Intelligent wall-climbing robot vehicle and intelligent repair method for repairing concrete cracks

Abstract

The present invention belongs to the technical field of concrete detection and repair, and specifically is an intelligent wall-climbing robot vehicle and an intelligent repair method for repairing concrete cracks. The intelligent wall-climbing robot vehicle includes: a robot vehicle synchronous positioning and mapping system, which is used to construct a three-dimensional environmental model of the wall to be repaired; a negative pressure adsorption mobile system, which is used to climb and move on the wall to be repaired; a crack detection system, which is used to identify the position and width of the crack; a crack pre-cleaning system, which is used to pre-clean the cracks identified by the crack detection system; a crack plant urease repair system, which is used to perform urease repair on the cracks according to the crack width or to fill and repair the cracks with urease; an intelligent control system, which is used to control the operation of the robot vehicle synchronous positioning and mapping system, the negative pressure adsorption mobile system, the crack detection system, the crack pre-cleaning system and the crack plant urease repair system. The present invention can realize automatic detection and green and efficient repair of cracks.

Description

Technical Field

[0001] The present invention belongs to the technical field of concrete detection and repair, and specifically relates to an intelligent wall-climbing robot vehicle and an intelligent repair method for repairing concrete cracks. Background Art

[0002] As one of the main building materials, concrete is widely used in various engineering structures. However, due to the inherent defects of concrete materials and the influence of the harsh external environment, most concrete structures will inevitably develop cracks during use. These cracks not only affect the appearance of the structure, but also seriously reduce its safety and durability. The existence of cracks provides a channel for the invasion of external moisture, chloride ions, etc., which in turn accelerates the deterioration process such as steel corrosion and concrete carbonization, resulting in a decrease in the bearing capacity of the structure, and even causing major safety accidents such as collapse, resulting in casualties and huge economic and property losses. Therefore, timely and effective detection and repair of concrete cracks to ensure the long-term stability of the structure has become a difficult problem that needs to be solved in the field of civil engineering.

[0003] At present, crack repair mainly relies on manual operation, and maintenance personnel adopt different repair methods according to the damage and degree of damage to the structure. However, traditional manual repair not only has high labor costs, but also faces limitations such as low efficiency and difficulty in covering large-area structures. Especially in high-rise buildings or complex structures, crack detection and repair are more difficult. First, it is necessary to confirm the location of the crack through detection equipment and measure its length and width. According to the test results, the filling grouting method, surface covering repair method and other methods are used for repair. However, for locations that are high or difficult to access, it is usually necessary to use lifting equipment to send workers and repair equipment to the designated area for crack detection and repair, which not only increases construction costs, but also poses safety hazards.

[0004] Although some automated equipment has been used for crack repair, such as the invention patent with publication number CN116575747A, which proposes an intelligent detection and automatic repair device through pre-laid tracks. The device uses a robotic arm to identify and repair cracks, and the discharge pipe automatically moves to the location of the wall crack to achieve intelligent spraying repair. However, this type of equipment usually relies on preset tracks or manual operation, and cannot fully realize flexible automated detection and repair. The labor cost is still high, and it cannot adapt to a variety of construction environments. In addition, the equipment only uses grouting to repair cracks, which is difficult to adapt to the repair needs of various crack types. Moreover, the wall cracks are vertical structures, and the overflow of slurry needs to be considered during grouting. Therefore, it is urgent to provide a new type of technical equipment that can intelligently and automatically identify and repair cracks. Summary of the invention

[0005] The purpose of the present invention is to provide an intelligent wall-climbing robot vehicle and an intelligent repair method for repairing concrete cracks, so as to realize automatic identification and targeted repair of cracks and improve the repair efficiency and effect.

[0006] To achieve the above-mentioned object, the present invention provides an intelligent wall-climbing robot vehicle for repairing concrete cracks, comprising a vehicle body and functional components arranged on the vehicle body, wherein the functional components include:

[0007] The robot vehicle synchronous positioning and mapping system is used to build a three-dimensional environmental model of the wall to be repaired;

[0008] A negative pressure adsorption moving system is used to drive the intelligent wall-climbing robot to climb and move on the wall to be repaired;

[0009] Crack detection system to identify the location and width of cracks;

[0010] A crack pre-cleaning system, used for pre-cleaning cracks identified by the crack detection system;

[0011] Crack plant urease repair system, used to perform urease repair on cracks or to fill and urease repair cracks according to crack width;

[0012] An intelligent control system is used to control the operation of the robot vehicle synchronous positioning and mapping system, the negative pressure adsorption movement system, the crack detection system, the crack pre-cleaning system and the crack plant urease repair system.

