Construction supervision inspection method and system based on unmanned aerial vehicle

Abstract

The invention relates to a construction supervision inspection method and system based on an unmanned aerial vehicle, and relates to the field of construction supervision inspection of the unmanned aerial vehicle. Acquiring construction site image information and matching a flight mode according to an environment wind power value; frame-selecting helmet information from the construction site image information based on preset helmet features; when the helmet information is inconsistent with the preset reference helmet information, giving an alarm and continuing to inspect; when the helmet information is consistent with the preset reference helmet information, collecting construction site audio information; if the construction site audio information is consistent with the preset reference audio information, matching is completed, and inspection is continued; and if the construction site audio information is inconsistent with the preset reference audio information, giving an alarm and controlling the unmanned aerial vehicle to continue inspection. According to the invention, potential safety hazards are reduced, and unmanned supervision is realized at the same time.

Description

Technical Field

[0001] The present invention relates to the field of construction supervision and inspection using unmanned aerial vehicles (UAVs), and in particular to a construction supervision and inspection method and system based on unmanned aerial vehicles (UAVs). Background Art

[0002] With the acceleration of urbanization, construction projects such as various buildings and facilities are increasing, and the safety and quality of construction have become the key to construction operations.

[0003] The construction process presents numerous uncertainties and risks. For example, excavation can cause settlement and displacement of the surrounding soil, potentially impacting the safety of surrounding buildings, roads, pipelines, and more. Therefore, strict supervision of construction sites is essential to ensure that safety, quality, and environmental standards are met.

[0004] In response to the above-mentioned technical issues, traditional supervision methods mainly rely on manual inspections, which are greatly affected by external factors and are prone to fatigue and laziness, making it difficult to detect safety hazards in a timely manner. Summary of the Invention

[0005] In order to reduce potential safety hazards and realize unmanned supervision, the present invention provides a construction supervision and inspection method and system based on drones.

[0006] In a first aspect, the present invention provides a construction supervision and inspection method based on a drone, which adopts the following technical solutions: A construction supervision and inspection method based on drones, comprising: S1: responding to a trigger instruction to collect ambient wind force values; S2: Collect construction site image information and match the flight mode according to the ambient wind value; S3: selecting helmet information from the construction site image information based on preset helmet features; S40: When the helmet information is inconsistent with the preset reference helmet information, an alarm is issued and the inspection continues; S41: When the helmet information is consistent with the preset reference helmet information, collecting the construction site audio information; S410: If the construction site audio information is consistent with the preset reference audio information, the matching is completed and the inspection continues; S411: If the construction site audio information is inconsistent with the preset baseline audio information, an alarm is issued and the drone is controlled to continue the inspection.

[0007] By adopting the above technical solution, the flight mode is matched according to the environmental wind value to cope with different situations. The helmet information is compared based on the construction site image information. If there is inconsistency, it means that the personnel are mismatched and an alarm is issued and the inspection continues. If the construction site audio information is inconsistent with the baseline audio information, it means that there is a problem with the construction. An alarm is issued and the inspection continues, thereby reducing safety hazards and realizing unmanned supervision.

[0008] Optional, including: S4100: Collect the tool distance difference between the helmet and the preset tool and the tool's movement frequency; S4101: Matching a reference amplitude curve based on helmet information; S4102: Obtaining a tool distance variation amplitude based on the tool distance difference and the motion frequency; S4103: Generate a construction amplitude curve during tool construction based on distance variation fitting; S41030: When the construction amplitude curve is consistent with the reference amplitude curve, the matching is completed and the inspection continues; S41031: When the construction amplitude curve is inconsistent with the reference amplitude curve, an alarm is issued and the drone is controlled to continue inspection.

[0009] By adopting the above technical solution, the distance difference between the helmet and the tool and the movement frequency of the tool are collected, and the construction amplitude curve is fitted and compared with the baseline amplitude curve. If there is inconsistency, it means that there is a problem with the construction action, which will affect construction safety or project quality. The drone will issue an alarm and continue to inspect, improving the accuracy of construction monitoring, thereby reducing safety hazards and realizing unmanned supervision.

[0010] Optional, including: S20: When the drone is performing an inspection, it is determined whether the ambient wind value is greater than a preset reference wind value; S200: When the ambient wind force value is not greater than the reference wind force value, the drone is controlled to switch to the hovering mode and perform inspection at the preset reference power; S210: When the ambient wind value is greater than the reference wind value, collecting the ambient wind direction and determining a shelter position within the construction site based on the construction site image information and the ambient wind direction; S211: identifying the spatial volume of the avoidance position based on the avoidance position, and selecting a avoidance position that is larger than a preset reference volume; S212: Determine a wind shelter position from the selected shelter positions based on the preset construction position, and match the wind shelter power of the UAV according to the ambient wind direction and the ambient wind force value; S213: Control the UAV to switch to the wind avoidance mode and move it to a wind avoidance location at the wind avoidance power for inspection.

[0011] By adopting the above technical solution, when the ambient wind value is not greater than the reference wind value, it means that the ambient wind will not affect the inspection; when the ambient wind value is greater than the reference wind value, the avoidance position is selected, the shelter power is matched according to the ambient wind direction and ambient wind value, and the flight mode is switched for inspection, thereby reducing safety hazards and realizing unmanned supervision.

[0012] Optional, including: S2110: When the spatial volume of the sheltering position is not greater than the reference volume, the wind force difference between the ambient wind force value and the reference wind force value is calculated; S2111: Get strong power based on wind force difference matching; S2112: Obtain the force angle based on the ambient wind direction and construction location; S2113: Control the drone to switch to strong mode and conduct inspections with strong power and strong angle.

[0013] By adopting the above technical solution, if the space volume of the sheltering position is insufficient, it means that there is no place to shelter, and it is necessary to conduct an on-site inspection, calculate the wind force difference between the ambient wind force value and the reference wind force value to match the strong power, determine the strong angle based on the ambient wind direction and the construction position, switch to the strong mode for inspection, and realize on-site supervision.

