Touch and visual detection device and method for tall tower hidden danger detection unmanned aerial vehicle

Abstract

The invention relates to the technical field of defect detection, in particular to a tall tower hidden danger detection unmanned aerial vehicle touch vision detection device which comprises an unmanned aerial vehicle, an external vision sensor and a touch vision device. The unmanned aerial vehicle carries the touch vision device and the external vision sensor to fly up and down, left and right; the touch vision device is a touch vision sensor and is composed of a sensing part, a thin film skin part and a skin fixing ring, and when the touch vision sensor is used, the surface of the touch vision sensor makes contact with a detection surface, the texture of the detection surface is fitted out, and whether flaws exist or not is detected; the external visual sensor directly acquires image data of the detection surface to obtain obvious flaw data of the detection part; through combination of the unmanned aerial vehicle and the touch sensing device, the problems of high cost, safety risk, poor environmental adaptability and the like in a traditional detection method are solved, the detection efficiency and accuracy are improved, and a reliable and economical solution is provided for safety maintenance of the tall tower.

Description

Technical Field

[0001] The present invention relates to the technical field of defect detection, and in particular to a drone tactile-visual detection device and method for detecting hidden dangers of high-rise towers. Background Art

[0002] Tall tower structures may experience vibration and shaking hazards during long-term use. If these hazards are not detected and handled in time, they may pose a serious threat to the safety of the structure. At present, most detection methods for tall towers rely on manual detection. This method requires workers to climb the tower body for inspection. It is not only costly and inefficient, but also has great safety risks and poses a potential threat to the personal safety of operators. Existing methods mostly use pure visual methods for detection. Visual detection uses photographic equipment to capture images of the surface of the tall tower, and then processes the images. The corresponding algorithms are used to detect whether there are gaps, cracks and other hidden dangers in the tall tower. At this stage, visual detection of surface defects is widely used, not only for tall towers, but also for roads. In cases such as surface crack detection, glass bottle defects, and material surface scratch detection, the quality of the data set obtained for defect detection is not high. Vision is often used in the actual production process. Due to changes in lighting conditions, object surface characteristics, shooting angles and other factors, the acquired images may have problems such as noise and blur. In addition, defects have various forms, and their size, color, texture and other characteristics vary greatly, which brings challenges to data annotation and model training. In addition, visual technology is greatly affected by the environment. When the environment is dark, effective fill light is required, and it is difficult to perform detection in haze or other low visibility conditions. Therefore, how to obtain clear and accurate image data without being affected by the environment and effectively process the data is an important issue facing defect detection.

[0003] At present, the research on tactile sensing technology is mainly focused on tasks in quasi-static processes. For example, when a robot performs relatively static operations such as grasping and carrying, it uses tactile sensors to obtain information such as the shape and hardness of an object. However, only a few researchers have applied tactile sensing to dynamic scenarios, such as the estimation of liquid properties and robot swinging. In these studies, tactile sensors are used to detect and analyze the viscosity, density and other properties of liquids, or to perceive changes in the external environment when the robot moves quickly. However, in the field of mobile robots, especially in drones, their application is still very limited. Although drones have broad application prospects, the technology of using tactile sensors to perceive and adapt to the environment during flight is still in its infancy and still needs further exploration and development.

[0004] Therefore, in order to address the problem that workers need to personally climb the tower to conduct inspections, which is not only costly and inefficient, but also poses great safety risks and poses a potential threat to the personal safety of operators, an innovative detection device and process method combining a drone and a tactile sensor device can be designed. Summary of the invention

[0005] In order to overcome the problem that workers need to climb the tower in person to conduct inspections, it is not only costly and inefficient, but also poses great safety risks and poses a potential threat to the personal safety of operators.

[0006] The technical solution of the present invention is: a drone touch-vision detection device for high-rise tower hidden danger detection, comprising a drone, an external visual sensor and a touch-vision device;

[0007] The drone includes two fixed landing gears, two clockwise propellers, two counterclockwise propellers, a laser radar, a GPS module, a camera module that can transmit images in real time, and a touch-vision module. The entire drone flies up, down, left, right, forward, and backward with the touch-vision device and external vision sensors.