[0013] Further, the crack plant urease repair system includes a crack repair component and a crack filling component;

[0014] The crack repair assembly includes a first mechanical injection arm carrying a cementing liquid and a plant urease solution, and is used to inject a plant urease-based repair agent into the crack;

[0015] The crack filling assembly includes a second mechanical injection arm carrying a filling agent, which is used to inject the filling agent into the crack.

[0016] Furthermore, when the width of the crack is less than 0.6 mm, the intelligent control system controls the crack repair component to inject a plant urease-based repair agent into the crack; when the width of the crack is greater than or equal to 0.6 mm, the intelligent control system controls the crack filling component to fill the crack with the filling agent, and then controls the crack repair component to inject a plant urease-based repair agent into the crack.

[0017] Furthermore, the crack detection system includes a trained crack recognition model based on a U-net deep convolutional neural network; the recognition process of the crack recognition model includes: extracting crack features through secondary convolution and pooling operations in the downsampling stage of the U-net deep convolutional neural network, copying the secondary convolution feature map to an upsampling area with the same number of channels for superposition, and completing a skip-layer connection operation; in the upsampling stage, using deconvolution and skip-layer connection operations to restore the small-size multi-channel feature map to the original input image size, and obtaining a crack morphology map;

[0018] Then, the crack point of the single-pixel crack morphology map is used as the seed point, and an eight-direction search is performed on the crack morphology map through the seed point until the crack boundary is reached. The number of pixels in the eight directions is counted and merged in the four directions of 0°, 45°, 90°, and 135°. The minimum number of pixels is taken as the crack width for calibration, and the crack width is obtained by multiplying it with the actual size corresponding to the pixel point.

[0019] Furthermore, the negative pressure adsorption mobile system includes a negative pressure suction cup and wheels arranged at the bottom of the intelligent wall-climbing robot vehicle;

[0020] The robot vehicle synchronous positioning and mapping system includes a camera, a gyroscope and a positioning module; the camera is arranged on the front side of the intelligent wall-climbing robot vehicle;

[0021] The crack pre-cleaning system includes a mechanical arm equipped with an air gun; one end of the mechanical arm is fixed in the cabin of the intelligent wall-climbing robot vehicle, and the other end extends out from the front side of the intelligent wall-climbing robot vehicle;

[0022] One end of the first mechanical injection arm and the second mechanical injection arm are installed side by side in the cabin of the intelligent wall-climbing robot vehicle, and the other end extends out from the front side of the intelligent wall-climbing robot vehicle; the cabin is also provided with a plant urease solution tank and a binder tank connected to the first mechanical injection arm, and a filler tank connected to the second mechanical injection arm.

[0023] The present invention also provides an intelligent repair method for concrete wall cracks, using any of the above-mentioned intelligent wall-climbing robots, comprising the following steps:

[0024] S1. According to the structural shape and geographical location of the wall to be repaired, a suitable position is selected to vertically attach the robot to the wall, and then the wall to be repaired is scanned to construct a three-dimensional environmental model of the wall to be repaired;

[0025] S2. The intelligent control system plans a moving route according to the three-dimensional environment model, and the intelligent wall-climbing robot moves on the wall to be repaired according to the moving route, and simultaneously continuously scans the wall to be repaired to obtain image information of the wall to be repaired;

[0026] S3, the crack detection system identifies the location of the crack according to the image information. When the crack is identified, the intelligent wall-climbing robot stops moving and starts the crack pre-cleaning system to pre-clean the crack;

[0027] S4, the crack detection system re-identifies the cracks after pre-cleaning and re-obtains the position and width of the cracks;

[0028] S5. The intelligent control system starts the crack plant urease repair system to perform urease repair on the crack or to fill and repair the crack according to the width of the crack;

[0029] S6. After the current cracks are repaired, the intelligent wall-climbing robot continues to move according to the moving route and repeats steps S3 to S5 until all cracks in the repair wall are repaired.

[0030] Furthermore, in step S5, when the width of the crack is less than 0.6 mm, the intelligent control system controls the crack repair component to inject a plant urease-based repair agent into the crack; when the width of the crack is greater than or equal to 0.6 mm, the intelligent control system controls the crack filling component to fill the crack with the filling agent, and then controls the crack repair component to inject a plant urease-based repair agent into the crack.