[0014] Optional, including: S201: Selecting a preset parking feature from the construction site image information, and defining the selected position as a landing position; S202: Obtaining an inspection direction and an inspection distance based on the construction position and the landing position; removing landing positions whose inspection distances do not fall within a preset identification interval based on the inspection direction; S203: determining the height coordinates of all objects between the construction position and the landing position after removal based on the construction site image information, removing positions where the height coordinates of objects are higher than the landing position; S204: Collecting the ambient wind direction and the UAV position, matching the flight power required to move from the UAV position to the remaining landing positions based on the ambient wind force and direction, and selecting the landing position with the lowest power from the matched flight powers and defining it as the parking position; S205: Control the UAV to switch to the parking mode and move it to the parking position at the flight power to land and conduct inspection.

[0015] By adopting the above technical solution, the landing position is determined according to the parking features selected by the frame, the positions that cannot be properly supervised are removed, and the parking position with the lowest power consumption is selected according to the ambient wind direction. The UAV is controlled to land and inspect in parking mode, thereby reducing the power consumption of the inspection.

[0016] Optional, including: S214: When moving toward the parking position, continuously collecting parking image information of the parking position; S215: identifying a preset reference static feature in the stationary image information using a preset reference static feature set; S216: Based on the reference static feature, the stationary image information is sequentially compared, and the distance difference between the reference static feature and the stationary position is sequentially calculated; S2160: When the distance difference is equal to the preset reference distance difference, the UAV continues to move at the flight power to the parking position, lands, and conducts inspection; S2161: When the distance difference is not equal to the reference distance difference, the distance differences are arranged in order of magnitude, and the maximum distance difference is selected; S21610: Obtain the adjustment range of the drone based on the maximum value matching of the distance difference; S21611: Get the adjustment direction of the drone based on the ambient wind direction; S21612: Control the drone to move to the parking position and land, and then conduct inspection after adjusting the amplitude and direction.

[0017] By adopting the above technical solution, the distance difference between the reference static feature and the parked position is determined by identifying the reference static feature, and the adjustment amplitude of the UAV when landing is obtained based on the maximum value of the distance difference, so as to control the UAV to patrol at the parked position with the adjustment amplitude, so as to eliminate the vibration of the parked position due to wind, reduce the impact on the patrol, improve the accuracy and reliability of the patrol, and thus reduce safety hazards.

[0018] Optional, including: S217: collecting the balance angle difference of the UAV at the parking position; S2170: If the balance angle difference is 0, control the drone to perform inspection; S2171: If the balance angle difference is not 0, the balance angle differences are arranged in order of magnitude, and the maximum balance angle difference is selected; S21710: Obtaining a jitter amplitude at the parking position based on the maximum value matching of the balance angle difference; S21711: Obtaining a matching amplitude corresponding to the drone based on the jitter amplitude, and controlling the drone to adjust at the parking position using the matching amplitude.

[0019] By adopting the above technical solution, after landing at the parking position, if the collected balance angle difference of the drone is not 0, it means that the drone is still shaking after landing. The maximum value of the balance angle difference is then screened out to match the shaking amplitude, and the drone is adjusted accordingly to eliminate the shaking and ensure the stability of the inspection.

[0020] Optional, including: S50: Determine the personnel position and the edge position based on the construction site image information, and when the personnel position falls into the edge position, collect the inspection position of the drone, the inspection position includes the target inspection position and the active inspection position; S51: Calculate the drone distance between the target inspection position and the active inspection position; S520: When the distance between the UAV and the human is less than the preset collision distance and the ambient wind value is not greater than the reference wind value, determining the movement direction of the human based on the human position and the edge position; S5200: Obtaining an interaction position based on the active inspection position and the target inspection position by matching the personnel movement direction; S5201: Obtain an interaction path based on the interaction position, and control the UAV to interact using the interaction path.

[0021] By adopting the above technical solution, when a person approaches the edge, the distance to the drone is calculated to determine whether there is a drone nearby. When the distance is too small and there is no wind, it means that a drone will be encountered when avoiding it. By determining the interaction position and controlling the interaction of the drone, the interactivity and safety of the inspection are improved, safety hazards are reduced, and unmanned supervision is achieved at the same time.

[0022] Optional, including: S521: When the distance between the UAV and the target inspection position and the active inspection position is less than the preset collision distance and the ambient wind value is greater than the reference wind value, a wind-shielding interactive position is selected from the avoidance positions based on the target inspection position and the active inspection position; S5210: Match the interactive power of the UAV based on the ambient wind value and the wind shelter interactive position; S5211: Matching the UAV’s wind avoidance interaction path based on the personnel’s position and wind avoidance interaction position; S5212: Control the UAV to complete the interaction at the wind avoidance interaction position using the wind avoidance interaction path and wind avoidance power, and then conduct an inspection.

[0023] By adopting the above technical solution, when the distance between drones is small and the wind is strong, the wind-sheltered interaction positions are screened, a suitable wind-sheltered interaction position is selected, and then the interaction power and interaction path are matched to control the drone to complete the interaction at the wind-sheltered interaction position, ensuring the consistency of supervision and achieving comprehensive and accurate supervision.

[0024] Secondly, this application provides a construction supervision and inspection system based on drones, which adopts the following technical solutions: An acquisition module is used to obtain environmental wind force values, construction site image information, and construction site audio information; A memory for storing a program for a construction supervision and inspection method based on a drone; The processor loads and executes the program in the memory.

