Impact drilling hole okinawa deviation visual monitoring device

Abstract

The application discloses an impact drilling Okinawa deviation visual monitoring device, and particularly relates to the technical field of pile foundation construction equipment.The device comprises an impact drilling host, an industrial camera, a cross laser calibrator, an opening and an operation box.One end of the industrial camera is provided with a cleaning mechanism for cleaning the camera lens.The cleaning mechanism is provided with a water jet pump and a water jet head, which actively spray water.The third driving motor, the threaded rod and the sliding block drive the brush to reciprocate and scrape, so that the camera lens is automatically cleaned by water spraying and physical scraping, which solves the industry common problem that the lens surface is quickly covered with mud and the camera loses the monitoring function in a short time due to the serious splashing of mud during the impact drilling construction.On this basis, the annular high-pressure target flushing ring flushes the red target on the top surface of the pile casing in a timed or image triggered manner, so that the deposited mud is timely flushed away, and the visual recognition failure caused by the target being covered with mud is effectively prevented.

Description

Technical Field

[0001] This application relates to the field of pile foundation construction equipment technology, and more specifically, to a visual monitoring device for the deviation of the impact drilling rope. Background Technology

[0002] In bored pile construction, the impact hammer is suspended from the pile frame by a steel wire rope and lifted by a winch before falling freely to impact and break the rock strata. In hard basalt, inclined rock strata, or areas with developed karst caves, uneven force on the impact hammer at the bottom of the hole can easily cause the steel wire rope to deviate. If the deviation continues to increase and is not detected in time, it will directly cause the pile hole to tilt, and in severe cases, it will lead to hammer jamming or hole collapse, resulting in the scrapping of the pile foundation and causing huge economic losses.

[0003] Traditional borehole deviation monitoring methods mainly rely on manual placement of plumb bobs for measurement, which has the following significant drawbacks: it cannot monitor continuously in real time, and deviations occurring during measurement intervals cannot be detected in time; it lacks accuracy and is greatly affected by on-site wind force and personnel operation; it has a delayed response, and by the time deviation occurs, manual detection, and measures are taken, the best opportunity for correction has often been missed; plumb bob measurement requires working close to the borehole opening, posing a safety hazard of personnel falling into the borehole.

[0004] In recent years, some existing technologies have attempted to incorporate industrial cameras into pile foundation construction monitoring, such as installing fixed cameras on the pile frame to collect images of the borehole opening. However, these solutions have revealed the following structural defects in practical engineering applications: the fixed installation of the cameras means they cannot compensate for the field of view angle due to the elastic deformation of the support pole or strong wind disturbances; when the drilling rig vibrates violently, the center of the pile hole deviates from the center of the image, leading to recognition failure; at the same time, the severe mud splashing during impact drilling quickly covers the lens surface with mud, and the lack of an effective active cleaning structure in existing technologies causes the camera to lose its monitoring function in a short time; furthermore, the connection method between the existing monitoring devices and the pile rig support pole is complex, inconvenient to install and disassemble, and difficult to quickly transfer between different pile locations.

[0005] Therefore, a visual monitoring device for the deviation of the impact drilling rope is proposed to address the above problems. Summary of the Invention

[0006] The purpose of this application is to provide a visual monitoring device for the deviation of the impact drilling rope.

[0007] The visual monitoring device for impact drilling rope deviation provided in this application adopts the following technical solution: A visual monitoring device for percussion drilling rope deviation includes a percussion drill main unit, an industrial camera, a crosshair laser calibrator, an aperture, and an operating box. A pile frame is fixedly connected to the surface of the percussion drill main unit. One side of the pile frame has a mounting mechanism for installing the industrial camera. The surface of the mounting mechanism has a rotating mechanism for rotating the industrial camera, and the surface of the rotating mechanism has an adjustment mechanism for adjusting the angle of the industrial camera. One end of the industrial camera has a cleaning mechanism for cleaning the camera lens. A winch is installed on the percussion drill main unit and the pile frame. A steel wire rope is attached to the winch, which is connected to the percussion hammer. The steel wire rope extends downwards into the aperture. A protective sleeve is installed at the aperture opening. Tracked wheels are installed at the bottom of the percussion drill main unit. A supplementary light and an audible and visual alarm are installed on the surface of the pile frame. The operating box contains a PLC control system and an embedded edge computing terminal. The industrial camera, rotating mechanism, adjustment mechanism, cleaning mechanism, crosshair laser calibrator, supplementary light, and audible and visual alarm are all electrically and / or signal-connected to the operating box. Preferably, the industrial camera captures images of the wire rope and the target at the opening, and the control box runs an offset recognition algorithm to calculate the real-time physical offset of the wire rope relative to the center of the opening. When the offset exceeds a preset threshold, the sound and light alarm is activated.

[0008] Preferably, the top surface of the casing is coated with standard red target markings, and the surface of the wire rope is coated with standard green target markings at intervals. The offset recognition algorithm in the control box uses HSV color space threshold segmentation or YOLOv8 deep learning model to identify the red and green targets and calculate the physical offset of the wire rope from the center of the opening. A ring-shaped high-pressure target flushing ring is provided on one side of the casing. The ring-shaped high-pressure target flushing ring has several high-pressure water spray holes evenly distributed on it. The ring-shaped high-pressure target flushing ring is connected to a water pump through a pipeline. The control box controls the ring-shaped high-pressure target flushing ring to perform high-pressure flushing on the red target markings on the top surface of the casing.