[0031] Further, the plant urease-based repair agent includes a cementing liquid and a plant urease solution;

[0032] The binder solution comprises 0.5-1.0 mol / L calcium chloride and 0.5-1.0 mol / L urea, and the molar ratio of calcium chloride to urea is 1:1; the plant urease solution is obtained by fully mixing soybean powder with a mixed solution of ethanol and water in a volume ratio of 30%:70% and then filtering;

[0033] The volume ratio of the cementing liquid to the plant urease solution is 1:1;

[0034] The filler is fine sand with a diameter less than or equal to 0.3 mm.

[0035] Further, in steps S1 and S2, the scanning is performed by a camera;

[0036] In step S2, while continuously scanning the wall to be repaired, the robot vehicle synchronous positioning and mapping system continuously establishes and improves the three-dimensional environment model;

[0037] In step S3, the pre-cleaning is to use an air gun to blow and clean the inside and surface of the crack;

[0038] In step S4, the position and width of the crack are obtained by the crack detection system.

[0039] In step S5, the intelligent control system determines the inclination state of the crack according to the three-dimensional environmental model, divides the crack into several sections according to the inclination state, and performs urease repair or filling and urease repair on each section in turn.

[0040] Furthermore, the crack detection system identifies the cracks through a crack recognition model based on a U-net deep convolutional neural network;

[0041] The U-net deep convolutional neural network first obtains a crack morphology map, then uses the crack points of the single-pixel crack morphology map as seed points, performs eight-directional search on the crack morphology map through the seed points, and stops at the crack boundary. The number of pixels in the eight directions is counted and merged in the four directions of 0°, 45°, 90°, and 135°. The minimum number of pixels is taken as the crack width for calibration, and the crack width is obtained by multiplying it with the actual size corresponding to the pixel point.

[0042] In general, the above technical solution conceived by the present invention has the following technical advantages compared with the prior art:

[0043] 1. The intelligent wall-climbing robot vehicle for repairing concrete cracks provided by the present invention performs three-dimensional modeling of the wall to be repaired through the robot vehicle synchronous positioning and mapping system, obtains the coordinate system of the current environment, and is convenient for rational planning of the robot vehicle's moving route and accurate positioning of the crack position. Through the recognition model of the crack detection system, the cracks can be accurately identified to achieve accurate and comprehensive repair. Through the crack plant urease repair system, the repair plan can be adaptively adjusted according to the crack width, so that the repair effect is better and more complete.

[0044] 2. The present invention is equipped with a crack intelligent identification system and a repair combination robot arm, and uses deep learning, plant urease repair technology and sand fixation technology to achieve automatic detection and green and efficient repair of cracks, significantly improving the flexibility, efficiency and environmental friendliness of crack repair. The robot is equipped with a crack pre-cleaning device, which can handle complex cracks, and combines two plant urease-based repair technologies (enzyme solution deposition calcium carbonate repair and consolidation sandblasting repair), effectively solving many limitations in the existing technology and providing a new technical means for the automated and intelligent repair of concrete structures.

[0045] 3. The present invention adopts a negative pressure adsorption wall-climbing robot vehicle with a compact and flexible design, which enables it to climb and complete repair work in dangerous locations such as vertical walls. It does not rely on track laying, greatly reduces dependence on manpower and material resources, and significantly improves repair efficiency.

[0046] 4. The present invention is based on the repair scenario of vertical walls. It determines the inclination angle and direction of the cracks according to the three-dimensional environmental model, and divides the cracks into three states: vertical, horizontal and inclined. The cracks are repaired in sections according to the three states, so that the repair liquid or filler can be better injected into the cracks, effectively preventing overflow, reducing waste and pollution to the wall. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A schematic diagram of the structure of the intelligent wall-climbing robot vehicle for repairing concrete cracks provided by the present invention after opening the cabin.

[0048] Figure 2 A schematic diagram of the three-dimensional structure of the intelligent wall-climbing robot vehicle for repairing concrete cracks provided by the present invention with the cabin closed.

[0049] Figure 3 A schematic diagram of the functional framework of the intelligent wall-climbing robot vehicle for repairing concrete cracks provided by the present invention.

[0050] Figure 4 An intelligent repair flow chart of the intelligent wall-climbing robot vehicle for repairing concrete cracks provided by the present invention.