[0025] In summary, this application includes at least one of the following beneficial technical effects: By matching helmets with construction workers to determine the construction worker's situation and whether the personnel are matched, the construction situation can be determined by comparing the audio information of the construction site with the reference audio information, and whether there are any construction problems, thereby reducing safety hazards and achieving unmanned supervision. The drone automatically matches the flight mode according to the ambient wind speed value and flexibly adjusts the flight strategy under different wind conditions to ensure the continuity and efficiency of the inspection mission. When it is difficult to find a suitable location, the strong mode is used to enhance the wind resistance and ensure the flexibility of the inspection, thus realizing unmanned supervision. The drone monitors the relationship between the personnel position and the edge position, calculates the drone distance, and identifies potential collision risks. When personnel are close to the edge, the drone interacts by planning the interaction position and interaction path. In strong wind conditions, the drone selects sheltered interaction positions based on the avoidance position to complete the inspection task and improve the continuity of the inspection. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 This is a method flow chart of a construction supervision and inspection method based on a drone according to an embodiment of the present invention; Figure 2 is a method flow chart of a construction action matching method according to an embodiment of the present invention; Figure 3 This is the method flow of the flight mode matching method of the embodiment of the present invention Figure 1 ; Figure 4 This is the method flow of the flight mode matching method of the embodiment of the present invention Figure 2 ; Figure 5 This is the method flow of the flight mode matching method of the embodiment of the present invention Figure 3 ; Figure 6 This is the method flow of the flight mode matching method of the embodiment of the present invention Figure 4 ; Figure 7 This is the method flow of the flight mode matching method of the embodiment of the present invention Figure 5 . DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0028] The embodiment of the present application discloses a construction supervision and inspection method based on a drone, which switches different flight modes according to different environmental wind speeds to supervise personnel at the construction site.

[0029] Reference Figure 1 , a construction supervision and inspection method based on drones, comprising the following steps: Step S1: Responding to a trigger instruction to collect environmental wind force values.

[0030] The trigger command refers to the command to start the drone inspection when the construction site starts, and the command is issued through the start button.

[0031] The ambient wind value refers to the wind force in the current environment, which is obtained through a wind sensor preset on the construction site. The wind sensor can collect the ambient wind value and ambient wind direction. The wind sensor is pre-set by technicians based on actual conditions and will not be described in detail here.

[0032] Step S2: Collect construction site image information and match the flight mode according to the ambient wind value.

[0033] Flight mode refers to the flight status of the drone, which includes wind avoidance mode, hovering mode, strong mode and parking mode.

[0034] The construction site image information refers to the construction site image data obtained through multiple cameras preset on the drone. The cameras are pre-set by technicians based on actual conditions and will not be described in detail here.

[0035] After collecting construction site image information, different flight modes are matched according to the different ambient wind values. When the ambient wind value is large, if there is a suitable sheltered location, the wind shelter mode is switched to avoid strong winds; if there is no suitable sheltered location, the strong mode is switched to change the flight power and improve wind resistance; if there is a suitable parking location, the parking mode is switched to move to the parking location to avoid and reduce power consumption.

[0036] Step S3: selecting helmet information from the construction site image information based on preset helmet features.

[0037] The helmet feature refers to the image feature of the helmet, which is pre-set by the technicians according to the actual situation and will not be described in detail here.

[0038] Helmet information refers to helmet-related data identified from the image, such as helmet color, shape, logo, and number.

[0039] By annotating a large number of images with helmet features, inputting the annotated images with helmet features into the YOLO large model, repeatedly stacking images using the Pytorch network architecture to extract helmet features, and calculating the error between the results and the data, it is put into use when the error is less than 1%.

[0040] The construction site image information is input into the large model. When the helmet features are recognized, the helmet features are marked and selected from the construction site image information. Features such as helmet color, shape, logo and number are identified in the selected image, which is the helmet information.

[0041] Step S40: When the helmet information is inconsistent with the preset reference helmet information, an alarm is issued and the inspection continues.

[0042] The baseline helmet information refers to the helmet information of the workers at the construction location, which is pre-set by the technicians based on the actual situation and will not be described in detail here.

[0043] The alarm is a warning signal that reminds construction workers when the helmet information does not match the construction location.

[0044] When the helmet information is inconsistent with the preset benchmark helmet information, it means that the personnel and the construction location do not match, and there may be safety hazards. The helmet will issue an alarm, and after issuing a warning, the inspection will continue.

[0045] Step S41: When the helmet information is consistent with the preset reference helmet information, the construction site audio information is collected.

[0046] Construction site audio information refers to the sound produced by corresponding workers during operations on the construction site.

[0047] When the helmet information is consistent with the preset reference helmet information, it means that the personnel on the construction site and the construction location are successfully matched, thereby controlling the helmet to collect audio information from the construction site.

[0048] Step S410: If the construction site audio information is consistent with the preset reference audio information, the matching is completed and the inspection continues.

[0049] The reference audio information refers to the normal sound produced by corresponding workers on the construction site during operation. It is pre-set by technical personnel based on actual conditions and will not be described in detail here.

[0050] If the construction site audio information is consistent with the preset baseline audio information, it means that there is no problem with the operation of the construction site personnel and the construction situation is normal. The matching is completed and the inspection continues.

[0051] Step S411: If the construction site audio information is inconsistent with the preset reference audio information, an alarm is issued and the drone is controlled to continue the inspection.

[0052] If the audio information from the construction site is inconsistent with the preset baseline audio information, it means that there may be construction errors and other problems during construction, which may easily cause safety hazards. An alarm will be issued through the helmet and the drone will be controlled to continue inspection.

[0053] After step S410, the following steps are also included: Step S4100: Collect the tool distance difference between the helmet and the preset tool and the tool movement frequency.

[0054] The tool distance difference refers to the distance difference between the helmet and the tool. The helmet and tool are pre-connected, and the distance between the two is collected by Bluetooth distance sensors preset on the helmet and tool. The model and installation location of the Bluetooth distance sensor are pre-set by technicians based on actual conditions and are not detailed here.

[0055] Movement frequency refers to the speed of the tool's movement during construction. The tool's movement frequency is collected by the preset acceleration sensor on the tool, output to the processor in the helmet, and uploaded to the cloud for storage through the processor.

[0056] Step S4101: Match a reference amplitude curve based on helmet information.

[0057] The reference amplitude curve refers to the reference amplitude curve of the tool movement corresponding to the personnel at the construction position. It is used to judge the construction situation of the personnel. It is pre-set by the technical personnel based on the actual situation and will not be described in detail here.

[0058] Each helmet has a preset reference amplitude curve, and the corresponding reference amplitude curve can be matched according to the determined helmet information.