[0008] The touch vision device is a touch vision sensor, which consists of three parts: a sensing part, a film skin part and a skin fixing ring. Gas is filled in from the inflation port at the rear to swell the film skin part. At the same time, the sensing part and the skin fixing ring work together to fix the bulging shape of the film skin part in a suitable position. The concentric ring in the middle of the sensing part is a built-in camera, and the outer circle of the ring is an LED light. The film skin part uses an elastic film with printed marks as the touch vision sensor surface, and an air pump is used to make it in a convex fixed shape. When the touch vision sensor is used, the touch vision sensor surface contacts the detection surface, and the point cloud data of the contact surface is generated by the algorithm, so as to fit the texture of the detection surface and detect whether there are defects.

[0009] The external visual sensor directly acquires the image data of the inspection surface and obtains the obvious defect data of the inspection part.

[0010] Preferably, the haptic-vision device uses an RGB matching method.

[0011] Preferably, the LED light adopts an RGB light strip. When the touch-vision device is in use, the RGB light strip illuminates the surface of the touch-vision sensor, and the built-in camera captures images and collects data.

[0012] Preferably, the elastic film is an ultra-thin elastic silicone film or a gel pad.

[0013] Preferably, when the touch vision sensor surface contacts an external object, the elastic film on the touch vision sensor surface will deform, reflecting the 3D shape and texture of the contact surface. At the same time, the printed mark will be displaced, and the force and texture data applied to the touch vision sensor surface can be obtained based on this displacement.

[0014] Preferably, both the external vision sensor and the tactile vision sensor use point cloud technology.

[0015] Preferably, the external visual sensor performs feature extraction through the YOLO series algorithm to obtain obvious defect data of the detection part.

[0016] A method for detecting hidden dangers of high-rise towers by using a drone for tactile vision detection includes the above-mentioned device for detecting hidden dangers of high-rise towers by using a drone for tactile vision detection, and the steps are as follows:

[0017] Step 1: Start the device

[0018] First, start the tactile sensing device, external visual sensor and drone to ensure that all equipment is operating normally and ready to enter the detection state;

[0019] Step 2: Operate the drone

[0020] Maneuver the drone to fly next to the tower, and use the flexibility and maneuverability of the drone to accurately control its position so that it can approach various inspection locations of the tower;

[0021] Step 3: Touch the tower surface with the touch sensor

[0022] Maneuver the drone to a distance where the tactile sensor can touch the surface of the towering tower, ensuring that the surface of the tactile sensor is in close contact with the tower's external structure to obtain accurate tactile data;

[0023] Step 4: Data Collection

[0024] When the touch-vision sensor is in contact with the tower surface, the tactile sensing device will detect the mechanical data and texture data of the contact surface in real time. These data include the vibration and shaking of the tower surface and the subtle changes in the surface texture. In addition, while the tactile detection is in progress, the external visual sensor on the drone can be used for visual detection to obtain the image data of the tower surface. In this way, two sets of data about the tower structure status can be obtained at the same time - one is tactile data and the other is visual data.

[0025] Step 5: Data processing and analysis

[0026] The collected data is transmitted to the ground control station or cloud server for processing. By analyzing the mechanical data and texture data, it can be identified whether the tall tower has the risk of vibration or shaking. Combined with the data from visual inspection, the inspection results can be further verified and refined to ensure the accuracy and reliability of the inspection.

[0027] Beneficial effects of the present invention:

[0028] 1. The combination of drones and tactile sensing devices solves the problems of high cost, safety risks and poor environmental adaptability in traditional detection methods. This method not only improves detection efficiency and accuracy, but also provides a reliable and economical solution for the safe maintenance of tall towers.

[0029] 2. Using drones as detection platforms eliminates the need for manual climbing of tall towers, greatly reducing labor costs and safety risks. Drones are flexible and maneuverable and can quickly reach various parts of the tower for detection, thereby improving the speed and efficiency of detection. Drones can also cover a wider area to ensure the comprehensiveness and accuracy of detection.