[0051] Throughout the drawings, the same reference numerals are used to denote the same elements or structures, wherein:

[0052] 1-Detection system equipped with a camera; 2-Robotic arm equipped with an air gun; 3-First robotic injection arm; 4-Second robotic injection arm; 5-Filling agent tank; 6-Cabin shield; 7-Urease solution tank; 8-Cementing liquid tank; 9-Car body; 10-Wheel; 11-Negative pressure suction cup. DETAILED DESCRIPTION

[0053] In order to make the purpose, technical scheme and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with the 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. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0054] See also Figure 1-4 The present invention provides an intelligent wall-climbing robot vehicle for repairing concrete cracks, comprising a body 9 and a functional component arranged on the body 9, wherein the functional component comprises:

[0055] The robot vehicle synchronous positioning and mapping system is used to build a three-dimensional environmental model of the wall to be repaired;

[0056] Negative pressure adsorption mobile system, used for climbing and moving on the wall to be repaired;

[0057] Crack detection system to identify the location and width of cracks;

[0058] A crack pre-cleaning system, used for pre-cleaning cracks identified by the crack detection system;

[0059] Crack plant urease repair system, used to perform urease repair on cracks or to fill and urease repair cracks according to crack width;

[0060] An intelligent control system is used to control the operation of the robot vehicle synchronous positioning and mapping system, the negative pressure adsorption movement system, the crack detection system, the crack pre-cleaning system and the crack plant urease repair system.

[0061] Specifically, the intelligent wall-climbing robot vehicle includes a cabin body and functional components, and the functional components include the above-mentioned negative pressure adsorption mobile system, crack pre-cleaning system, crack detection system, crack plant urease repair system, robot vehicle synchronous positioning and mapping system and robot vehicle intelligent control system. Among them, the negative pressure adsorption mobile system includes a negative pressure suction cup 11 and wheels 10 arranged at the bottom of the cabin body.

[0062] The robot vehicle synchronous positioning and mapping system includes a camera, a gyroscope and a positioning module; the camera (such as Figure 1 The detection system 1) equipped with a camera is arranged on the front side of the cabin body of the intelligent wall-climbing robot vehicle for scanning and detection.

[0063] The crack pre-cleaning system includes a robotic arm 2 equipped with an air gun; one end of the robotic arm is fixed in the cabin of the intelligent wall-climbing robot vehicle, and the other end extends out in front of the cabin body of the intelligent wall-climbing robot vehicle; the front side is the front side of the intelligent wall-climbing robot vehicle, that is, the direction of wheel movement.

[0064] The crack plant urease repair system includes a crack repair component and a crack filling component; the crack repair component includes a first mechanical injection arm 3 carrying a binder and a plant urease solution, which is used to inject a plant urease-based repair agent into the crack; the crack filling component includes a second mechanical injection arm 4 carrying a filling agent, which is used to inject the filling agent into the crack.

[0065] One end of the first mechanical injection arm 3 and the second mechanical injection arm 4 are installed in parallel in the cabin of the intelligent wall-climbing robot vehicle, and the other end extends out from the front side of the intelligent wall-climbing robot vehicle; the cabin is also provided with a plant urease solution tank 7 and a cementing liquid tank 8 connected to the first mechanical injection arm 3, which are transported to the first mechanical injection arm 3 through a built-in micro water pump; the second mechanical injection arm 4 is connected to the filler tank 5, which is transported to the second mechanical injection arm 4 through a built-in micro water pump. The filler tank 5, urease solution tank 7 and cementing liquid tank 8 are all fixed inside the cabin body and can be unloaded when they need to be replaced; the cabin baffle 6 is installed on the rear side of the cabin body. When the robot vehicle is in working state, the baffle 6 is retracted, the cabin is opened, and the mechanical arm and injection arm are extended to work. Figure 1 When the robot car is not in working state, the cabin shield 6 is unfolded, the robot arm and the injection arm are retracted, and the cabin is closed, such as Figure 2 .

[0066] With such an arrangement, the negative pressure suction cup 11 and the wheel 10 cooperate to freely climb and move on the vertical wall without relying on track laying, which greatly simplifies the operation process. The wall to be repaired is three-dimensionally modeled through the robot vehicle synchronous positioning and mapping system to obtain the coordinate system of the current environment, which is convenient for reasonable planning of the movement route of the robot vehicle and accurate positioning of the crack position. When the crack needs to be cleaned or repaired, according to the position of the crack in the coordinate system, the robot vehicle can be controlled to move so that the mechanical arm or injection arm is aligned with the corresponding position of the crack for cleaning and repair. Through the recognition model of the crack detection system, the crack can be accurately identified to achieve accurate and comprehensive repair. The crack plant urease repair system can adaptively adjust the repair plan according to the width of the crack, so that the repair effect is better and more complete.

[0067] In particular, when the width of the crack is less than 0.6 mm, the intelligent control system controls the crack repair component to inject a repair agent based on plant urease into the crack; when the width of the crack is greater than or equal to 0.6 mm, the intelligent control system controls the crack filling component to fill the crack with the filling agent, and then controls the crack repair component to inject the repair agent based on plant urease into the crack. In this way, targeted repair can be performed according to the degree of cracking, and the repair is more flexible and scientific.