[0059] Step S4102: Obtain the tool distance variation amplitude based on the tool distance difference and the motion frequency.

[0060] The distance variation amplitude refers to the variation amplitude of the distance when the tool moves, including the frequency, phase and other parameters of the amplitude.

[0061] By inputting the tool distance difference and movement frequency into a database, the corresponding distance variation amplitude is obtained. The database pre-stores the relationship between the tool distance difference and the corresponding distance variation amplitude corresponding to the movement frequency. A larger tool distance difference indicates a greater distance between the helmet and the tool; a faster movement frequency indicates a faster speed at which the tool is being moved by the operator. The corresponding distance variation amplitude can be obtained based on the tool distance difference and movement frequency.

[0062] Step S4103: Generate a construction amplitude curve during tool construction based on distance variation amplitude fitting.

[0063] The construction amplitude curve refers to the amplitude curve of the tool being moved during construction.

[0064] Taking time as the horizontal coordinate and the distance change amplitude as the vertical coordinate, the change rate at the unit time point is obtained in sequence, and the adjacent change rates are connected to obtain the construction amplitude curve.

[0065] Step S41030: When the construction amplitude curve is consistent with the reference amplitude curve, the matching is completed and the inspection continues.

[0066] When the construction amplitude curve is consistent with the reference amplitude curve, it means that the construction action is correct and there is no safety hazard. The matching is completed and the inspection continues.

[0067] Step S41031: When the construction amplitude curve is inconsistent with the reference amplitude curve, the helmet will sound an alarm and control the drone to continue inspection.

[0068] When the construction amplitude curve is inconsistent with the benchmark amplitude curve, it indicates that the construction action is wrong and there is a safety hazard. An alarm will be issued through the helmet and the drone will be controlled to continue the inspection.

[0069] The flight mode includes a wind avoidance mode and a hovering mode. The flight mode matching method in step S2 includes: Step S20: When the drone is performing an inspection, it is determined whether the ambient wind value is greater than a preset reference wind value.

[0070] The baseline wind speed value refers to the wind speed in a calm state. In this application, a calm state is also defined as a wind speed between 0 and 3.5 m / s. The baseline wind speed value is used to determine whether the wind environment can affect drone inspections. It is pre-set by technicians based on actual conditions and is not detailed here.

[0071] Step S200: When the ambient wind force value is not greater than the reference wind force value, the drone is controlled to switch to a hovering mode and perform inspections at a preset reference power.

[0072] Hover mode refers to the mode in which the drone remains stationary in the air.

[0073] The baseline power refers to the operating power of the drone to maintain hovering in a windless state. It is pre-set by technicians based on actual conditions and will not be described in detail here.

[0074] When the ambient wind value is not greater than the reference wind value, it means that the ambient wind will not affect the flight of the drone. The drone is controlled to switch to hovering mode and perform inspections at the reference power.

[0075] Step S210: When the ambient wind value is greater than the reference wind value, the ambient wind direction is collected and a sheltering position within the construction site is determined based on the construction site image information and the ambient wind direction.

[0076] Ambient wind direction refers to the current wind direction, which is collected by pre-installed wind sensors on the construction site. Avoidance position refers to the position where the drone can avoid strong winds.

[0077] When the ambient wind speed is greater than the baseline wind speed, it indicates that the ambient wind will affect the drone's hovering, and the ambient wind direction is collected. Preset locations are selected in the construction site image information. These locations can reduce the impact of wind on the drone's flight. Corresponding locations are matched based on different wind directions and wind speed values. Experimental data is recorded to generate a database for later retrieval.

[0078] When the ambient wind force value and the ambient wind direction are known, this data is input into a pre-stored database to obtain the coordinates of several corresponding positions, which are the avoidance positions.

[0079] Step S211: identifying the spatial volume of the avoidance position based on the avoidance position, and selecting a avoidance position that is larger than a preset reference volume.

[0080] Spatial volume refers to the size of the space at the dodging position. The drone uses binocular stereo vision to identify the spatial volume at the dodging position.

[0081] The reference volume refers to the volumetric space that can reduce the impact of wind on the drone. It is pre-set by technicians based on actual conditions and will not be described in detail here.

[0082] By selecting a position larger than the reference volume among the sheltering positions, a position that satisfies the requirement of reducing the impact of wind on the drone is found, thereby screening the sheltering positions and obtaining a position where the drone can take shelter from the wind.

[0083] Step S212: Determine a sheltered location from the selected sheltered locations based on the preset construction location, and match the sheltered power of the UAV according to the ambient wind direction and ambient wind force value.

[0084] The construction location refers to the construction area on the construction site, which is pre-set by technicians based on actual conditions and is not detailed here. The wind shelter location refers to the final shelter location selected. The wind shelter power refers to the power required to move the drone to the wind shelter location in wind shelter mode.

[0085] Among the selected sheltering locations, determine the location closest to the construction location, which is the sheltering location.

[0086] The shelter power is determined by inputting the ambient wind direction and wind speed values ​​into a preset shelter power database. This database is pre-configured by technicians based on actual conditions and contains the relationship between the ambient wind direction, wind speed, and shelter power. If the direction of movement differs from the ambient wind direction, the greater the ambient wind speed, the greater the required shelter power. If the direction of movement and the ambient wind direction are the same, the greater the ambient wind speed, the less required shelter power. The actual parameters are pre-configured by technicians based on actual conditions and are not detailed here.

[0087] Step S213: Control the UAV to switch to the wind avoidance mode and move to a wind avoidance position at the wind avoidance power to perform inspection.

[0088] Control the drone to switch to wind-avoidance mode and move it to a wind-avoidance location with wind-avoidance power for inspection.

[0089] The flight mode also includes a strong mode. After step S211, the flight mode matching method further includes: Step S2110: When the spatial volume of the sheltering position is not greater than the reference volume, the wind force difference between the ambient wind force value and the reference wind force value is calculated.

[0090] Wind differential refers to the difference between the ambient wind value and the reference wind value. It is calculated by subtracting the reference wind value from the ambient wind value. A larger wind differential indicates a stronger wind.