[0030] 3. Apply tactile sensors to dynamic scenes and perform detection through contact with the surface of tall towers. This method is less affected by ambient light and can work effectively even in dark or low-visibility environments. Tactile sensors can detect subtle structural changes and physical properties to achieve a detailed assessment of the tower structure. In addition, through different types of sensors, for the details that need to be detected, sensors with smaller surfaces can be used, and RGB matching methods can be used for fine detection. This method can capture subtle changes and is suitable for checking details such as tiny cracks and corrosion spots in the tower structure. For large areas that need to be detected, sensors with larger surfaces can be used, and point cloud technology can be used for detection. Point cloud technology can quickly scan and generate three-dimensional models, conduct a comprehensive assessment of the overall condition of large areas, and can be customized for different detection needs to ensure the accuracy and comprehensiveness of the detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 Shown is a first three-dimensional structural schematic diagram of a drone touch-vision detection device for high-rise tower hidden danger detection of the present invention;

[0032] Figure 2 Shown is a schematic diagram of the three-dimensional structure of the touch-vision device in the touch-vision detection device for high-rise tower hidden danger detection drone of the present invention;

[0033] Figure 3 Shown is a schematic diagram of the three-dimensional structure of the sensing part of the high-rise tower hidden danger detection drone touch-vision detection device of the present invention;

[0034] Figure 4 Shown is a flow chart of a method for collecting data using a touch-vision device in a touch-vision detection method for detecting hidden dangers of a high-rise tower using an unmanned aerial vehicle;

[0035] Figure 5 Shown is a flow chart of the external visual sensor data collection method in the high-rise tower hidden danger detection drone visual detection method of the present invention.

[0036] Explanation of the reference numerals: 1. touch-vision device; 2. drone; 3. external vision sensor; 101. sensing part; 102. film skin part; 103. skin fixing ring; 1011. built-in camera; 1012. LED light. DETAILED DESCRIPTION

[0037] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0038] See also Figure 1-Figure 5 , the present invention provides an embodiment: a high-rise tower hidden danger detection drone touch-vision detection device, including a drone 2, an external visual sensor 3 and a touch-vision device 1;

[0039] The drone 2 includes two fixed landing gears, two clockwise propellers, two counterclockwise propellers, a laser radar, a GPS module, a camera module capable of transmitting images in real time, and a touch-vision module. The entire drone 2 flies up, down, left, right, forward, and backward with the touch-vision device 1 and the external vision sensor 3.

[0040] The touch vision device 1 is a touch vision sensor, which consists of three parts: a sensing part 101, a film skin part 102 and a skin fixing ring 103. Gas is filled in from the inflation port at the rear to swell the film skin part 102. At the same time, the sensing part 101 and the skin fixing ring 103 work together to fix the swollen shape of the film skin part 102 in a suitable position. The concentric ring in the middle of the sensing part 101 is a built-in camera 1011, and the outer circle of the ring is an LED light 1012. The film skin part 102 uses an elastic film with printed marks as the touch vision sensor surface, and an air pump is used to make it in a convex fixed shape. When the touch vision sensor is used, the touch vision sensor surface contacts the detection surface, and the point cloud data of the contact surface is generated by an algorithm, so as to fit the texture of the detection surface and detect whether there are defects.

[0041] The external visual sensor 3 directly acquires the image data of the detection surface and obtains the obvious defect data of the detection part.

[0042] Preferably, the touch-vision device 1 uses an RGB matching method.

[0043] Preferably, the LED lamp 1012 uses an RGB light strip. When the touch-vision device is in use, the RGB light strip illuminates the surface of the touch-vision sensor, and the built-in camera 1011 captures images and collects data.

[0044] Preferably, the elastic film is an ultra-thin elastic silicone film or a gel pad.

[0045] Preferably, when the touch vision sensor surface contacts an external object, the elastic film on the touch vision sensor surface will deform, reflecting the 3D shape and texture of the contact surface. At the same time, the printed mark will be displaced, and the force and texture data applied to the touch vision sensor surface can be obtained based on this displacement.

[0046] Preferably, both the external vision sensor 3 and the touch vision sensor use point cloud technology.

[0047] Preferably, the external visual sensor 3 performs feature extraction through the YOLO series algorithm to obtain obvious defect data of the detection part.