[0068] The crack detection system includes a trained crack recognition model based on a U-net deep convolutional neural network; the recognition process of the crack recognition model includes: extracting crack features through secondary convolution and pooling operations in the downsampling stage of the U-net deep convolutional neural network, copying the secondary convolution feature map to an upsampling area with the same number of channels for superposition, and completing a skip-layer connection operation; in the upsampling stage, using deconvolution and skip-layer connection operations to restore the small-size multi-channel feature map to the original input image size, and obtaining a crack morphology map;

[0069] Then, the crack point of the single-pixel crack morphology map is used as the seed point, and an eight-direction search is performed on the crack morphology map through the seed point until the crack boundary is reached. The number of pixels in the eight directions is counted and merged in the four directions of 0°, 45°, 90°, and 135°. The minimum number of pixels is taken as the crack width for calibration, and the crack width is obtained by multiplying it with the actual size corresponding to the pixel point.

[0070] The training of the crack recognition model is to classify 40,000 3-channel crack images of 227×277 pixels into training set and validation set in a ratio of 7:3, and standardize and enlarge the data set to highlight the crack contrast. Then, the training and validation are carried out according to the above recognition process, so as to continuously optimize the recognition model.

[0071] See also Figure 4 The present invention also provides an intelligent repair method for concrete wall cracks, using the above-mentioned intelligent wall climbing robot, comprising the following steps:

[0072] S1. According to the structural shape and geographical location of the wall to be repaired, a suitable position is selected to vertically attach the robot to the wall, and then the wall to be repaired is scanned to construct a three-dimensional environmental model of the wall to be repaired;

[0073] S2. The intelligent control system plans a moving route according to the three-dimensional environment model, and the intelligent wall-climbing robot moves on the wall to be repaired according to the moving route, and simultaneously continuously scans the wall to be repaired to obtain image information of the wall to be repaired;

[0074] S3, the crack detection system identifies the location of the crack according to the image information. When the crack is identified, the intelligent wall-climbing robot stops moving and starts the crack pre-cleaning system to pre-clean the crack;

[0075] S4, the crack detection system re-identifies the cracks after pre-cleaning and re-obtains the position and width of the cracks;

[0076] S5. The intelligent control system starts the crack plant urease repair system to perform urease repair on the crack or to fill and repair the crack according to the width of the crack;

[0077] S6. After the current cracks are repaired, the intelligent wall-climbing robot continues to move according to the moving route and repeats steps S3 to S5 until all cracks in the repair wall are repaired.

[0078] In step S5, when the width of the crack is less than 0.6 mm, the intelligent control system controls the crack repair component to inject a plant urease-based repair agent into the crack; when the width of the crack is greater than or equal to 0.6 mm, the intelligent control system controls the crack filling component to fill the crack with the filling agent, and then controls the crack repair component to inject a plant urease-based repair agent into the crack.

[0079] The plant urease-based repair agent includes a binder and a plant urease solution; the binder comprises 0.5-1.0 mol / L calcium chloride and 0.5-1.0 mol / L urea, and the molar ratio of the calcium chloride to the urea is 1:1; the plant urease solution is obtained by fully mixing soybean powder with a mixed solution of ethanol and water in a volume ratio of 30%:70% and then filtering; the mixed volume ratio of the binder and the plant urease solution is 1:1;

[0080] The filler is fine sand with a diameter less than or equal to 0.3 mm.

[0081] In steps S1 and S2, the scanning is performed by a camera; in step S2, while the wall to be repaired is continuously scanned synchronously, the robot vehicle synchronous positioning and mapping system continuously establishes and improves the three-dimensional environment model.

[0082] In step S3, the pre-cleaning is to use an air gun to blow and clean the inside and surface of the crack; in step S4, the position and width of the crack are obtained by the crack detection system.

[0083] The crack detection system identifies the cracks through a crack recognition model based on a U-net deep convolutional neural network; the U-net deep convolutional neural network first obtains a crack morphology map, then uses the crack points of the single-pixel crack morphology map as seed points, performs eight-directional search on the crack morphology map through the seed points, and stops at the crack boundary, counts the number of pixels in the eight directions and merges them in the four directions of 0°, 45°, 90°, and 135°, takes the minimum number of pixels as the crack width for calibration, and multiplies it by the actual size corresponding to the pixel to obtain the crack width.

