A machine intelligence-based autonomous decision-making support system for water environment cleaning

Abstract

The invention belongs to the technical field of intelligent environmental protection. The invention provides an autonomous decision-making auxiliary system for water environment cleaning based on machine intelligence. The image of the bottom of a ship is collected to obtain a grid area image. The grid area image is divided into a plurality of cleaning areas. The attachment data in each cleaning area is processed. The attachment data includes an attachment distribution value and an attachment area ratio. The attachment distribution value and the attachment area ratio are processed to obtain an attachment degree value. The preliminary planning of a cleaning path is completed according to the attachment degree value. Therefore, by analyzing the proportion and distribution of attachments in each cleaning area, the area proportion and aggregation degree of attachments in the cleaning area are comprehensively evaluated. According to the size of the attachment degree value corresponding to each cleaning area, the preliminary planning of the cleaning path is completed.

Description

Technical Field

[0001] The present invention belongs to the technical field of intelligent environmental protection, and in particular is a water environment cleaning autonomous decision-making auxiliary system based on machine intelligence. Background Art

[0002] Marine microorganisms attached to the bottom of the ship for a long time will secrete chemicals that corrode the bottom of the ship. A large number of organisms and dirt attached to the bottom of the ship will increase the weight of the ship, which will affect the speed and increase fuel consumption. Microorganisms corrode the paint on the bottom of the ship and accelerate the wear and tear of the hull. Therefore, with the assistance of the water environment cleaning autonomous decision-making assistance system based on machine intelligence, the data of attachments in each required cleaning area on the bottom of the ship is collected, analyzed and processed, and the cleaning route is determined to plan the cleaning of each required cleaning area on the bottom of the ship.

[0003] In the prior art, the cleaning route is obtained by acquiring data of attachments in each required cleaning area on the bottom of the ship and analyzing and processing the data of attachments in each required cleaning area on the bottom of the ship. However, after the cleaning route is preliminarily planned, the position distance of each required cleaning area in the preliminary cleaning route is random, and the distance between two adjacent cleaning areas in the preliminary cleaning route is uncertain. Therefore, it is necessary to analyze whether the order of the cleaning areas in the preliminary cleaning path on the cleaning path is reasonable. If it is unreasonable, it is necessary to optimize it, compare the path deviation value corresponding to the optimized cleaning path with the path deviation value corresponding to the preliminary cleaning path, determine the final cleaning path, and realize decision assistance in determining the optimal path.

[0004] To this end, the present invention provides a water environment cleaning autonomous decision-making auxiliary system based on machine intelligence. Summary of the invention

[0005] In order to make up for the shortcomings of the prior art, the problem of incomplete cleaning of floating objects due to changes in the distribution of floating objects in the required cleaning area proposed in the background art is solved.

[0006] The technical solution adopted by the present invention to solve its technical problem is:

[0007] An autonomous decision-making support system for water environment cleaning based on machine intelligence. The specific process is as follows:

[0008] Preliminary planning module: collect images of the bottom of the ship to obtain grid area images, divide the grid area images into several cleaning areas, process the attachment data in each cleaning area, where the attachment data includes the attachment area and the coordinates of the attachment, obtain the attachment distribution value and the attachment area ratio respectively, process the attachment distribution value and the attachment area ratio to obtain the attachment degree value, and complete the preliminary planning of the cleaning path according to the attachment degree value;

[0009] Evaluation planning module: Based on the preliminary cleaning path, obtain the non-recent cleaning quantity ratio and the non-recent cleaning degree value, process the non-recent cleaning quantity ratio and the non-recent cleaning degree value, output the planning analysis value, and compare it with the threshold to generate a path analysis signal;

[0010] Wherein, the path analysis signal includes a path analysis signal or a path non-analysis signal;

[0011] Path analysis module: Based on the path analysis signal, the cleaning path is re-planned to obtain a reference cleaning path. Based on the deviation analysis of the reference cleaning path and the preliminary cleaning path, the path deviation performance value is obtained, and the final cleaning path is determined based on the path deviation performance value.

[0012] As a further technical solution of the present invention: the method for obtaining the adhesion area ratio is:

[0013] Obtain the area of ​​the attachments in the cleaning area, add up the areas of the attachments in the cleaning area, and obtain the total area of ​​the attachments;

[0014] The ratio of the total area of ​​the attachments to the area of ​​the cleaning region is calculated to obtain the attachment area ratio.

[0015] As a further technical solution of the present invention: the method for obtaining the attachment distribution value is:

[0016] The intersection of the two intersecting boundary lines at the edge of the cleaning area is taken as the origin of the two-dimensional coordinate system. A straight line in the horizontal direction is drawn with the origin as the starting point as the X axis of the two-dimensional coordinate system. A straight line in the vertical direction is drawn with the origin as the starting point as the Y axis of the two-dimensional coordinate system to obtain the cleaning area coordinate system.