[0048] A method for detecting hidden dangers of high-rise towers by using a drone for tactile vision detection includes the above-mentioned device for detecting hidden dangers of high-rise towers by using a drone for tactile vision detection, and the steps are as follows:

[0049] Step 1: Start the device

[0050] First, start the tactile sensing device 1, the external visual sensor 3 and the drone 2 to ensure that all devices are operating normally and ready to enter the detection state;

[0051] Step 2: Operate the drone 2

[0052] The drone 2 is operated to fly next to the tall tower, and the flexibility and maneuverability of the drone 2 are used to accurately control its position so that it can approach various inspection locations of the tall tower;

[0053] Step 3: Touch the tower surface with the touch sensor

[0054] The drone 2 is operated to a distance where the tactile sensor can touch the surface of the towering tower, ensuring that the surface of the tactile sensor is in close contact with the external structure of the tower, the RGB light bar illuminates the surface of the tactile sensor, and the built-in camera 1011 captures images and collects data, thereby obtaining accurate tactile data;

[0055] Step 4: Data Collection

[0056] When the touch-visual sensor is in contact with the tower surface, the tactile sensing device 1 detects the mechanical data and texture data of the contact surface in real time. These data include the vibration and shaking of the tower surface and the subtle changes in the surface texture. In addition, while the tactile detection is in progress, the external visual sensor 3 on the drone 2 can be used to perform visual detection to obtain image data of the tower surface. In this way, two sets of data on the tower structure status can be obtained at the same time - one is tactile data and the other is visual data.

[0057] Step 5: Data processing and analysis

[0058] The collected data is transmitted to the ground control station or cloud server for processing. By analyzing the mechanical data and texture data, it can be identified whether the tall tower has the risk of vibration or shaking. Combined with the data from visual inspection, the inspection results can be further verified and refined to ensure the accuracy and reliability of the inspection.

[0059] When working, the steps are as follows:

[0060] Step 1: Start the device

[0061] First, start the tactile sensing device 1, the external visual sensor 3 and the drone 2 to ensure that all devices are operating normally and ready to enter the detection state;

[0062] Step 2: Operate the drone 2

[0063] The drone 2 is operated to fly next to the tall tower, and the flexibility and maneuverability of the drone 2 are used to accurately control its position so that it can approach various inspection locations of the tall tower;

[0064] Step 3: Touch the tower surface with the touch sensor

[0065] Drone 2 is operated to a distance where the tactile sensor can touch the surface of the tall tower, ensuring that the surface of the tactile sensor is in close contact with the outer structure of the tower, thereby obtaining accurate tactile data;

[0066] Step 4: Data Collection

[0067] When the touch-vision sensor is in contact with the tower surface, the elastic film on the touch-vision sensor surface will be deformed, reflecting the 3D shape and texture of the contact surface. At the same time, the printed mark will be displaced. The tactile sensing device 1 will detect the mechanical data and texture data of the contact surface in real time. These data include the vibration and shaking of the tower surface and the subtle changes in the surface texture. In addition, while the tactile detection is being performed, the external visual sensor 3 on the drone 2 can be used for visual detection to obtain the image data of the tower surface. In this way, two sets of data on the state of the tower structure can be obtained at the same time - one is tactile data and the other is visual data.

[0068] Step 5: Data processing and analysis

[0069] The collected data is transmitted to the ground control station or cloud server for processing. By analyzing the mechanical data and texture data, it can be identified whether the tall tower has the risk of vibration or shaking. Combined with the data from visual inspection, the inspection results can be further verified and refined to ensure the accuracy and reliability of the inspection.

[0070] Through the above steps, the drone 2 and the tactile sensor device 1 are combined to solve the technical problem that workers need to climb the tower body in person for inspection, which is not only costly and inefficient, but also has great safety risks and poses a potential threat to the personal safety of operators.

[0071] The embodiments of the present invention are described in detail above with reference to the accompanying drawings, but the present invention is not limited to the above embodiments, and various changes can be made within the knowledge scope of those skilled in the art without departing from the spirit of the present invention.