[0084] Specifically, the intelligent repair method includes the following steps:

[0085] Step 1: First, according to the structural shape and geographical location of the wall to be repaired, select a suitable location to attach the robot vertically to the wall, turn on the power, and start the camera and sonar of the robot to perform the first scan to obtain the terrain features of the surrounding environment and build a three-dimensional environment coordinate system model. The intelligent control system receives the three-dimensional environment coordinate system model and selects a suitable route to start work through the automatic driving planning application in the intelligent control system. When the robot moves and works, the camera and sonar continue to scan, and continuously build and improve the three-dimensional environment coordinate system model.

[0086] Step 2: Train and apply the crack recognition model. 40,000 crack images with a size of 227×277 pixels and 3 channels are classified into training sets and validation sets according to the ratio of 7:3. The images are standardized and enlarged to highlight the crack contrast. The U-net deep convolutional neural network is used to identify cracks and accurately detect crack morphological data. In the downsampling stage of the U-net deep convolutional neural network, the crack features are extracted through secondary convolution and pooling operations. The secondary convolution feature map is copied to the upsampling area with the same number of channels for superposition to complete the skip-layer connection operation. In the upsampling stage, deconvolution and skip-layer connection operations are used to restore the small-size multi-channel feature map to the original input image size to obtain the binary crack morphological data.

[0087] The crack binary morphology map obtained by the U-net deep convolutional neural network is used to obtain the crack binary central axis morphology map through the crack thinning algorithm. The crack pixels in the binary morphology map are classified into endpoints, intersections and simple pixels through the connection relationship with the neighboring pixels. Under the premise of ensuring that the removal of pixels does not affect the original connectivity of the structure and does not generate new components or holes, the pixels are iterated one by one to determine whether to remove them. After all pixels are iterated, a single-pixel crack morphology map is obtained. The crack width is obtained by searching in eight directions (0°, 45°, 90°, 135°, 180°, 225°, 270°, and 315°). The crack points in the single-pixel crack morphology map are used as seed points. The eight-direction search is performed on the crack morphology map through the seed points until the crack boundary is reached. The number of pixels in the eight directions is counted and merged in the four directions of 0°, 45°, 90°, and 135°. The minimum number of pixels in the four directions is taken as the crack width for calibration, and the crack width is obtained by multiplying it with the actual size corresponding to the pixel point. The above completes the model training for crack identification and crack width detection.

[0088] As the robot moves, its camera group will continuously take photos according to the program, upload the photos to the intelligent control system through the synchronization device, automatically identify the cracks using the trained recognition model, and calculate the crack width, and then synchronously transmit the crack location and width information to the intelligent control device. When the robot recognizes a crack, it will stop moving and start the cleaning and repair program.

[0089] Step 3: The crack pre-cleaning device of the robot vehicle starts working. The robot arm 2 equipped with an air gun extends from the cabin. According to the crack position marked by the intelligent control system, it automatically selects a suitable position to approach the crack under the action of the negative pressure suction cup 11 and the wheel 10, and uses the air gun to clean the inside and surface of the crack to reduce the influence of dust and impurities on the crack morphology. Then, the camera re-identifies the cleaned crack and recalculates the characteristic parameters of the crack.

[0090] Step 4: After cleaning, the robot will mobilize the camera to further scan the cracks marked by the intelligent control system. The width is obtained through the recognition model, exported and uploaded to the intelligent control system, and summarized according to the obtained crack morphology information.