[0009] Preferably, the fill light is a ring-shaped high-brightness LED fill light, which is arranged around the outer periphery of the industrial camera lens. The fill light has a built-in photosensitive sensor that automatically turns on when the ambient light is lower than a preset brightness threshold.

[0010] Preferably, two cross laser calibrators are provided, which are symmetrically installed on the horizontal sides of the front end face of the industrial camera housing. The distance between the two cross laser calibrators is fixed and known. The laser emission direction is towards the top surface of the protective cylinder at an angle downward, projecting two fixed-distance linear laser crosshairs. The operation box corrects the image pixel scale in real time according to the pixel distance change of the two laser crosshairs in the image.

[0011] Preferably, the installation mechanism includes a connecting frame, a T-slot, a T-block, a threaded hole, a bolt rod, and a connecting cover. The connecting frame is fixed to the upright of the pile frame. The surface of the connecting frame has a T-slot vertically formed. The T-block is slidably disposed in the T-slot. The surface of the connecting frame has a threaded hole. The inner wall of the threaded hole is threaded with a bolt rod. One end of the bolt rod is fixedly connected to the connecting cover. One end of the bolt rod cooperates with the T-block.

[0012] Preferably, the rotating mechanism includes a fixed sleeve, a first drive motor, a worm, a worm wheel, and a connecting rod. The fixed sleeve is fixedly connected to the surface of the T-block. The first drive motor is installed inside the fixed sleeve. The output shaft of the first drive motor is connected to the worm. The worm meshes with the worm wheel for transmission. The connecting rod is fixedly connected to the center of the worm wheel.

[0013] Preferably, the adjustment mechanism includes a first fixed frame, a second drive motor, a fixed rod, a first bevel gear, a second bevel gear, a rotating rod, and a second fixed frame. One end of the connecting rod is fixedly connected to the first fixed frame. The second drive motor is fixedly connected inside the first fixed frame. One end of the second drive motor is fixedly connected to the fixed rod. One end of the fixed rod is fixedly connected to the first bevel gear. The surface of the first bevel gear is meshed with the second bevel gear. The rotating rod is fixedly connected inside the second bevel gear. The rotating rod is rotatably connected to the first fixed frame. The surface of the rotating rod is fixedly connected to the second fixed frame. An industrial camera is fixedly connected to the surface of the second fixed frame.

[0014] Preferably, the cleaning mechanism includes a water pump, a water tank, a water nozzle, a chute, a third drive motor, a threaded rod, a slider, and a brush. A water pump is located on one side of the industrial camera, and a water tank is located on the surface of the water pump. A water nozzle is located at one end of the water pump, and the water spray direction of the nozzle is towards the mirror surface of the industrial camera. A chute is formed inside one end of the industrial camera, and a third drive motor is located at one end of the inner wall of the chute. A threaded rod is fixedly connected to one end of the third drive motor, and the industrial camera is rotatably connected to one end of the threaded rod. The slider is threadedly engaged with the threaded rod and slidably disposed within the chute. The brush is fixed to the slider and abuts against the mirror surface of the industrial camera.

[0015] Preferably, the operation box also includes an image clarity self-diagnosis module, which evaluates the cleanliness of the industrial camera lens in real time. When the image clarity is lower than a preset threshold, the cleaning mechanism is automatically triggered to perform water spraying and scraping actions. The operation box also includes a data fusion arbitration module, which receives external sensor signals from the level gauge and / or mud density gauge inside the opening. When the wire rope deviation exceeds a preset threshold and the liquid level drop rate is ≥20% and the mud density drop rate is ≥10%, it is determined to be a hole collapse deviation and an emergency stop is triggered. When the liquid level and density are normal, it is determined to be a vibration deviation and the emergency stop signal is blocked.

[0016] The technical effects and advantages of this application are as follows: Compared with existing technologies, this visual monitoring device for impact drilling rope deviation achieves rapid installation and removal of the industrial camera on the pile frame pole by quickly locking the T-slot and T-block bolt rod of the installation mechanism. This significantly shortens the installation and removal time for transfer between pile positions. The device is stable and reliable with excellent vibration resistance, effectively solving the defects of existing monitoring devices in the background technology, such as complex connection methods, inconvenient installation and removal, and difficulty in rapid transfer between pile positions. Simultaneously, through the worm gear transmission of the rotation mechanism and the bevel gear reversing transmission of the adjustment mechanism, dual-degree-of-freedom electric adjustment of the industrial camera's horizontal rotation and pitch swing is achieved. When the pile frame pole is subjected to impact elastic deformation or strong wind disturbance, causing the pile hole center to move out of the field of view, the PLC control system automatically drives dual-axis linkage tracking to ensure that the pile hole center is always located in the center of the image field of view. This effectively solves the technical problems of the background technology, such as the inability to compensate for the field of view angle when the camera is fixedly installed, and the failure of recognition due to the pile hole center deviating from the center of the image when the drilling rig vibrates violently.