[0091] When the spatial volume of the avoidance position is not greater than the reference volume, it means that there is no place for the drone to avoid, and it is necessary to conduct an inspection at the original position to calculate the wind difference.

[0092] Step S2111: Obtain strong power based on wind force difference matching.

[0093] Powerful power refers to the power of the drone in powerful mode.

[0094] The greater the wind force difference is, the greater the required high power is. The relationship between the high power and the wind force difference is stored in advance, so the corresponding high power can be obtained according to the wind force difference.

[0095] Step S2112: Obtain the force angle based on the ambient wind direction and the construction location.

[0096] The force angle refers to the tilt direction of the drone when the drone is in force mode.

[0097] The ambient wind direction is opposite to the direction of the strong angle. The farther the construction location is, the larger the strong angle is. The size relationship of the strong angle corresponding to the construction location is pre-stored, so the corresponding strong angle size can be obtained according to the construction location.

[0098] Step S2113: Control the drone to switch to the strong mode and perform inspections with strong power and strong angle.

[0099] Powerful mode refers to the mode in which the drone operates at high power in strong winds.

[0100] Control the drone to switch to strong mode, and use strong power and strong angle to make the drone inspect the construction location at its original position. The motors in different positions can be controlled to output different power to keep the drone hovering.

[0101] The flight mode also includes a parking mode. After step S20, the flight mode matching method further includes: Step S201: A preset parking feature is selected from the construction site image information, and the selected position is defined as the landing position.

[0102] The parking features refer to image features in the image that are suitable for drone landing, such as images of positions such as brackets and platforms. They are pre-set by technicians based on actual conditions and will not be described in detail here.

[0103] Landing location refers to the location where the drone can land.

[0104] By feeding a large number of images with stop features into the YOLO model, the Pytorch network architecture is used to repeatedly stack the images to extract the stop features. The error between the results and the data is calculated and used when the error is less than 1%. Construction site imagery is fed into the model. When a stop feature is identified, it is marked and boxed out from the construction site imagery. The boxed location is the landing location.

[0105] Step S202: obtaining an inspection direction and an inspection distance based on the construction position and the landing position; and removing landing positions whose inspection distances do not fall within a preset identification interval based on the inspection direction.

[0106] Inspection direction refers to the direction in which the drone inspects the construction site from all landing locations, and is a collection of multiple directions.

[0107] The inspection distance refers to the distance between the drone's construction location and landing location. It is a collection of multiple distance values, calculated by calculating the distance from all landing locations to the construction location.

[0108] The identification interval refers to the preset inspection distance range, which means that the supervision image will not be unclear or out of focus within the range. It is pre-set by technical personnel based on actual conditions and will not be elaborated here.

[0109] Based on the inspection direction, the landing positions whose inspection distance does not fall within the identification interval are removed, thereby ensuring that the drone can collect clear images when supervising the construction location.

[0110] Step S203: determining the height coordinates of all objects between the construction position and the landing position after removal based on the construction site image information, and removing positions where the height coordinates of objects are higher than the landing position.

[0111] The height coordinates refer to the height positions of all objects between the construction location and the landing location after removal.

[0112] The construction site imagery includes a preset reference object. By detecting the edge of the reference object and calculating its pixel width, and since the actual length of the reference object is known, the height coordinates of all objects can be inferred based on the proportional relationship between the pixels and the actual length. The reference object and the target must be in the same plane during shooting, and the camera position must be fixed.

[0113] The positions where the height coordinates of objects are higher than the landing position are removed from the construction site image information to prevent the drone from being unable to monitor the construction situation due to obstruction by the objects. The positions after removal are the positions where the drone can normally recognize the helmet image after it stops.

[0114] Step S204: Collect the ambient wind direction and the drone position, match the flight power required to move from the drone position to the remaining landing positions based on the ambient wind force value and the ambient wind direction, select the landing position with the lowest power from the matched flight powers, and define it as the parking position.

[0115] Flight power refers to the power required for the drone to fly to its parking position.

[0116] Flight power is determined by entering the ambient wind direction and wind speed into a pre-set flight power database. This database is pre-configured by technicians based on actual conditions and contains the relationship between ambient wind direction, wind speed, and flight power. If the direction of travel differs from the ambient wind direction, the higher the ambient wind speed, the higher the required flight power. If the direction of travel and the ambient wind direction are the same, the higher the ambient wind speed, the lower the required flight power. The actual parameters are pre-configured by technicians based on actual conditions and are not detailed here.

[0117] The landing position with the lowest power is selected from the matched flight powers. This position is the most energy-saving and is defined as the parking position.

[0118] Step S205: Control the UAV to switch to the parking mode and move to the parking position at the flight power to land and perform inspection.

[0119] Parking mode refers to the mode in which the drone lands at a parking position to reduce consumption. At this time, the propellers of the drone do not rotate, and the mechanical claws at the bottom of the drone will grasp objects at the parking position, such as poles and brackets.

[0120] Control the drone to switch to parking mode, move it to the parking position with flight power, land it, and shoot with the camera to complete the inspection work.

[0121] After step S213, the method further includes: Step S214: When moving to the parking position, continuously collecting parking image information of the parking position.

[0122] The stop image information refers to the image data of the stop position, which is a collection of images collected multiple times continuously.

[0123] When moving to the parking position, the camera on the drone that does not shoot the helmet continuously captures and collects parking image information of the parking position.

[0124] Step S215: identifying a preset reference static feature in the stationary image information using a preset reference static feature set.

[0125] The reference static features refer to the static feature images in the construction site, such as thick tree trunks, brackets fixed on the concrete floor, etc., which are pre-set by technical personnel based on actual conditions and will not be described in detail here.

[0126] The reference static feature set refers to a set of static feature images in the construction site, which is pre-set by technical personnel based on actual conditions and will not be described in detail here.