[0091] Step 5: The intelligent recognition system automatically determines the crack type based on the above crack width information: If the crack width is less than 0.6 mm, the robot will start the crack repair component, and the first mechanical injection arm 3 will absorb the cementing liquid and the plant urease solution through a micro water pump and mix them in the first mechanical injection arm 3. In order to prevent the injected liquid from gathering at the bottom of the crack due to factors such as gravity, the intelligent control system adopts the strategy of multi-injection site injection and pressurized injection. Injection at multiple injection sites means that the intelligent control system will determine the inclination angle and direction of the crack according to the three-dimensional environmental coordinate system model and the marked crack position, and then select a reasonable position to set the injection site: if it is determined to be a vertical crack type (the inclination angle is 90°), the intelligent control system will evenly divide the crack into several segments according to the length in the vertical direction, and use the highest end point of each segment as the injection site, and the injection order of the injection site is from bottom to top; if it is determined to be a horizontal crack type (the inclination angle is 0°), the intelligent control system will evenly divide the crack into several segments according to the length in the horizontal direction, and use the midpoint of each segment as the injection site, and the injection order of the injection site is from left to right; if it is determined to be an inclined crack type (the inclination angle is greater than 0° and less than 90°), the intelligent control system will perform irregular curve fitting on the crack, generate a line segment function, and evenly divide the line segment function into several segments according to the length, and use the highest point of each segment as the injection site, and the injection order of the injection site is from bottom to top. In addition, the intelligent control system will determine whether the injection liquid fills the crack according to the length and width of the crack in the crack summary information, as well as the real-time image obtained by the robot mobilizing the camera to shoot the injection position during the injection process, and then accurately control the injection volume. The intelligent control system takes the filling of a crack in the line segment where the injection site is located as an indication to start the injection of the next injection site, and continuously performs the above operation until the entire crack is filled. The first mechanical arm 3 works according to the above strategy. The ingredients of the cementing liquid formula are calcium chloride (0.5-1.0 mol / L) and urea (0.5-1.0 mol / L), and the molar ratio of calcium chloride to urea is 1:1. The plant urease solution is made by mixing soybean powder and 30% ethanol solution thoroughly and filtering impurities; if the crack width is greater than 0.6 mm, the robot will start the crack filling component and the crack repair component respectively. First, the crack filling component is started, and the second mechanical injection arm 4 carrying the filler absorbs the filler through a micro water pump. Similar to the control strategy for repairing cracks with a width less than 0.6 mm, in order to avoid the injected filler from gathering at the bottom of the crack due to factors such as gravity, the intelligent control system also adopts the strategy of multi-injection site injection and pressurized injection. The second mechanical injection arm 4 works according to the above strategy. After that, the crack repair component starts the first mechanical injection arm 3 to suck the plant urease solution and the cementing liquid through a micro water pump, and then selects the same position to inject into the crack. When the filling agent injection dose is sufficient (that is, the filling agent fills the crack), it only needs to start the first mechanical injection arm 3 to inject the plant urease solution and the cementing liquid mixed solution with the same strategy until the crack is filled.The filler is generally fine sand, that is, the diameter of the sand is less than or equal to 0.3mm.

[0092] Step 6: After the repair operation is completed, the robot will scan the crack for the second time, mobilize the camera to detect the appearance of the crack after repair, and compare it with the crack before repair. The intelligent recognition system determines whether the crack has been repaired based on the comparison. If the repair is not completed, it will continue to step 5 or report to the operator, who can remotely monitor and control the robot.

[0093] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent wall-climbing robot vehicle for repairing concrete cracks, characterized in that: The vehicle body comprises a vehicle body and a functional component arranged on the vehicle body, wherein the functional component comprises: The robot vehicle synchronous positioning and mapping system is used to build a three-dimensional environmental model of the wall to be repaired; A negative pressure adsorption moving system is used to drive the intelligent wall-climbing robot to climb and move on the wall to be repaired; A crack detection system for identifying the position and width of cracks; the crack detection system includes a trained crack recognition model based on a U-net deep convolutional neural network; the recognition process of the crack recognition model includes: extracting crack features through secondary convolution and pooling operations in the downsampling stage of the U-net deep convolutional neural network, copying the secondary convolution feature map to an upsampling area with the same number of channels for superposition, and completing a skip-layer connection operation; in the upsampling stage, using deconvolution and skip-layer connection operations to restore the small-size multi-channel feature map to the original input image size, and obtaining a crack morphology map; Then, the crack point of the single-pixel crack morphology map is used as the seed point, and the eight-direction search is performed on the crack morphology map through the seed point until the crack boundary is reached. The number of pixels in the eight directions is counted and merged in the four directions of 0°, 45°, 90°, and 135°. The minimum number of pixels is taken as the crack width for calibration, and the crack width is obtained by multiplying it with the actual size corresponding to the pixel point. A crack pre-cleaning system, used for pre-cleaning cracks identified by the crack detection system; Crack plant urease repair system, used to perform urease repair on cracks or to fill and urease repair cracks according to crack width; An intelligent control system is used to control the operation of the robot vehicle synchronous positioning and mapping system, the negative pressure adsorption movement system, the crack detection system, the crack pre-cleaning system and the crack plant urease repair system.

2. The intelligent wall-climbing robot vehicle for repairing concrete cracks according to claim 1 is characterized in that: The crack plant urease repair system comprises a crack repair component and a crack filling component; The crack repair assembly includes a first mechanical injection arm carrying a cementing liquid and a plant urease solution, and is used to inject a plant urease-based repair agent into the crack; The crack filling assembly includes a second mechanical injection arm carrying a filling agent, which is used to inject the filling agent into the crack.

3. The intelligent wall-climbing robot vehicle for repairing concrete cracks according to claim 2 is characterized in that: When the width of the crack is less than 0.6 mm, the intelligent control system controls the crack repair component to inject a plant urease-based repair agent into the crack; when the width of the crack is greater than or equal to 0.6 mm, the intelligent control system controls the crack filling component to fill the crack with the filling agent, and then controls the crack repair component to inject a plant urease-based repair agent into the crack.