[0017] Compared with existing technologies, this visual monitoring device for impact drilling rope deviation achieves dual cleaning of the camera lens through automatic water spraying and physical scraping. This is achieved by actively spraying water from the cleaning mechanism's water pump and nozzles, combined with a third drive motor, threaded rod, and slider-driven brush. This completely solves the common industry problem in the prior art where severe mud splashing during impact drilling causes the lens surface to be quickly covered by mud, resulting in the camera losing its monitoring function for a short period. Furthermore, a ring-shaped high-pressure target flushing ring performs timed or image-triggered high-pressure flushing on the red target on the top surface of the casing, promptly washing away the deposited mud and effectively preventing the target from being covered by mud, thus avoiding visual recognition failure. Combined with a built-in photosensitive sensor in the supplementary light, it achieves automatic ambient light perception and PWM dimming control, automatically supplementing light in low light conditions and automatically shutting off in high light conditions, and also has overheat protection. From lens cleaning and target flushing to nighttime supplementary lighting, it achieves end-to-end anti-fouling and all-weather protection, ensuring long-term continuous and uninterrupted monitoring.

[0018] Compared with existing technologies, this visual monitoring device for displacement of impact drilling ropes projects laser crosshairs at fixed intervals using a crosshair laser calibrator. The control box corrects the image pixel scale in real time based on the pixel distance between the two laser crosshairs in the image, effectively compensating for perspective distortion caused by the elastic deformation of the pole, temperature changes, and camera posture adjustments, ensuring that the displacement measurement accuracy remains stable within 5mm. At the recognition algorithm level, a dual-mode redundant recognition mechanism using HSV color space threshold segmentation and the YOLOv8 deep learning model is employed. Under normal operating conditions, HSV segmentation is prioritized for fast and low-power recognition. When recognition fails, it automatically switches to the YOLOv8 model to accelerate inference using the NPU. The two algorithms serve as backups for each other, ensuring uninterrupted displacement recognition under various harsh conditions. Meanwhile, the image clarity self-diagnosis module evaluates the cleanliness of the mirror surface in real time, realizing automatic perception of dirt on the mirror surface and triggering autonomous cleaning; the data fusion arbitration module cross-validates visual deviation data with wireless sensor data from the borehole level gauge and mud density gauge, effectively eliminating the defects of pure visual monitoring that are prone to false alarms under drilling rig vibration and high wind conditions, improving the accuracy of borehole collapse warning to over 99%, greatly reducing unnecessary downtime losses caused by false alarms, and realizing all-weather, high-precision, and high-reliability intelligent operation of percussion borehole rope deviation monitoring. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the overall structure of this application; Figure 2 This is a schematic diagram of the structure of the casing and the annular high-pressure target flushing ring of this application. Figure 3 This is a schematic diagram of the structure of the impact drill main unit and the crawler wheel in this application. Figure 4 This is a schematic diagram of the structure of the industrial camera and the cross laser calibrator used in this application. Figure 5 This is a schematic diagram of the installation mechanism of this application; Figure 6 This is a schematic diagram of the rotating mechanism of this application; Figure 7 This is a schematic diagram of the regulating mechanism of this application; Figure 8 This is a schematic diagram of the structure of the cleanup organization in this application; Figure 9 This is a diagram of the YOLOv8 network structure for this application.

[0020] The attached figures are labeled as follows: 1. Impact drill main unit; 2. Pile frame; 3. Winch; 4. Wire rope; 5. Industrial camera; 6. Installation mechanism; 601. Connecting frame; 602. T-slot; 603. T-block; 604. Threaded hole; 605. Bolt rod; 606. Connecting cover; 7. Rotating mechanism; 701. Fixing sleeve; 702. First drive motor; 703. Worm gear; 704. Worm wheel; 705. Connecting rod; 8. Adjusting mechanism; 801. First fixing frame; 802. Second drive motor; 803. Fixing rod; 8 04. First bevel gear; 805. Second bevel gear; 806. Rotating rod; 807. Second fixed frame; 9. Cleaning mechanism; 901. Water pump; 902. Water tank; 903. Spray head; 904. Slide groove; 905. Third drive motor; 906. Threaded rod; 907. Slider; 908. Brush; 10. Impact hammer; 11. Track wheel; 12. Cross laser calibrator; 13. Supplemental light; 14. Opening; 15. Casing; 16. Annular high-pressure target flushing ring; 17. Audible and visual alarm; 18. Control box. Detailed Implementation

