Ship cleaning robot path guiding method based on visual identification

Abstract

The invention discloses a ship cleaning robot path guiding method based on visual identification, and relates to the technical field of ship cleaning robots, and the method comprises the steps: obtaining ship wall region image data, and carrying out the analysis and calculation, and obtaining a ship wall stain index of a ship wall region; according to the method, the emphasized cleaning area, the general cleaning area and the no-clean area are accurately distinguished in combination with ship wall stain index calculation and a double-threshold division rule, the problem that traditional cleaning paths are not completely covered is solved, differential cleaning is achieved, invalid operation is avoided, secondly, a plurality of cleaning paths are generated by means of a path generation algorithm, and the cleaning efficiency is improved. The optimal path is screened by comprehensively considering the total cleaning path distance, the ship wall area surface curvature and the stain degree based on the path optimization index, in addition, the visual sensor continuously collects images of the current position, the path deviation index is calculated in real time, and the movement track is dynamically adjusted. The defects that a traditional robot lacks an effective visual feedback and intelligent decision-making mechanism, and deviation is prone to being generated in path tracking are overcome.

Description

Technical Field

[0001] This invention relates to the field of ship cleaning robot technology, and in particular to a path guidance method for ship cleaning robots based on vision recognition. Background Technology

[0002] In the field of modern water transport and marine engineering, ships, as core transportation tools, are exposed to complex marine environments for extended periods. Their hull surfaces are susceptible to marine organism attachment, corrosion, and pollutant accumulation, which not only increases navigation resistance and reduces fuel efficiency but may also accelerate metal structure corrosion, threatening ship safety. Therefore, regular and efficient cleaning and maintenance of ships has become a key aspect of ensuring ship lifespan and operational efficiency. Against this backdrop, cleaning robots, with their advantage of being able to replace manual labor in high-risk and complex cleaning operations, can effectively reduce safety risks such as falls and chemical contact. At the same time, through automated operations, they significantly improve cleaning efficiency and quality consistency, making them an important development direction in the field of ship maintenance.

[0003] However, traditional path planning methods rely heavily on preset programs or manual remote control, lacking the ability to perceive and adapt to the actual cleaning environment in real time. This results in fixed cleaning paths, incomplete coverage, and difficulty in dynamically avoiding obstacles. In addition, due to the lack of effective visual feedback and intelligent decision-making mechanisms, traditional robots are prone to deviations when tracking paths on complex ship hull surfaces, thus affecting the overall cleaning efficiency and quality.

[0004] To address the aforementioned technical deficiencies, a solution is proposed. Summary of the Invention

[0005] The purpose of this invention is to address the problems of traditional path planning methods, which rely heavily on preset programs or manual remote control and lack real-time perception and adaptability to the actual cleaning environment. This results in fixed cleaning paths, incomplete coverage, and difficulty in dynamically avoiding obstacles. Furthermore, due to the lack of effective visual feedback and intelligent decision-making mechanisms, traditional robots are prone to deviations when tracking paths on complex ship surfaces, thus affecting the overall cleaning efficiency and quality.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a path guidance method for a ship cleaning robot based on visual recognition, comprising the following steps:

[0007] Step 1: The ship cleaning robot collects images of the ship surface in real time through its onboard vision sensors and performs image preprocessing to divide the ship surface into several ship wall areas. It then acquires image data of the ship wall areas and performs analysis and calculation to obtain the ship wall stain index of the ship wall areas.

[0008] Step 2: Generate several cleaning paths using a path generation algorithm, obtain optimal data using image recognition technology, and analyze and calculate the path optimization index of the cleaning robot. Through comparative analysis, select the optimal cleaning path for the robot.

[0009] Step 3: Based on the robot's optimal cleaning path, generate control commands using a PID control algorithm to drive the robot to crawl and clean along the optimal cleaning path. At the same time, maintain the connection with the operator terminal through wires to achieve remote monitoring.

[0010] Step 4: When the robot is performing the cleaning task, the vision sensor continuously collects images of the current position, provides real-time feedback on the deviation between the actual path and the planned path, and dynamically adjusts the subsequent motion trajectory to ensure the accuracy of path guidance.