[0127] By feeding a large number of images with baseline static features into the YOLO model, and then repeatedly stacking the images using a PyTorch network architecture, a model capable of identifying the baseline static features from stationary images was developed. A database with an error of less than 1% was then used. By feeding the stationary images into the trained database, features consistent with those in the baseline static feature set were identified and marked. These marked features became the baseline static features.

[0128] Step S216: Based on the reference static feature, the stationary image information is sequentially compared, and the distance difference between the reference static feature and the stationary position is sequentially calculated.

[0129] The distance difference value refers to a set of distance differences between the reference stationary feature and the parked position, and is obtained by calculating the distance difference between the reference stationary feature and the parked position.

[0130] The information of two adjacent stationary images is compared in sequence, and the distance difference between the reference static feature and the stationary position is calculated in sequence. Whether there is jitter is determined by whether the value before and after the distance difference changes.

[0131] Step S2160: When the distance difference is equal to the preset reference distance difference, the UAV continues to move to the parking position at the flight power, lands, and conducts inspection.

[0132] The reference distance difference refers to the standard distance difference between the reference stationary feature and the parking position, which is preset by technicians based on actual conditions and will not be described in detail here.

[0133] When the distance difference is equal to the preset reference distance difference, it means that the parking position has not shaken, and the drone continues to move to the parking position at the flight power and lands for inspection.

[0134] In step S2161, when the distance difference is not equal to the reference distance difference, the distance differences are arranged in order of size, and the maximum distance difference is selected.

[0135] The maximum value refers to the maximum or minimum value among the distance differences.

[0136] When the distance difference is not equal to the reference distance difference, it indicates that there is jitter in the parking position. The distance differences are arranged in order of size, and the maximum distance difference is selected.

[0137] Step S21610: Obtain the adjustment amplitude of the UAV based on the maximum value matching of the distance difference.

[0138] The adjustment amplitude refers to the vibration amplitude of the drone, which is controlled by the vibrator and propeller preset in the drone.

[0139] The adjustment range is determined by inputting the maximum distance difference value into a preset adjustment range database. This database is pre-configured by a technician based on actual conditions and contains the relationship between the maximum distance difference value and the adjustment range. The larger the maximum distance difference value, the larger the adjustment range. The actual parameters are pre-configured by the technician based on actual conditions and are not detailed here.

[0140] Step S21611: Obtaining the adjustment direction of the drone based on the ambient wind direction.

[0141] Adjustment direction refers to the direction in which the drone is adjusted.

[0142] The adjustment direction of the drone is opposite to the ambient wind direction.

[0143] Step S21612: Control the drone to move to the parking position and land, and conduct inspection after adjusting the amplitude and direction.

[0144] The drone is controlled to move to its parking position and land, adjusting its amplitude and direction to prevent vibration from the parking position from affecting the drone's inspection. When the wind pushes the drone and its parking position to the left, the vibrator and propeller control the drone to tilt to the right, keeping the camera facing the helmet and maintaining stability. The vibrator and propeller control methods are set by the staff based on actual conditions and are not detailed here.

[0145] After landing at the parking position, it also includes: Step S217: collecting the balance angle difference of the UAV at the parking position.

[0146] The balance angle difference refers to the tilt angle of the drone when it is in the parked position.

[0147] The data is collected through the angle sensor preset on the drone. The angle sensor is pre-set by technicians according to actual conditions and will not be described in detail here.

[0148] Step S2170: If the balance angle difference is 0, control the drone to perform inspection.

[0149] A balance angle difference of 0 means the drone is in a completely level state.

[0150] If the balance angle difference is 0, it means that there is no shaking at the parking position, and the drone can continue to be controlled for inspection.

[0151] Step S2171: If the balance angle difference is not 0, the balance angle differences are arranged in order of magnitude, and the maximum balance angle difference is selected.

[0152] The maximum value refers to the maximum or minimum value of the balance angle difference.

[0153] If the balance angle difference is not 0, it means that there is jitter in the parking position. The balance angle differences are arranged in order of size, and the maximum balance angle difference is selected.

[0154] Step S21710: Obtain the jitter amplitude of the parking position based on the maximum value matching of the balance angle difference.

[0155] The jitter amplitude refers to the degree of shaking of the drone at the parked position.

[0156] The jitter amplitude is determined by inputting the maximum balance angle difference into a preset jitter amplitude database. This database is pre-configured by technicians based on actual conditions and contains the relationship between the maximum balance angle difference and the jitter amplitude. The greater the maximum balance angle difference, the greater the jitter amplitude. The actual parameters are pre-configured by technicians based on actual conditions and are not detailed here.

[0157] Step S21711: Obtaining a matching amplitude corresponding to the drone based on the jitter amplitude, and controlling the drone to adjust at the parking position using the matching amplitude.

[0158] Matching amplitude refers to the amplitude used to adjust the balance of the drone at the parked position. The balance amplitude is synchronously controlled by the vibrator and propeller, which will not be explained here. Matching amplitude is exactly the opposite of jitter amplitude.

[0159] The drone is controlled to adjust at the parking position with matching amplitude to eliminate the interference of vibration on inspection and supervision.

[0160] When multiple drones are flying synchronously on a construction site, there will be interactions between the drones. The interaction method includes the following steps: Step S50: Determine the personnel position and edge position based on the construction site image information. When the personnel position falls into the edge position, collect the inspection position of the drone. The inspection position includes the target inspection position and the active inspection position.

[0161] The edge position refers to the boundary position of the image in the construction site image information. The personnel position is the location of the helmet information, which is determined by identifying the helmet features in the construction site image information.

[0162] The inspection position refers to the location of the drone during inspection, which is sent by the GPS chip preset on the drone. The GPS chip is pre-set on the drone by technicians based on actual conditions and will not be described in detail here.

[0163] The target inspection position refers to the location of the target drone that needs to be moved. The active inspection position refers to the location of other drones except the target drone.

[0164] When the personnel position falls into the edge position, it means that the personnel needs to move, so the movement of the drone needs to be controlled to prevent the personnel from being out of the supervision range, and the inspection position is collected through the GPS chip.

[0165] Step S51: Calculate the drone distance between the target inspection position and the active inspection position.