[0021] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] Example 1 like Figures 1 to 9The device shown is a visual monitoring device for the deviation of the impact drilling rope. It includes an impact drill main unit 1, an industrial camera 5, a cross laser calibrator 12, an opening 14, and an operation box 18. A pile frame 2 is fixedly connected to the surface of the impact drill main unit 1. A mounting mechanism 6 is provided on one side of the pile frame 2 to facilitate the installation of the industrial camera 5. A rotating mechanism 7 is provided on the surface of the mounting mechanism 6 to rotate the industrial camera 5. An adjustment mechanism 8 is provided on the surface of the rotating mechanism 7 to adjust the angle of the industrial camera 5. A cleaning mechanism 9 is provided at one end of the industrial camera 5 to clean the camera lens. A winch 3 is provided on the impact drill main unit 1 and the pile frame 2. A steel wire rope 4 is provided on the winch 3. The winch 3 is connected to the impact hammer 10 through the steel wire rope 4. The steel wire rope 4 hangs down and extends into the opening 14. A protective sleeve 15 is provided at the opening of the opening 14. The protective sleeve 15 is a steel cylinder buried below the ground and its top is above the ground. A certain height is used to protect the borehole opening from collapse. The bottom of the impact drill 1 is equipped with tracked wheels 11 for the machine to move and relocate on the construction site. The surface of the pile frame 2 is equipped with supplementary lights 13 and audible and visual alarms 17. The control box 18 is equipped with a PLC control system and an embedded edge computing terminal. The industrial camera 5, rotating mechanism 7, adjusting mechanism 8, cleaning mechanism 9, cross laser calibrator 12, supplementary lights 13, and audible and visual alarms 17 are all electrically and / or signal connected to the control box 18. The industrial camera 5 and its accessories are quickly installed on the upright of the pile frame 2 through the mounting mechanism 6. The rotating mechanism 7 and adjusting mechanism 8 realize the dual-degree-of-freedom electric adjustment of the camera's horizontal rotation and pitch swing. The cleaning mechanism 9 realizes the automatic water spray cleaning and scraping of the mirror surface. Combined with the PLC and edge computing terminal in the control box 18, the fully automatic high-precision real-time monitoring of the wire rope 4's deviation is realized.

[0023] In a preferred embodiment, the industrial camera 5 captures images of the wire rope 4 and the target at the opening 14. The industrial camera 5 is an industrial-grade color camera with a resolution of no less than 1920×1080 pixels and a frame rate of no less than 30fps. The lens is vertically downward, aimed at the top surface area of ​​the protective cylinder 15 of the opening 14. The camera transmits the video stream in real time to the embedded edge computing terminal in the operation box 18. The operation box 18 runs an offset recognition algorithm to calculate the real-time physical offset of the wire rope 4 relative to the center of the opening 14. When the offset exceeds a preset threshold, the audible and visual alarm 17 is activated. The execution steps of the offset recognition algorithm are as follows: The top surface of the protective cylinder 15 is coated with a standard red ring mark, and the surface of the wire rope 4 is coated with standard green marks at intervals. The algorithm uses the HSV color space threshold segmentation method to extract the center of the fitted ellipse in the red area as the center coordinate of the opening 14, and extracts the fitted straight line in the green area as the projected straight line of the wire rope 4. The vertical pixel distance from the center of the opening 14 to the fitted straight line of the wire rope is calculated and converted into the actual physical offset by combining it with a pre-calibrated scale. When the offset exceeds the preset threshold of 150mm and lasts for more than 3 seconds, the PLC control system outputs an alarm signal, the audible and visual alarm 17 sounds an alarm, and the touch screen interface of the control box 18 displays the alarm information. If the offset exceeds 200mm, the system directly outputs a dry contact emergency stop signal to the brake control system of the winch 3, automatically stopping the lifting action of the impact hammer. All offset data is stored in the local SD card at 1-second intervals and uploaded to the cloud server in real time via a 4G or 5G module.

[0024] As a preferred embodiment, the top surface of the casing 15 is coated with a standard red target mark, and the surface of the wire rope 4 is coated with standard green target marks at intervals. The red target mark is coated as follows: after the casing 15 is installed in place, a standard red ring mark with a width of 15cm is coated around the top surface of the casing 15 with red anti-rust paint. After the paint dries, it forms a water-resistant and wear-resistant durable target. The green target marking method for the steel wire rope 4 is as follows: In the section of the steel wire rope 4 near the connection end of the impact hammer 10, within a range of approximately 10m upwards from the lifting ring of the impact hammer 10, a 5cm wide ring of green fluorescent paint is applied every 25cm, forming an alternating light and dark coding sequence for the industrial camera 5 to identify the position and orientation of the steel wire rope 4. The offset recognition algorithm in the operation box 18 uses HSV color space threshold segmentation or a YOLOv8 deep learning model to identify the red and green targets, calculating the physical offset of the steel wire rope 4 from the center of the opening 14. The specific process of using HSV color space threshold segmentation is as follows: The image captured by the industrial camera 5 is first converted from the RGB color space to the HSV color space by the embedded edge computing terminal in the operation box 18. After the conversion, the system performs pixel-level filtering on the image according to the preset HSV threshold range. The HSV range of the red target is set as follows: hue H is 0 to 10 degrees or 170 to 180 degrees, saturation S is 200 to 255, and lightness V is 200 to 255. The HSV range for the green target is set to hue (H) between 50 and 75 degrees, saturation (S) between 200 and 255, and brightness (V) between 200 and 255. Red and green binarized mask images are generated through threshold filtering. For the red binarized image, the system extracts all white pixel coordinates and fits an ellipse using the least squares method, taking the center coordinates of the ellipse as the center coordinates (X0, Y0) of the opening 14. For the green binarized image, the system extracts all white pixel coordinates and fits a straight line using the Hough transform, obtaining the projection equation of the steel wire rope 4 in the image: Ax + By + C = 0. Simultaneously, a YOLOv8 deep learning model is used as an auxiliary recognition scheme, automatically switching when the HSV threshold segmentation confidence decreases due to severe mud contamination or drastic changes in lighting. The YOLOv8 model is pre-trained on image datasets covering different lighting conditions, different levels of mud contamination, and different offset states. After training, it is deployed in the embedded edge computing terminal within the operation box 18, utilizing the NPU for inference acceleration, with single-frame inference time controlled within 50ms. During inference, the model outputs the bounding boxes and segmentation masks of the red target area and the green marked area. The system extracts contour points from the segmentation mask and performs ellipse fitting and line fitting, executing the same offset calculation process as HSV segmentation.HSV and YOLOv8 form a dual-mode redundant identification mechanism. Under normal operating conditions, HSV segmentation is used first, and the YOLOv8 model is automatically switched when identification fails to ensure uninterrupted monitoring. A ring-shaped high-pressure target flushing ring 16 is provided on one side of the casing 15. Several high-pressure water jets are evenly distributed on the ring-shaped high-pressure target flushing ring 16. The ring-shaped high-pressure target flushing ring 16 is connected to a water pump via a pipeline. The control box 18 controls the ring-shaped high-pressure target flushing ring 16 to perform high-pressure flushing on the red target markings on the top surface of the casing 15. The ring-shaped high-pressure target flushing ring 16 is connected to the water pump outlet via a high-pressure hose, with a normally closed solenoid valve connected in series in between. The control box 18 controls the opening and closing of the solenoid valve. The control box 18 automatically starts the flushing program every 15 to 30 minutes, or triggers flushing when the number of red pixels in the image is lower than a preset threshold. During flushing, high-pressure water is sprayed radially along the top surface of the casing 15, washing away the deposited mud and exposing the clear red target.