[0011] Furthermore, the image data includes the dimensions of the ship surface image and the pixel value data of the ship surface image, and the preferred data includes the surface curvature of the ship wall area and the total distance data of the cleaning path.

[0012] Furthermore, image preprocessing includes Gaussian filtering for noise reduction and histogram equalization to enhance contrast, in order to improve image quality.

[0013] Furthermore, the calculation process for the ship's wall stain index in the hull area is as follows:

[0014] S11. Obtain the dimensions of the ship's surface image and the pixel value data of the ship's surface image, and perform analysis and calculation;

[0015] S12. Calculate the ship wall stain index G for the ship wall area according to the following formula:

[0016]

[0017] Where M is the horizontal dimension of the processed ship surface image, N is the vertical dimension of the processed ship surface image, and S... (x,y) Let (x, y) be the pixel value of the ship's surface image at coordinates (x, y). The standard pixel value of the ship surface image at coordinates (x,y) is a preset value. The ship wall stain index is used to reflect the severity of stains in each ship wall area.

[0018] S13. Obtain the preset upper limit threshold G for ship wall stains. top Lower limit threshold G for ship wall stains low The analysis was compared with the ship wall stain index G in the ship wall area. When G≥G top If the stains on the ship's hull are severe, it indicates that the area is designated as a key cleaning area.

[0019] S14, when G∈(Glow G top If the stains on the ship's wall area are of moderate severity, it is classified as a general cleaning area.

[0020] S15, when G≤G th If the stains are low, it indicates that the stains on the ship's side are not severe and the area is designated as a no-clean zone.

[0021] Furthermore, the calculation process for the path optimization index of the cleaning robot is as follows:

[0022] S21. Obtain and analyze the surface curvature of the ship's wall area and the total distance of the cleaning path.

[0023] S22. Calculate the path optimization index F of the cleaning robot according to the following formula:

[0024]

[0025] Where L is the total path length of the generated cleaning path, L avg Let C be the average distance of all generated cleaning paths, n be the number of ship hull regions included in the generated cleaning paths, and C be the average distance of all generated cleaning paths. i Let C be the surface curvature of the i-th hull region. max C represents the maximum surface curvature of the hull region encompassed by the cleaning path. min G is the minimum surface curvature in the hull region encompassed by the cleaning path. i Let G be the wall stain index of the i-th wall region. e The preset standard ship wall stain index;

[0026] S23. The path optimization index of the cleaning robot is used to reflect the cleaning complexity of each cleaning path. The path optimization index of each cleaning path is obtained and sorted in ascending order according to the value. The cleaning path with the first ranking is selected as the optimal cleaning path of the robot.

[0027] Furthermore, the calculation process for the deviation between the actual path and the planned path is as follows:

[0028] S31. Obtain and analyze the actual path and planned path data of the robot.

[0029] S32. Calculate the path deviation index E of the robot's movement according to the following formula: Where T is the total number of path deviation samples, e t Let t be the deviation between the robot's actual position and its planned position at the t-th sampling time.

[0030] S33. Obtain the preset path deviation threshold E thA comparative analysis was conducted with the path deviation index E of the robot's movement. When E ≥ E th If the actual path of the cleaning robot deviates significantly from the planned path, then dynamic adjustments to the subsequent movement trajectory of the cleaning robot are necessary.

[0031] S34. When E <E th If the actual path of the cleaning robot deviates little from the planned path, then there is no need to dynamically adjust the subsequent movement trajectory of the cleaning robot.