[0166] Drone distance refers to the distance between the target inspection location and the active inspection location.

[0167] The distance between the target inspection position and the active inspection position is the drone distance. For example, if the target inspection position is (X2, Y2) and the active inspection position is (X1, Y1), then the drone distance L = √((X1-X2) 2 + (Y1-Y2) 2 ).

[0168] Step S520: When the distance between the UAV and the human is less than the preset collision distance and the ambient wind value is not greater than the reference wind value, the moving direction of the human is determined based on the human position and the edge position.

[0169] The collision distance refers to the safe distance for judging whether the drone will collide. It is pre-set by technicians based on actual conditions and will not be explained here.

[0170] The movement direction of personnel refers to the movement direction of construction personnel, which is the direction from the personnel position to the edge position.

[0171] When the distance between the drones is less than the preset collision distance and the ambient wind value is not greater than the baseline wind value, it means that the target drone will collide with surrounding drones when moving. In the absence of wind, the direction of the person's movement is determined based on the person's position and the edge position.

[0172] Step S5200: obtaining an interaction position based on matching of the active inspection position and the target inspection position with the movement direction of the personnel.

[0173] The interaction location refers to the location where the target UAV interacts with other UAVs.

[0174] A coordinate system is constructed based on the active inspection position and the target inspection position. The target inspection position moves synchronously with the direction of the personnel's movement. The active inspection position is defined as a fixed position, and the target inspection position moves closer to the active inspection position. When the two are within a preset safety distance, the coordinates are considered the interaction position, where the two drones interact. The safety distance is pre-set by technicians based on actual conditions and is not detailed here.

[0175] Step S5201: Obtain an interaction path based on the interaction position, and control the drone to interact using the interaction path.

[0176] An interaction path is the path along which drones interact. Once the target drone reaches the interaction location, the interaction path allows other drones to avoid the target drone, allowing them to fly relative to each other. This avoidance strategy involves other drones flying downward while the target drone flies upward, allowing the two drones to fly relative to each other.

[0177] After step S51, the method further includes: Step S521: When the distance between the UAV and the vehicle is less than the preset collision distance and the ambient wind value is greater than the reference wind value, a wind-shielding interactive position is selected from the avoidance positions based on the target inspection position and the active inspection position.

[0178] A wind-sheltered interaction location refers to a location where interaction is possible in strong winds. Since there are certain risks when drones interact in strong winds, it is recommended to seek locations with relatively low winds to improve safety.

[0179] Taking the target inspection position and the active inspection position as two points, the position where the sum of the distances to the target inspection position and the active inspection position in the avoidance position is the shortest is the sheltering interaction position.

[0180] When the distance between the drones is less than the preset collision distance and the ambient wind value is greater than the baseline wind value, it means that the target drone will collide with the surrounding drones when moving. Under windy conditions, based on the target inspection position and the active inspection position, a windproof interaction position is selected from the avoidance positions.

[0181] Step S5210: Match the interaction power of the UAV based on the ambient wind value and the wind shelter interaction position.

[0182] Interaction power refers to the power of the drone during interaction.

[0183] The interactive power is determined by matching the ambient wind speed and the wind shelter interaction position into a preset interactive power database. This database is pre-configured by technicians based on actual conditions and contains the relationship between ambient wind speed, wind shelter interaction position, and interactive power. The greater the ambient wind speed, the greater the interactive power required to counteract the ambient wind. The farther the wind shelter interaction position is, the greater the required travel distance, resulting in a greater interactive power. The actual parameters are pre-configured by technicians based on actual conditions and are not detailed here.

[0184] Step S5211: Matching the wind avoidance interaction path of the UAV based on the personnel position and the wind avoidance interaction position.

[0185] The wind-avoiding interaction path refers to the path along which the drone interacts in strong winds.

[0186] A wind-avoidance interaction path is one in which, after the target drone reaches the wind-avoidance interaction position, other drones avoid the target drone, thereby achieving relative flight. This mutual avoidance strategy involves other drones flying downward while the target drone flies upward, allowing the two drones to fly relative to each other. While in the wind-avoidance interaction position, the target drone can still recognize the helmet's features.

[0187] Step S5212: Control the UAV to complete the interaction at the wind shelter interaction position using the wind shelter interaction path and wind shelter power, and then conduct an inspection.

[0188] Wind avoidance power refers to the power of the drone in wind avoidance mode.

[0189] Control the drone to interact at the windproof interaction position using the windproof interaction path and windproof power to avoid the impact of strong wind on the interactive operation. After the interaction is completed, conduct inspections.

[0190] If there is no wind-shielding interaction position, the UAV is controlled to interact with the wind-shielding power at twice the safety distance using the method in step S5201.

[0191] Based on the same inventive concept, an embodiment of the present invention provides a construction supervision and inspection system based on a drone, comprising: An acquisition module is used to obtain environmental wind force values, construction site image information, and construction site audio information; A memory for storing a program for a construction supervision and inspection method based on a drone; The processor loads and executes the program in the memory.

[0192] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiment. All technical solutions based on the concept of the present invention are within the scope of protection of the present invention. It should be noted that for those skilled in the art, various improvements and modifications that do not depart from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A construction supervision and inspection method based on drones, characterized in that: include: S1: responding to a trigger instruction to collect ambient wind force values; S2: Collect construction site image information and match the flight mode according to the ambient wind value; S3: selecting helmet information from the construction site image information based on preset helmet features; S40: When the helmet information is inconsistent with the preset reference helmet information, an alarm is issued and the inspection continues; S41: When the helmet information is consistent with the preset reference helmet information, collecting the construction site audio information; S410: If the construction site audio information is consistent with the preset reference audio information, the matching is completed and the inspection continues; S411: If the construction site audio information is inconsistent with the preset baseline audio information, an alarm is issued and the drone is controlled to continue the inspection.