[0025] In a preferred embodiment, the supplementary light 13 is a ring-shaped high-brightness LED supplementary light, surrounding the lens of the industrial camera 5. The supplementary light 13 has a built-in photosensor that automatically turns on when the ambient light is below a preset brightness threshold, turns off when the ambient light is above 100 lux, automatically turns on at 100% power when the ambient light is below 50 lux, and outputs 30% to 70% power using PWM dimming when the ambient light is between 50 and 100 lux. The supplementary light 13 has an overheat protection function; it automatically reduces power when the temperature exceeds 65°C and automatically turns off and triggers an alarm when the temperature exceeds 75°C.

[0026] In a preferred embodiment, two cross laser calibrators 12 are provided, symmetrically installed on both sides of the front end face of the industrial camera 5 housing. The distance between the two cross laser calibrators 12 is fixed and known. The laser emission direction is directed towards the top surface of the protective cylinder 15 at a downward angle, projecting two fixed-distance linear laser crosshairs. The operation box 18 corrects the image pixel scale in real time based on the pixel distance changes of the two laser crosshairs in the image. The two cross laser calibrators 12 project two fixed-distance cross laser marks onto the top surface of the protective cylinder 15. The edge computing terminal inside the operation box 18 detects the pixel distance of the two laser crosshairs in the image in real time. Knowing the physical distance between the two lasers, the real-time scale is calculated to be equal to the physical distance divided by the pixel distance. The system replaces the initially calibrated fixed scale with this real-time scale, thereby achieving dynamic compensation for perspective distortion and changes in installation posture, ensuring that the measurement accuracy is stable within 5mm.

[0027] In a preferred embodiment, the installation mechanism 6 includes a connecting frame 601, a T-slot 602, a T-block 603, a threaded hole 604, a bolt rod 605, and a connecting cover 606. The connecting frame 601 is fixed to the upright of the pile frame 2. The surface of the connecting frame 601 has a vertically formed T-slot 602. The T-block 603 is slidably disposed in the T-slot 602. The surface of the connecting frame 601 has a threaded hole 604. The inner wall of the threaded hole 604 is threadedly connected to the bolt rod 605. One end of the bolt rod 605 is fixedly connected to the connecting cover 606. One end of the bolt rod 605 cooperates with the T-block 603. The T-block 603 is inserted into the T-slot 602. The bolt rod 605 is turned by the connecting cover 606. The bolt rod 605 passes through the threaded hole 604 and is threadedly connected to the T-block 603, so that the T-block 603 is fixed in the connecting frame 601.

[0028] In a preferred embodiment, the rotating mechanism 7 includes a fixed sleeve 701, a first drive motor 702, a worm gear 703, a worm wheel 704, and a connecting rod 705. The fixed sleeve 701 is fixedly connected to the surface of the T-block 603. The first drive motor 702 is installed inside the fixed sleeve 701. The output shaft of the first drive motor 702 is connected to the worm gear 703, and the worm gear 703 meshes with the worm wheel 704 for transmission. The connecting rod 705 is fixedly connected to the center of the worm wheel 704. The first drive motor 702 drives the worm gear 703 to rotate, which in turn drives the worm wheel 704 to rotate at a reduced speed. The worm wheel 704 then drives the connecting rod 705, the adjusting mechanism 8 fixed thereon, and the industrial camera 5 to rotate horizontally around the vertical axis. The worm gear transmission pair has a self-locking characteristic, and the angle remains stable after the motor is powered off.