Claims

1. A high-rise tower hidden danger detection drone touch-vision detection device, comprising a drone (2), an external visual sensor (3) and a touch-vision device (1); characterized in that: The drone (2) includes two fixed landing gears, two clockwise propellers, two counterclockwise propellers, a laser radar, a GPS module, a camera module capable of transmitting images in real time, and a touch-vision module. The entire drone (2) flies up, down, left, right, forward, and backward with the touch-vision device (1) and the external vision sensor (3); The touch vision device (1) is a touch vision sensor, which consists of three parts: a sensing part (101), a thin film skin part (102) and a skin fixing ring (103). Gas is filled in from the inflation port at the rear to swell the thin film skin part (102). At the same time, the sensing part (101) and the skin fixing ring (103) work together to fix the swollen shape of the thin film skin part (102) at a suitable position. The concentric ring in the middle of the sensing part (101) is a built-in camera (1011), and the outer ring is an LED light (1012). The thin film skin part (102) uses an elastic film with a printed mark as the touch vision sensor surface, and an air pump is used to make it in a convex fixed shape. When the touch vision sensor is used, the touch vision sensor surface contacts the detection surface, and the point cloud data of the contact surface is generated by an algorithm, so as to fit the texture of the detection surface and detect whether there are defects. The external visual sensor (3) directly acquires the image data of the detection surface to obtain the obvious defect data of the detection part.

2. The high tower hidden danger detection drone tactile visual detection device according to claim 1 is characterized by: The touch-vision device (1) uses an RGB matching method.

3. The high-rise tower hidden danger detection drone tactile visual detection device according to claim 1 is characterized by: The LED light (1012) uses an RGB light strip. When the touch-vision device is in use, the RGB light strip illuminates the surface of the touch-vision sensor, and the built-in camera (1011) captures images and collects data.

4. The high-rise tower hidden danger detection drone tactile visual detection device according to claim 1 is characterized by: The elastic membrane is an ultra-thin elastic silicone film or a gel pad.

5. The high-rise tower hidden danger detection drone tactile visual detection device according to claim 1 is characterized by: When the surface of the touch vision sensor contacts an external object, the elastic film on the surface of the touch vision sensor will deform, reflecting the 3D shape and texture of the contact surface. At the same time, the printed mark will be displaced. Based on this displacement, the force and texture data applied to the surface of the touch vision sensor can be obtained.

6. The high-rise tower hidden danger detection drone tactile visual detection device according to claim 1 is characterized by: Both the external vision sensor (3) and the touch vision sensor use point cloud technology.

7. The high-rise tower hidden danger detection drone tactile visual detection device according to claim 1 is characterized by: The external visual sensor (3) performs feature extraction through the YOLO series algorithm to obtain obvious defect data of the detected part.

8. A drone visual inspection method for high tower hidden danger detection is characterized by The method comprises the high tower hidden danger detection drone touch-vision detection device according to claim 1, wherein the steps are as follows: Step 1: Start the device First, start the tactile sensing device (1), the external visual sensor (3) and the drone (2) to ensure that all devices are operating normally and ready to enter the detection state; Step 2: Operate the drone (2) The drone (2) is operated to fly next to the tall tower, and the flexibility and maneuverability of the drone (2) are used to accurately control its position so that it can approach various inspection locations of the tall tower; Step 3: Touch the tower surface with the touch sensor Maneuvering the drone (2) to a distance where the tactile sensor can touch the surface of the towering tower, ensuring that the surface of the tactile sensor is in close contact with the external structure of the tower, thereby obtaining accurate tactile data; Step 4: Data Collection When the touch-visual sensor is in contact with the tower surface, the touch-sensing device (1) detects the mechanical data and texture data of the contact surface in real time. The data includes the vibration and shaking of the tower surface and the subtle changes of the surface texture. In addition, while the touch-sensing device is in contact with the tower surface, the external visual sensor (3) on the drone (2) can be used to perform visual detection to obtain image data of the tower surface. In this way, two sets of data on the tower structure status can be obtained at the same time - one set is the touch data and the other set is the visual data. Step 5: Data processing and analysis The collected data is transmitted to the ground control station or cloud server for processing. By analyzing the mechanical data and texture data, it can be identified whether the tall tower has the risk of vibration or shaking. Combined with the data from visual inspection, the inspection results can be further verified and refined to ensure the accuracy and reliability of the inspection.

Images