[0032] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0033] This method for guiding ship cleaning robots based on vision recognition acquires real-time images of the ship's surface using a vision sensor. Image quality is optimized through Gaussian filtering for noise reduction and histogram equalization preprocessing. Combined with the calculation of the ship's wall dirt index and a dual-threshold classification rule, it accurately distinguishes between areas requiring intensive cleaning, general cleaning, and areas that do not require cleaning. This solves the problem of incomplete coverage in traditional cleaning paths, achieves differentiated cleaning, avoids ineffective work, and significantly improves cleaning targeting and resource utilization. Secondly, multiple cleaning paths are generated using a path generation algorithm. A path selection index is used to comprehensively consider the total distance of the cleaning path, the surface curvature of the ship's wall area, and the degree of dirt on the selected path. The optimal path overcomes the limitations of traditional path planning, which relies on preset programs and lacks real-time environmental awareness, and can dynamically adapt to complex ship surface environments. At the same time, it generates precise control commands through PID control algorithms to drive the robot's crawling, and connects to the operating end via wires to achieve remote monitoring, enhancing operational controllability and safety, and reducing the risk of human intervention. In addition, the visual sensor continuously collects images of the current position, calculates the path deviation index in real time, and dynamically adjusts the motion trajectory, making up for the shortcomings of traditional robots that lack effective visual feedback and intelligent decision-making mechanisms and are prone to deviations in path tracking, ensuring the accuracy of path guidance and comprehensively improving the efficiency of ship cleaning. Attached Figure Description

[0034] Figure 1 A schematic diagram of the method flow of the present invention is shown. Detailed Implementation

[0035] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0036] Example:

[0037] like Figure 1As shown, a path guidance method for a ship cleaning robot based on vision recognition is proposed. First, the ship cleaning robot acquires real-time images of the ship's surface using its onboard vision sensors. Through image preprocessing, the ship's surface is divided into several wall regions, and image data of these wall regions is acquired and analyzed to derive the wall dirt index for each region. It should be noted that the image preprocessing includes Gaussian filtering for noise reduction and histogram equalization to enhance contrast, thereby improving image quality. The image data includes the dimensions of the ship's surface image and its pixel values.

[0038] The calculation process for the ship's wall stain index in the hull area is as follows:

[0039] S11. Obtain the dimensions of the ship's surface image and the pixel value data of the ship's surface image, and perform analysis and calculation;

[0040] S12. Calculate the ship wall stain index G for the ship wall area according to the following formula:

[0041]

[0042] Where M is the horizontal dimension of the processed ship surface image, N is the vertical dimension of the processed ship surface image, and S... (x,y) Let (x, y) be the pixel value of the ship's surface image at coordinates (x, y). The standard pixel value of the ship surface image is set at coordinates (x,y). The ship wall stain index is used to reflect the severity of stains in each ship wall area. The higher the value of the ship wall stain index, the higher the severity of stains in the ship wall area.

[0043] S13. Obtain the preset upper limit threshold G for ship wall stains. top Lower limit threshold G for ship wall stains low The analysis was compared with the ship wall stain index G in the ship wall area. When G≥G top If the stains on the ship's hull are severe, it indicates that the area is designated as a key cleaning area and requires a second cleaning by a cleaning robot.

[0044] S14, when G∈(G low G top If the stains on the ship's hull are normal, it indicates that the stains are of moderate severity and the area is classified as a general cleaning area, requiring a cleaning robot to perform a single cleaning.

[0045] S15, when G≤G th If the stains on the ship's hull are low, it indicates that the area is not heavily soiled and is classified as a no-clean zone, meaning there is no need to send a cleaning robot to clean it.

[0046] Then, several cleaning paths are generated through a path generation algorithm. Optimal data is obtained through image recognition technology. The optimal data includes the surface curvature of the ship's wall area and the total distance of the cleaning path. The path optimization index of the cleaning robot is calculated and analyzed. The optimal cleaning path of the robot is selected through comparative analysis.

[0047] The calculation process for the path optimization index of the cleaning robot is as follows:

[0048] S21. Obtain and analyze the surface curvature of the ship's wall area and the total distance of the cleaning path.