[0029] In a preferred embodiment, the adjusting mechanism 8 includes a first fixed frame 801, a second drive motor 802, a fixed rod 803, a first bevel gear 804, a second bevel gear 805, a rotating rod 806, and a second fixed frame 807. One end of the connecting rod 705 is fixedly connected to the first fixed frame 801. The second drive motor 802 is fixedly connected inside the first fixed frame 801. One end of the second drive motor 802 is fixedly connected to the fixed rod 803. One end of the fixed rod 803 is fixedly connected to the first bevel gear 804. The surface of the first bevel gear 804 is meshed with the second bevel gear 805. An internal fixed connection to the industrial camera 5 is a rotating rod 806, which is rotatably connected to the first fixed frame 801. A second fixed frame 807 is fixedly connected to the surface of the rotating rod 806, and an industrial camera 5 is fixedly connected to the surface of the second fixed frame 807. When the second drive motor 802 is started, it drives the first bevel gear 804 to rotate through the fixed rod 803. The first bevel gear 804 drives the second bevel gear 805 to rotate 90 degrees in a different direction, so that the second bevel gear 805 drives the second fixed frame 807 and the industrial camera 5 to tilt and swing around the horizontal axis through the rotating rod 806. The tilt angle range is from -15 degrees to +15 degrees.

[0030] In a preferred embodiment, the cleaning mechanism 9 includes a water pump 901, a water tank 902, a water nozzle 903, a slide 904, a third drive motor 905, a threaded rod 906, a slider 907, and a brush 908. The water pump 901 is located on one side of the industrial camera 5, and the water tank 902 is located on the surface of the water pump 901. A water nozzle 903 is located at one end of the water pump 901, with the water spray direction of the nozzle 903 facing the mirror surface of the industrial camera 5. The nozzle 903 is a fan-shaped atomizing nozzle, fixed to the upper left corner of the industrial camera 5 housing by an L-shaped bracket, with the nozzle direction aligned with the center of the mirror surface and the water spray direction facing the mirror surface of the industrial camera 5. The head 5 has a groove 904 inside, and a third drive motor 905 is provided on one end of the inner wall of the groove 904. A threaded rod 906 is fixedly connected to one end of the third drive motor 905, and an industrial camera 5 is rotatably connected to one end of the threaded rod 906. The slider 907 is threadedly engaged with the threaded rod 906 and is slidably disposed in the groove 904. The brush 908 is fixed on the slider 907 and abuts against the mirror surface of the industrial camera 5. The third drive motor 905 drives the threaded rod 906 to rotate in both forward and reverse directions, causing the slider 907 to reciprocate along the groove 904. The brush 908 then reciprocates to scrape the mirror surface, and water is sprayed by the water spray head 903 to clean the mirror surface.

[0031] In a preferred embodiment, the operation box 18 also includes an image clarity self-diagnosis module. This module evaluates the cleanliness of the industrial camera 5's mirror surface in real time. When the image clarity is below a preset threshold, the cleaning mechanism 9 is automatically triggered to perform water spraying and scraping actions. This module is a software program running on an embedded edge computing terminal, which uses the Brenner gradient evaluation function to calculate the clarity score of the mirror's ROI region. When the clarity score is below the preset threshold for three consecutive frames, the mirror surface is determined to be dirty, and the cleaning mechanism 9 is automatically triggered to perform water spraying and scraping actions: first, the water pump 901 is started to spray water for 1 second, and after a delay of 0.5 seconds, the third drive motor 905 is started to drive the brush 908 to scrape back and forth 3 times. After completion, the clarity is re-evaluated. If the hole fails to recover after two cycles, a cleaning failure alarm is triggered. The control box 18 also contains a data fusion arbitration module, which receives external sensor signals from the level gauge and / or mud density gauge inside the opening 14. When the wire rope 4 deviates beyond a preset threshold and the level drop rate is ≥20% and the mud density drop rate is ≥10%, it is determined to be a hole collapse deviation and an emergency stop is triggered. When the level and density are normal, it is determined to be a vibration deviation and the emergency stop signal is blocked. When the wire rope 4 deviates beyond the preset threshold, the module simultaneously reads the hole level and mud density data. If the level drop rate is not less than 20% and the mud density drop rate is not less than 10%, it is determined to be a hole collapse deviation, triggering an emergency stop and an audible and visual alarm. If both the level and density are normal, it is determined to be a vibration deviation, only recording the data and driving the pan-tilt unit angle compensation, blocking the emergency stop signal and not triggering a stop. All arbitration results are recorded and uploaded to the cloud at 1-second intervals, forming a complete data chain for tracing construction quality.

