of which is: ； in, Rate the image quality; These are the camera configuration coefficients, obtained from the camera calibration file, with dimensions similar to... Maintain consistency; The camera's tilt angle relative to the ground ranges from 0 to 90 degrees and is obtained from the UAV monitoring platform; cos is the cosine function. This is the current flight altitude of the drone; To prevent the denominator from being zero by a minimal constant for flight altitude, and Dimensions are consistent; The resolution control index is determined by historical sample calibration or expert experience. The motion fuzzing intensity adjustment coefficient, with the dimension of time, is determined by simulation; For the drone's flight speed, The arc length coordinates of the drone along the preset inspection trajectory are obtained by the drone monitoring platform.

2. The bridge detection method balancing image quality and UAV power consumption according to claim 1, characterized in that, Step S2 integrates three types of flight energy consumption: speed-related thrust power, control compensation due to acceleration, and energy loss caused by external wind disturbances, to calculate the UAV's flight energy consumption, expressed as: ； ； in, Let x be the energy cost per unit path at point x; This is the equivalent coefficient for the basic hovering power of a drone, with the dimension being power. To prevent the denominator from being zero, a minimal constant is used to calculate the flight speed of drones. Dimensions are consistent; The acceleration of the UAV is determined implicitly by the rate of change of velocity, based on the chain rule of spatial domain. The disturbance wind speed is obtained from the wind speed sensor; To control smoothness penalty items; , , , These are the kinetic energy consumption weighting coefficient, the operating load energy consumption coefficient, the wind disturbance loss coefficient, and the acceleration penalty coefficient, respectively, and are also used for matching dimensions; This is the equivalent transformation expression of the time derivative of acceleration at x.

3. The bridge detection method balancing image quality and UAV power consumption according to claim 2, characterized in that, The specific implementation method of step S3 includes the following steps: S3.

1. Establish a regional weighting function to evaluate the priority of task regions. The expression is: ； in, For regional weighting functions; The adjustment coefficient for the kth critical region is determined in conjunction with the task configuration parameters; The attenuation width of the kth critical region is obtained from historical samples and its dimension is consistent with x; K is the total number of critical regions. This represents the center position of the kth critical region in the arc-length coordinate system. S3.

2. Based on the region weighting function, construct the coupling relationship between image frame density and structural priority, expressed as: ； in, Image frame density; The density adjustment coefficient is determined by expert experience; This represents the camera's actual frame rate.

4. The bridge detection method balancing image quality and UAV power consumption according to claim 3, characterized in that, Step S4 uses exponential decay to quantify the image sampling sufficiency or coverage approximation at each location, constructing an image sampling sufficiency coefficient model, the expression of which is: ； in, To ensure sufficient sampling; The adjustment coefficient corresponding to sufficiency is determined by historical sample calibration or expert experience, with the dimension being length, to ensure that the exponential term is dimensionless.

5. The bridge detection method balancing image quality and UAV power consumption according to claim 4, characterized in that, Step S5 analyzes the sensitivity to image changes based on the second derivative of the prior structural complexity continuous scoring function, expressed as: ； in, Sensitivity to image changes; The adjustment coefficient, which is related to the sensitivity to image changes, is determined by expert experience and its dimension is area. This is a continuous scoring function for prior structural complexity.

6. The bridge detection method balancing image quality and UAV power consumption according to claim 5, characterized in that, Step S6 performs a comprehensive value assessment on the images acquired at different path points to obtain the total value of the global image information, expressed as: ； ； in, Rate the value of the image at the current location; The total value of global image information; The total length of the detection path is determined by the detection personnel. The sensitivity weighting coefficient is obtained from historical task data.

7. The bridge detection method balancing image quality and UAV power consumption according to claim 6, characterized in that, The specific implementation method of step S7 includes the following steps: S7.

1. Establish a multi-objective task benefit index, expressed as follows: ； in, For multi-objective task benefit indicators, To prevent energy consumption from having a denominator of zero, a minimum constant is preset by technicians, with dimensions consistent with... Maintain consistency; S7.

2. Establish constraints and solve for the optimal flight speed, optimal flight altitude, and optimal camera frame rate of the UAV during bridge inspection. The expressions are as follows: ； ； in, , These are the minimum and maximum flight speeds, respectively, set by staff in conjunction with safety flight parameters for critical and non-critical areas. The wind speed coupling safety factor is determined by simulation calibration or expert experience. , These are the minimum and maximum values ​​for flight altitude, respectively, set by staff in conjunction with safety flight parameters for critical and non-critical areas. , These are the minimum and maximum values ​​of the camera frame rate, which are set by staff in conjunction with the safety flight parameters for critical and non-critical areas, respectively. The maximum permissible acceleration is set by staff in conjunction with the safe flight parameters for both critical and non-critical areas. Let be the optimal flight speed, and be the solution value of the equation; is the optimal flight altitude, and is the solution value of the equation; is the optimal camera frame rate, and is the solution value of the equation.

8. A bridge inspection system that balances image quality and UAV power consumption, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed, implements the steps of a bridge detection method balancing image quality and UAV power consumption as described in any one of claims 1-7.

Images