[0049] S22. Calculate the path optimization index F of the cleaning robot according to the following formula:

[0050]

[0051] Where L is the total path length of the generated cleaning path, L avg Let C be the average distance of all generated cleaning paths, n be the number of ship hull regions included in the generated cleaning paths, and C be the average distance of all generated cleaning paths. i Let C be the surface curvature of the i-th hull region. max C represents the maximum surface curvature of the hull region encompassed by the cleaning path. min G is the minimum surface curvature in the hull region encompassed by the cleaning path. i Let G be the wall stain index of the i-th wall region. e The preset standard ship wall stain index;

[0052] S23. The path selection index of the cleaning robot is used to reflect the cleaning complexity of each cleaning path. The larger the value of the path selection index, the higher the cleaning complexity of the cleaning path. Obtain the path selection index of each cleaning path and sort them in ascending order according to the value. Select the cleaning path with the first ranking as the optimal cleaning path of the robot.

[0053] Then, based on the robot's optimal cleaning path, control commands are generated through a PID control algorithm to drive the robot to crawl and clean along the optimal cleaning path. At the same time, the connection with the operator is maintained through wires to achieve remote monitoring.

[0054] Finally, when the robot is performing the cleaning task, the vision sensor continuously collects images of the current position, provides real-time feedback on the deviation between the actual path and the planned path, and dynamically adjusts the subsequent movement trajectory to ensure the accuracy of path guidance.

[0055] The calculation process for the deviation between the actual path and the planned path is as follows:

[0056] S31. Obtain and analyze the actual path and planned path data of the robot.

[0057] S32. Calculate the path deviation index E of the robot's movement according to the following formula: Where T is the total number of path deviation samples, e t Let t be the deviation between the robot's actual position and its planned position at the t-th sampling time.

[0058] S33. Obtain the preset path deviation threshold E th A comparative analysis was conducted with the path deviation index E of the robot's movement. When E ≥ E th If the actual path of the cleaning robot deviates significantly from the planned path, then dynamic adjustments to the subsequent movement trajectory of the cleaning robot are necessary.

[0059] S34. When E <E th If the actual path of the cleaning robot deviates little from the planned path, then there is no need to dynamically adjust the subsequent movement trajectory of the cleaning robot.

[0060] This invention uses a visual sensor to acquire real-time images of the ship's surface. After Gaussian filtering for noise reduction and histogram equalization preprocessing to optimize image quality, and combined with the calculation of the ship's wall dirt index and a dual-threshold classification rule, it accurately distinguishes between areas requiring intensive cleaning, general cleaning, and areas that do not require cleaning. This solves the problem of incomplete coverage in traditional cleaning paths, achieves differentiated cleaning, avoids ineffective work, and significantly improves cleaning targeting and resource utilization. Secondly, it uses a path generation algorithm to generate multiple cleaning paths. Based on a path optimization index, it comprehensively considers the total path distance, the surface curvature of the ship's wall area, and the degree of dirt to select the optimal path. This overcomes the limitations of traditional path planning, which relies on preset programs and lacks real-time environmental awareness, and can dynamically adapt to complex ship surface environments. Simultaneously, a PID control algorithm generates precise control commands to drive the robot's crawling, and a wire connects to the operating terminal for remote monitoring, enhancing operational controllability and safety, and reducing the risk of human intervention. Furthermore, the visual sensor continuously acquires images of the current position, calculates the path deviation index in real time, and dynamically adjusts the movement trajectory. This compensates for the shortcomings of traditional robots, such as the lack of effective visual feedback and intelligent decision-making mechanisms, and the tendency for path tracking to deviate, ensuring accurate path guidance and comprehensively improving the efficiency of ship cleaning.

[0061] The size of the interval and threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by those skilled in the art for each set of sample data; as long as it does not affect the ratio between the parameter and the quantized value.

[0062] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0063] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A path guidance method for a ship cleaning robot based on vision recognition, characterized in that, Includes the following steps: Step 1: The ship cleaning robot collects images of the ship surface in real time through its onboard vision sensors and performs image preprocessing to divide the ship surface into several ship wall areas. It then acquires image data of the ship wall areas and performs analysis and calculation to obtain the ship wall stain index of the ship wall areas. Step 2: Generate several cleaning paths using a path generation algorithm, obtain optimal data using image recognition technology, and analyze and calculate the path optimization index of the cleaning robot. Through comparative analysis, select the optimal cleaning path for the robot. Step 3: Based on the robot's optimal cleaning path, generate control commands using a PID control algorithm to drive the robot to crawl and clean along the optimal cleaning path. At the same time, maintain the connection with the operator terminal through wires to achieve remote monitoring. Step 4: When the robot is performing the cleaning task, the vision sensor continuously collects images of the current position, provides real-time feedback on the deviation between the actual path and the planned path, and dynamically adjusts the subsequent motion trajectory to ensure the accuracy of path guidance.