[0032] The working process of this application is as follows: After the impact drill host 1 is moved to the construction pile position by the track wheels 11, the T-block 603 is slid into the T-slot 602 and locked by the bolt rod 605 to complete the installation mechanism 6. Then, the initial calibration is performed - the operation box 18 identifies the laser crosshairs projected by the two cross laser calibrators 12 on the top surface of the casing 15, calculates the initial pixel scale, and at the same time, the supplementary light 13 automatically adjusts the supplementary light intensity according to the photosensitive sensor; during the construction process, the industrial camera 5 collects the video stream of the opening 14 in real time at a frame rate of not less than 30fps, and the edge computing terminal in the operation box 18 preferentially uses HSV color space threshold segmentation. (After converting the RGB image to HSV space, pixel filtering is performed based on the red target H:0~10 / S:200~255 / V:200~255 and the green target H:50~75 / S:200~255 / V:200~255. The center coordinates of the pile hole are obtained by fitting the red ellipse, and the steel wire rope projection line is obtained by fitting the green line using Hough transform.) When the confidence of HSV segmentation decreases due to mud contamination or changes in lighting, it automatically switches to the YOLOv8 deep learning model to accelerate dual-mode redundant recognition using NPU inference. At the same time, the cross laser calibrator 12 detects the pixel distance between the two laser crosses in the image in real time. The system changes and dynamically corrects the scale, calculates the vertical pixel distance from the center of the pile hole to the straight line of the wire rope, and converts it into a physical offset Dr. When Dr exceeds 150mm and lasts for 3 seconds, an audible and visual alarm 17 is triggered. When it exceeds 200mm, a dry contact emergency stop signal is directly output to the winch 3 brake control system to automatically stop the lifting of the impact hammer 10. At the same time, when the offset continues to increase and causes the center of the pile hole to move out of the center of the field of view, the first drive motor 702 of the rotating mechanism 7 drives the industrial camera 5 to rotate horizontally through the worm gear 703 and worm wheel 704. The second drive motor 802 of the adjusting mechanism 8 drives the industrial camera 5 to rotate horizontally through the first bevel gear 804 and the second bevel gear 805. The camera tilts and swings with a 0-degree reversal, achieving dual-degree-of-freedom active focus tracking; the image clarity self-diagnosis module uses the Brenner gradient evaluation function to monitor the cleanliness of the mirror surface in real time. When the clarity score is below the threshold for 3 consecutive frames, the cleaning mechanism 9 is automatically triggered. The water pump 901 sprays water onto the mirror surface through the water nozzle 903. After 1 second, the third drive motor 905 drives the threaded rod 906 to make the slider 907 drive the brush 908 to scrape back and forth 3 times. At the same time, the operation box 18 automatically opens the annular high-pressure target flushing ring 16 every 15 to 30 minutes or when the number of red pixels in the image is insufficient to perform high-pressure flushing on the top surface of the protective cylinder 15 to expose a clear red target.The data fusion arbitration module synchronously receives signals from the borehole level gauge and mud density gauge. When Dr exceeds 150mm and the level drop rate is ≥20% and the mud density drop rate is ≥10%, it is determined to be a borehole collapse deviation, triggering an emergency stop. When the level and density are normal, it is determined to be a vibration deviation, only recording the data and driving the gimbal angle compensation, shielding the emergency stop signal. All deviation data, level and density data, and arbitration results are stored on the local SD card at 1-second intervals and uploaded to the cloud in real time via a 4G or 5G module. Managers can remotely monitor via an APP or mini-program, with a display delay of less than 50ms. After a single borehole is completed, loosening the bolt rod 605 allows the T-block 603 to slide out of the T-slot 602, thus disassembling the entire visual monitoring module. It can then be transferred to the next pile location via the track wheels 11 for reinstallation and continued operation. The supplementary light 13 automatically turns on when the ambient illuminance is below 50 lux, achieving uninterrupted monitoring around the clock. The above describes the working principle of this impact drilling rope deviation visual monitoring device.

Claims

1. A visual monitoring device for the deviation of the drilling rope in an impact drill, comprising an impact drill main unit (1), an industrial camera (5), a cross laser calibrator (12), an opening (14), and an operation box (18), wherein a pile frame (2) is fixedly connected to the surface of the impact drill main unit (1), characterized in that: The pile frame (2) has a mounting mechanism (6) on one side for installing an industrial camera (5). The surface of the mounting mechanism (6) has a rotating mechanism (7) for rotating the industrial camera (5). The surface of the rotating mechanism (7) has an adjusting mechanism (8) for adjusting the angle of the industrial camera (5). One end of the industrial camera (5) has a cleaning mechanism (9) for cleaning the camera lens. The impact drill host (1) and the pile frame (2) are equipped with a winch (3). The winch (3) is equipped with a wire rope (4). The winch (3) is connected to the impact hammer (10) via the wire rope (4). The steel wire rope (4) extends downward into the opening (14), and a protective sleeve (15) is provided at the opening of the opening (14). The bottom of the impact drill host (1) is provided with a track wheel (11). The surface of the pile frame (2) is provided with a supplementary light (13) and an audible and visual alarm (17). The operation box (18) is provided with a PLC control system and an embedded edge computing terminal. The industrial camera (5), rotating mechanism (7), adjusting mechanism (8), cleaning mechanism (9), cross laser calibrator (12), supplementary light (13), and audible and visual alarm (17) are all electrically connected and / or signal connected to the operation box (18).

2. The visual monitoring device for percussion drilling rope deviation according to claim 1, characterized in that: The industrial camera (5) captures images of the wire rope (4) and the target at the opening (14). The operation box (18) runs an offset recognition algorithm to calculate the real-time physical offset of the wire rope (4) relative to the center of the opening (14). When the offset exceeds a preset threshold, the sound and light alarm (17) is activated.

3. The visual monitoring device for percussion drilling rope deviation according to claim 1, characterized in that: The top surface of the casing (15) is coated with a standard red target mark, and the surface of the wire rope (4) is coated with a standard green target mark at intervals. The offset recognition algorithm in the operation box (18) uses HSV color space threshold segmentation or YOLOv8 deep learning model to identify the red and green targets and calculate the physical offset of the wire rope (4) to the center of the opening (14). A ring high-pressure target flushing ring (16) is provided on one side of the casing (15). Several high-pressure water spray holes are evenly distributed on the ring high-pressure target flushing ring (16). The ring high-pressure target flushing ring (16) is connected to a water pump through a pipeline. The operation box (18) controls the ring high-pressure target flushing ring (16) to perform high-pressure flushing on the red target mark on the top surface of the casing (15).