[0017] In the cleaning area coordinate system, obtain the midpoint coordinates in the cleaning area, and mark the midpoint coordinates in the cleaning area as the cleaning midpoint coordinates ( , ), where z represents the midpoint;

[0018] Get the coordinates of each attachment in the cleaning area coordinate system ( , ), where n represents the number of all attachments in the cleaning area coordinate system;

[0019] Get the coordinates of each attachment in the cleaning area coordinate system ( , ) and the cleaning midpoint coordinates ( , ), marked as the attachment distance value;

[0020] The coordinates of each attachment in the cleaning area coordinate system ( , ) and the cleaning midpoint coordinates ( , ) into the formula: , calculate and obtain the floating distance value L;

[0021] All floating object distance values ​​L are added and averaged to obtain the mean floating distance, the perimeter of the cleaning area is measured, and the mean floating distance is ratioed with the perimeter of the cleaning area to obtain the attachment distribution value.

[0022] As a further technical solution of the present invention: the method for obtaining the preliminary cleaning path is:

[0023] The attachment area ratio is calculated with the attachment distribution value to obtain the attachment degree value;

[0024] Based on the adhesion degree value, the adhesion degree value corresponding to each key cleaning area is obtained, the adhesion degree value corresponding to each key cleaning area is compared, and the cleaning is carried out in order from large to small until the bottom attachments in the cleaning area are cleaned and a preliminary cleaning path is obtained.

[0025] As a further technical solution of the present invention: the planning analysis value is obtained by:

[0026] Substitute the non-recent cleaning quantity ratio and non-recent cleaning degree values ​​into the formula: , calculate the planning analysis value H, where It is expressed as the ratio of the number of non-recent cleanings, and L is expressed as the non-recent cleaning degree value. , Expressed as a preset scale factor, The value can be 0.4662. The value can be 0.5338.

[0027] As a further technical solution of the present invention: the method for obtaining the non-recent cleaning quantity ratio is:

[0028] When the preliminary cleaning path is obtained, all cleaning areas are combined to obtain multiple cleaning area groups;

[0029] Based on any cleaning area group;

[0030] In the cleaning area group, obtain the distance between two cleaning areas, marked as cleaning distance;

[0031] In the cleaning area group, the cleaning area with the highest ranking is marked as the front cleaning area, and the distances between the front cleaning area and all other cleaning areas are obtained and marked as other cleaning distances;

[0032] If all other cleaning distances are greater than the cleaning distance, the cleaning area group is marked as the nearest cleaning area group;

[0033] If all other cleaning distances are not greater than the cleaning distance, the cleaning area group is marked as a non-closest cleaning area group;

[0034] The number of non-recent cleaning area groups is counted and the ratio is processed with the total number of cleaning area groups to obtain the non-recent cleaning number ratio, which is marked as SL.

[0035] As a further technical solution of the present invention: the method for obtaining the non-recent cleaning degree value is:

[0036] Based on any non-recently cleaned area group;

[0037] Obtain the closest cleaning area adjacent to the front cleaning area in the non-closest cleaning area group, and measure the distance between the front cleaning area and the adjacent closest cleaning area, marking it as the closest cleaning distance;

[0038] Perform difference processing on the cleaning distance of the non-closest cleaning area group and the closest cleaning distance, and take the absolute value of the difference to obtain the cleaning distance difference of the non-closest cleaning area group, and sum and average the cleaning distance differences of all non-closest cleaning area groups to obtain the mean cleaning distance difference;

[0039] Get the cleaning distances of all cleaning area groups and sum them up to get the total cleaning distance;

[0040] The mean cleaning distance difference is ratioed to the total cleaning distance to obtain the non-closest cleaning degree value, which is marked as CD.

[0041] As a further technical solution of the present invention: the planning analysis value is compared with the planning analysis threshold, and the comparison process is as follows:

[0042] If the planning analysis value is greater than or equal to the planning analysis threshold, a path analysis signal is generated;

[0043] If the planning analysis value is less than the planning analysis threshold, a path is generated without analyzing the signal;

[0044] If the signal is not analyzed based on the path, the preliminary cleaning path is used as the final cleaning path.