2. The path guidance method for a ship cleaning robot based on vision recognition according to claim 1, characterized in that, The image data includes the dimensions of the ship surface image and the pixel value data of the ship surface image. The preferred data includes the surface curvature of the ship wall area and the total distance of the cleaning path.

3. The path guidance method for a ship cleaning robot based on vision recognition according to claim 1, characterized in that, Image preprocessing includes Gaussian filtering for noise reduction and histogram equalization to enhance contrast, in order to improve image quality.

4. The path guidance method for a ship cleaning robot based on vision recognition according to claim 1, characterized in that, The calculation process for the ship's wall stain index in the hull area is as follows: S11. Obtain the dimensions of the ship's surface image and the pixel value data of the ship's surface image, and perform analysis and calculation; S12. Calculate the ship wall stain index G for the ship wall area according to the following formula: Where M is the horizontal dimension of the processed ship surface image, N is the vertical dimension of the processed ship surface image, and S... (x,y) Let (x, y) be the pixel value of the ship's surface image at coordinates (x, y). The standard pixel value of the ship surface image at coordinates (x,y) is a preset value. The ship wall stain index is used to reflect the severity of stains in each ship wall area. S13. Obtain the preset upper limit threshold G for ship wall stains. top Lower limit threshold G for ship wall stains low The analysis was compared with the ship wall stain index G in the ship wall area. When G≥G top If the stains on the ship's hull are severe, it indicates that the area is designated as a key cleaning area. S14, when G∈(G low G top If the stains on the ship's wall area are of moderate severity, it is classified as a general cleaning area. S15, when G≤G th If the stains are low, it indicates that the stains on the ship's side are not severe and the area is designated as a no-clean zone.

5. The path guidance method for a ship cleaning robot based on vision recognition according to claim 1, characterized in that, The calculation process for the path optimization index of the cleaning robot is as follows: S21. Obtain and analyze the surface curvature of the ship's wall area and the total distance of the cleaning path. S22. Calculate the path optimization index F of the cleaning robot according to the following formula: Where L is the total path length of the generated cleaning path, L avg Let C be the average distance of all generated cleaning paths, n be the number of ship hull regions included in the generated cleaning paths, and C be the average distance of all generated cleaning paths. i Let C be the surface curvature of the i-th hull region. max C represents the maximum surface curvature of the hull region encompassed by the cleaning path. min G is the minimum surface curvature in the hull region encompassed by the cleaning path. i Let G be the wall stain index of the i-th wall region. e The preset standard ship wall stain index; S23. The path optimization index of the cleaning robot is used to reflect the cleaning complexity of each cleaning path. The path optimization index of each cleaning path is obtained and sorted in ascending order according to the value. The cleaning path with the first ranking is selected as the optimal cleaning path of the robot.

6. The path guidance method for a ship cleaning robot based on vision recognition according to claim 1, characterized in that, The calculation process for the deviation between the actual path and the planned path is as follows: S31. Obtain and analyze the actual path and planned path data of the robot. S32. Calculate the path deviation index E of the robot's movement according to the following formula: Where T is the total number of path deviation samples, e t Let t be the deviation between the robot's actual position and its planned position at the t-th sampling time. S33. Obtain the preset path deviation threshold E th A comparative analysis was conducted with the path deviation index E of the robot's movement. When E ≥ E th If the actual path of the cleaning robot deviates significantly from the planned path, then dynamic adjustments to the subsequent movement trajectory of the cleaning robot are necessary. S34. When E <E th If the actual path of the cleaning robot deviates little from the planned path, then there is no need to dynamically adjust the subsequent movement trajectory of the cleaning robot.

Images