4. The visual monitoring device for percussion drilling rope deviation according to claim 1, characterized in that: The fill light (13) is a ring-shaped high-brightness LED fill light, which is arranged around the outer periphery of the lens of the industrial camera (5). The fill light (13) has a built-in photosensitive sensor that automatically turns on when the ambient light is lower than the preset brightness threshold.

5. The visual monitoring device for percussion drilling rope deviation according to claim 1, characterized in that: Two cross laser calibrators (12) are provided and are symmetrically installed on the horizontal sides of the front end face of the industrial camera (5) housing. The distance between the two cross laser calibrators (12) is fixed and known. The laser emission direction is towards the top surface of the protective cylinder (15) at an angle downward, projecting two fixed-distance linear laser crosshairs. The operation box (18) corrects the image pixel scale in real time according to the pixel distance change of the two laser crosshairs in the image.

6. The visual monitoring device for percussion drilling rope deviation according to claim 1, characterized in that: The installation mechanism (6) includes a connecting frame (601), a T-slot (602), a T-block (603), a threaded hole (604), a bolt rod (605), and a connecting cover (606). The connecting frame (601) is fixed on the upright of the pile frame (2). The surface of the connecting frame (601) is provided with a T-slot (602) along the vertical direction. The T-block (603) is slidably disposed in the T-slot (602). The surface of the connecting frame (601) is provided with a threaded hole (604). The inner wall of the threaded hole (604) is threadedly connected to a bolt rod (605). One end of the bolt rod (605) is fixedly connected to the connecting cover (606). One end of the bolt rod (605) cooperates with the T-block (603).

7. The visual monitoring device for percussion drilling rope deviation according to claim 6, characterized in that: The rotating mechanism (7) includes a fixed sleeve (701), a first drive motor (702), a worm (703), a worm wheel (704), and a connecting rod (705). The fixed sleeve (701) is fixedly connected to the surface of the T-block (603). The first drive motor (702) is installed inside the fixed sleeve (701). The output shaft of the first drive motor (702) is connected to the worm (703). The worm (703) meshes with the worm wheel (704) for transmission. The connecting rod (705) is fixedly connected to the center of the worm wheel (704).

8. The visual monitoring device for percussion drilling rope deviation according to claim 7, characterized in that: The adjustment mechanism (8) includes a first fixed frame (801), a second drive motor (802), a fixed rod (803), a first bevel gear (804), a second bevel gear (805), a rotating rod (806), and a second fixed frame (807). One end of the connecting rod (705) is fixedly connected to the first fixed frame (801). The second drive motor (802) is fixedly connected inside the first fixed frame (801). One end of the second drive motor (802) is fixedly connected to the fixed rod (803). One end of the fixed rod (803) is fixedly connected to the first bevel gear (804). The surface of the first bevel gear (804) is meshed with the second bevel gear (805). The interior of the second bevel gear (805) is fixedly connected to the rotating rod (806). The rotating rod (806) is rotatably connected to the first fixed frame (801). The surface of the rotating rod (806) is fixedly connected to the second fixed frame (807). The surface of the second fixed frame (807) is fixedly connected to the industrial camera (5).

9. The visual monitoring device for percussion drilling rope deviation according to claim 1, characterized in that: The cleaning mechanism (9) includes a water pump (901), a water tank (902), a water nozzle (903), a chute (904), a third drive motor (905), a threaded rod (906), a slider (907), and a brush (908). A water pump (901) is located on one side of the industrial camera (5). A water tank (902) is located on the surface of the water pump (901). A water nozzle (903) is located at one end of the water pump (901), and the water spray direction of the water nozzle (903) is towards the mirror surface of the industrial camera (5). The industrial camera (5) has a sliding groove (904) inside one end. A third drive motor (905) is provided at one end of the inner wall of the sliding groove (904). A threaded rod (906) is fixedly connected to one end of the third drive motor (905). The industrial camera (5) is rotatably connected to one end of the threaded rod (906). The slider (907) is threadedly engaged with the threaded rod (906) and slidably disposed in the sliding groove (904). The brush (908) is fixed on the slider (907) and abuts against the mirror surface of the industrial camera (5).

10. The visual monitoring device for percussion drilling rope deviation according to claim 1, characterized in that: The operation box (18) is also equipped with an image clarity self-diagnosis module. The image clarity self-diagnosis module evaluates the cleanliness of the industrial camera (5) mirror in real time. When the image clarity is lower than the preset threshold, the cleaning mechanism (9) is automatically triggered to perform water spraying and scraping actions. The operation box (18) is also equipped with a data fusion arbitration module. The data fusion arbitration module is used to receive external sensor signals from the liquid level gauge and / or mud density gauge in the opening (14). When the wire rope (4) deviates beyond the preset threshold and the liquid level drop rate is ≥20% and the mud density drop rate is ≥10%, it is determined to be a hole collapse deviation and an emergency stop is triggered. When the liquid level and density are normal, it is determined to be a vibration deviation and the emergency stop signal is blocked.

Images