4. The intelligent wall-climbing robot vehicle for repairing concrete cracks according to claim 2 is characterized in that: The negative pressure adsorption mobile system includes a negative pressure suction cup and wheels arranged at the bottom of the intelligent wall-climbing robot vehicle; The robot vehicle synchronous positioning and mapping system includes a camera, a gyroscope and a positioning module; the camera is arranged on the front side of the intelligent wall-climbing robot vehicle; The crack pre-cleaning system includes a mechanical arm equipped with an air gun; one end of the mechanical arm is fixed in the cabin of the intelligent wall-climbing robot vehicle, and the other end extends out from the front side of the intelligent wall-climbing robot vehicle; One end of the first mechanical injection arm and the second mechanical injection arm are installed side by side in the cabin of the intelligent wall-climbing robot vehicle, and the other end extends out from the front side of the intelligent wall-climbing robot vehicle; the cabin is also provided with a plant urease solution tank and a binder tank connected to the first mechanical injection arm, and a filler tank connected to the second mechanical injection arm.

5. An intelligent repair method for concrete wall cracks, characterized in that: The intelligent wall-climbing robot vehicle according to any one of claims 1 to 4 comprises the following steps: S1. According to the structural shape and geographical location of the wall to be repaired, a position is selected to attach the robot vertically to the wall, and then the wall to be repaired is scanned to construct a three-dimensional environmental model of the wall to be repaired; S2. The intelligent control system plans a moving route according to the three-dimensional environment model, and the intelligent wall-climbing robot moves on the wall to be repaired according to the moving route, and simultaneously continuously scans the wall to be repaired to obtain image information of the wall to be repaired; S3, the crack detection system identifies the location of the crack according to the image information. When the crack is identified, the intelligent wall-climbing robot stops moving and starts the crack pre-cleaning system to pre-clean the crack; S4, the crack detection system recognizes the cracks after pre-cleaning again, and re-acquires the position and width of the cracks; the crack detection system recognizes the cracks through a crack recognition model based on a U-net deep convolutional neural network; The U-net deep convolutional neural network first obtains a crack morphology map, then uses the crack point of the single-pixel crack morphology map as a seed point, performs an eight-direction search on the crack morphology map through the seed point, and stops at the crack boundary, counts the number of pixels in the eight directions and merges them in the four directions of 0°, 45°, 90°, and 135°, takes the minimum number of pixels as the crack width for calibration, and multiplies it by the actual size corresponding to the pixel to obtain the crack width; S5. The intelligent control system starts the crack plant urease repair system to perform urease repair on the crack or to fill and repair the crack according to the width of the crack; S6. After the current cracks are repaired, the intelligent wall-climbing robot continues to move according to the moving route and repeats steps S3 to S5 until all cracks in the repair wall are repaired.

6. The intelligent repair method for concrete wall cracks according to claim 5 is characterized in that: In step S5, when the width of the crack is less than 0.6 mm, the intelligent control system controls the crack repair component to inject a plant urease-based repair agent into the crack; when the width of the crack is greater than or equal to 0.6 mm, the intelligent control system controls the crack filling component to fill the crack with the filling agent, and then controls the crack repair component to inject a plant urease-based repair agent into the crack.

7. The intelligent repair method for concrete wall cracks according to claim 6 is characterized in that: The plant urease-based repair agent comprises a cementing liquid and a plant urease solution; The binder solution comprises 0.5-1.0 mol / L calcium chloride and 0.5-1.0 mol / L urea, and the molar ratio of calcium chloride to urea is 1:1; the plant urease solution is obtained by fully mixing soybean powder with a mixed solution of ethanol and water in a volume ratio of 30%:70% and then filtering; The mixing volume ratio of the cementing liquid and the plant urease solution is 1:1; The filler is fine sand with a diameter less than or equal to 0.3 mm.

8. The intelligent repair method for concrete wall cracks according to claim 5 is characterized in that: In steps S1 and S2, the scanning is performed by a camera; In step S2, while continuously scanning the wall to be repaired, the robot vehicle synchronous positioning and mapping system continuously establishes and improves the three-dimensional environment model; In step S3, the pre-cleaning is to use an air gun to blow and clean the inside and surface of the crack; In step S4, the position and width of the crack are obtained by the crack detection system; In step S5, the intelligent control system determines the inclination state of the crack according to the three-dimensional environmental model, divides the crack into several sections according to the inclination state, and performs urease repair or filling and urease repair on each section in turn.

Images