2. A construction supervision and inspection method based on drone according to claim 1, characterized in that: After step S410, the following steps are also included: S4100: Collect the tool distance difference between the helmet and the preset tool and the tool's movement frequency; S4101: Matching a reference amplitude curve based on helmet information; S4102: Obtaining a tool distance variation amplitude based on the tool distance difference and the motion frequency; S4103: Generate a construction amplitude curve during tool construction based on distance variation fitting; S41030: When the construction amplitude curve is consistent with the reference amplitude curve, the matching is completed and the inspection continues; S41031: When the construction amplitude curve is inconsistent with the reference amplitude curve, an alarm is issued and the drone is controlled to continue inspection.

3. The construction supervision and inspection method based on drone according to claim 1 is characterized in that: The flight mode includes a wind avoidance mode and a hovering mode. The flight mode matching method in step S2 includes: S20: When the drone is performing an inspection, it is determined whether the ambient wind value is greater than a preset reference wind value; S200: When the ambient wind force value is not greater than the reference wind force value, the drone is controlled to switch to the hovering mode and perform inspection at the preset reference power; S210: When the ambient wind value is greater than the reference wind value, collecting the ambient wind direction and determining a shelter position within the construction site based on the construction site image information and the ambient wind direction; S211: identifying the spatial volume of the avoidance position based on the avoidance position, and selecting a avoidance position that is larger than a preset reference volume; S212: Determine a wind shelter position from the selected shelter positions based on the preset construction position, and match the wind shelter power of the UAV according to the ambient wind direction and the ambient wind force value; S213: Control the UAV to switch to the wind avoidance mode and move it to a wind avoidance location at the wind avoidance power for inspection.

4. A construction supervision and inspection method based on drone according to claim 3, characterized in that: The flight mode also includes a strong mode. After step S211, the flight mode matching method further includes: S2110: When the spatial volume of the sheltering position is not greater than the reference volume, the wind force difference between the ambient wind force value and the reference wind force value is calculated; S2111: Get strong power based on wind force difference matching; S2112: Obtain the force angle based on the ambient wind direction and construction location; S2113: Control the drone to switch to strong mode and conduct inspections with strong power and strong angle.

5. The construction supervision and inspection method based on drone according to claim 3 is characterized in that: The flight mode also includes a parking mode. After step S20, the flight mode matching method further includes: S201: Selecting a preset parking feature from the construction site image information, and defining the selected position as a landing position; S202: Obtaining an inspection direction and an inspection distance based on the construction position and the landing position; removing landing positions whose inspection distances do not fall within a preset identification interval based on the inspection direction; S203: determining the height coordinates of all objects between the construction position and the landing position after removal based on the construction site image information, removing positions where the height coordinates of objects are higher than the landing position; S204: Collecting the ambient wind direction and the UAV position, matching the flight power required to move from the UAV position to the remaining landing positions based on the ambient wind force and direction, and selecting the landing position with the lowest power from the matched flight powers and defining it as the parking position; S205: Control the UAV to switch to the parking mode and move it to the parking position at the flight power to land and conduct inspection.

6. The construction supervision and inspection method based on drone according to claim 1 is characterized in that: After step S213, the method further includes: S214: When moving toward the parking position, continuously collecting parking image information of the parking position; S215: identifying a preset reference static feature in the stationary image information using a preset reference static feature set; S216: Based on the reference static feature, the stationary image information is sequentially compared, and the distance difference between the reference static feature and the stationary position is sequentially calculated; S2160: When the distance difference is equal to the preset reference distance difference, the UAV continues to move at the flight power to the parking position, lands, and conducts inspection; S2161: When the distance difference is not equal to the reference distance difference, the distance differences are arranged in order of magnitude, and the maximum distance difference is selected; S21610: Obtain the adjustment range of the drone based on the maximum value matching of the distance difference; S21611: Get the adjustment direction of the drone based on the ambient wind direction; S21612: Control the drone to move to the parking position and land, and then conduct inspection after adjusting the amplitude and direction.

7. A construction supervision and inspection method based on drone according to claim 6, characterized in that: After landing at the parking position, it also includes: S217: collecting the balance angle difference of the UAV at the parking position; S2170: If the balance angle difference is 0, control the drone to perform inspection; S2171: If the balance angle difference is not 0, the balance angle differences are arranged in order of magnitude, and the maximum balance angle difference is selected; S21710: Obtaining a jitter amplitude at the parking position based on the maximum value matching of the balance angle difference; S21711: Obtaining a matching amplitude corresponding to the drone based on the jitter amplitude, and controlling the drone to adjust at the parking position using the matching amplitude.

8. The construction supervision and inspection method based on drone according to claim 3 is characterized in that: Also includes: S50: Determine the personnel position and the edge position based on the construction site image information, and when the personnel position falls into the edge position, collect the inspection position of the drone, the inspection position includes the target inspection position and the active inspection position; S51: Calculate the drone distance between the target inspection position and the active inspection position; S520: When the distance between the UAV and the human is less than the preset collision distance and the ambient wind value is not greater than the reference wind value, determining the movement direction of the human based on the human position and the edge position; S5200: Obtaining an interaction position based on the active inspection position and the target inspection position by matching the personnel movement direction; S5201: Obtain an interaction path based on the interaction position, and control the UAV to interact using the interaction path.

9. The construction supervision and inspection method based on drone according to claim 8 is characterized in that: After step S51, the method further includes: S521: When the distance between the UAV and the target inspection position and the active inspection position is less than the preset collision distance and the ambient wind value is greater than the reference wind value, a wind-shielding interactive position is selected from the avoidance positions based on the target inspection position and the active inspection position; S5210: Match the interactive power of the UAV based on the ambient wind value and the wind shelter interactive position; S5211: Matching the UAV’s wind avoidance interaction path based on the personnel’s position and wind avoidance interaction position; S5212: Control the UAV to complete the interaction at the wind avoidance interaction position using the wind avoidance interaction path and wind avoidance power, and then conduct an inspection.

10. A construction supervision and inspection system based on drones, characterized in that: include: An acquisition module is used to obtain environmental wind force values, construction site image information, and construction site audio information; A memory for storing a program of a construction supervision and inspection method based on a drone according to any one of claims 1 to 9; The processor loads and executes the program in the memory.

Images