[0045] As a further technical solution of the present invention: the reference cleaning path is obtained in the following manner:

[0046] Combining the first cleaning area with other cleaning areas respectively to obtain a cleaning area unit group;

[0047] Based on any cleaning area unit group;

[0048] In the cleaning area unit group, the adhesion degree values ​​corresponding to the two cleaning areas are subjected to difference processing, and the absolute value of the difference is taken to obtain the adhesion difference corresponding to the cleaning area unit group;

[0049] The front cleaning area in the cleaning area unit group is marked as the front cleaning area, and the adhesion difference corresponding to the cleaning area unit group is processed with the adhesion degree value of the front cleaning area to obtain the adhesion proximity value;

[0050] In the cleaning area unit group, the distance between two cleaning areas is obtained to obtain the cleaning distance of the cleaning area unit group, the cleaning distances of all cleaning area unit groups are obtained and summed to obtain the total cleaning distance, and the cleaning distance of the cleaning area unit group is processed by ratio with the total cleaning distance to obtain the cleaning distance ratio;

[0051] The attachment proximity value and the cleaning distance ratio are summed to obtain the priority value of the cleaning area unit group;

[0052] Compare the priority values ​​of all cleaning area unit groups, and select the cleaning area unit group corresponding to the minimum priority value as the target unit group;

[0053] Two cleaning areas in the target unit group are regarded as adjacent cleaning areas, wherein the positions of the adjacent cleaning areas in the reference cleaning path are adjacent;

[0054] When the cleaning area adjacent to the first cleaning area is determined, the cleaning area adjacent to the first cleaning area is used again as the second cleaning area;

[0055] The second cleaning area is processed in the same manner as the first cleaning area, and cleaning areas adjacent to the second cleaning area are determined until all cleaning areas are determined, thereby obtaining a reference cleaning path.

[0056] As a further technical solution of the present invention, the specific process of performing deviation analysis on the reference cleaning path and the preliminary cleaning path to obtain the optimized cleaning path is as follows:

[0057] In the reference cleaning path and the preliminary cleaning path, the cleaning areas with the same ranking are integrated into cleaning groups with the same ranking;

[0058] In the same ranking cleaning group, if the positions of the two cleaning areas in the grid area image are different, the same ranking cleaning group is marked as a non-overlapping group; if the positions of the two cleaning areas in the grid area image are the same, the same ranking cleaning group is marked as an overlapping group;

[0059] Count the number of non-overlapping groups and compare them with the number of cleaned groups with the same ranking to obtain the path deviation performance value;

[0060] comparing the path deviation performance value to a path deviation performance threshold;

[0061] If the path deviation performance value is greater than the path deviation performance threshold, the reference cleaning path is used as the final cleaning path;

[0062] If the path deviation performance value is less than or equal to the path deviation performance threshold, the reference cleaning path or the preliminary cleaning path is used as the final cleaning path.

[0063] The beneficial effects of the present invention are as follows:

[0064] (1) The present invention collects images of the bottom of the ship to obtain a grid area image, divides the grid area image into a plurality of cleaning areas, processes the attachment data in each cleaning area, the attachment data includes an attachment distribution value and an attachment area ratio, processes the attachment distribution value and the attachment area ratio to obtain an attachment degree value, and completes the preliminary planning of the cleaning path according to the attachment degree value, thereby comprehensively evaluating the area proportion and aggregation degree of the attachments in the cleaning area by analyzing the proportion and distribution of the attachments in each cleaning area, and completing the preliminary planning of the cleaning path according to the size of the attachment degree value corresponding to each cleaning area;

[0065] (2) Based on the preliminary cleaning path, the present invention obtains the non-recent cleaning quantity ratio and the non-recent cleaning degree value, substitutes the non-recent cleaning quantity ratio and the non-recent cleaning degree value into the formula, and outputs the planning analysis value. If the planning analysis value is greater than or equal to the planning analysis threshold, it means that the sorting of the cleaning areas in the preliminary cleaning path on the path is relatively unreasonable, thereby reflecting the rationality of the distance sorting of the cleaning areas in the preliminary cleaning path by analyzing the non-recent cleaning quantity ratio and the non-recent cleaning degree value;

[0066] (3) The present invention optimizes the preliminary cleaning path based on the unreasonable path signal to obtain the optimized cleaning path, obtains the optimized reasonable value, compares the path deviation value corresponding to the optimized cleaning path with the path deviation threshold, and determines the final cleaning path. Thus, the final cleaning path is determined by comparing the size of the optimized reasonable value and the path deviation value, thereby realizing decision assistance in determining the final path. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] The present invention will be further described below in conjunction with the accompanying drawings.

[0068] Figure 1 It is a module diagram of a water environment cleaning autonomous decision-making auxiliary system based on machine intelligence according to an embodiment of the present invention;

[0069] Figure 2It is a flow chart of a method for determining a final cleaning path in a water environment cleaning autonomous decision-making support system based on machine intelligence according to an embodiment of the present invention. DETAILED DESCRIPTION

[0070] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0071] Example 1

[0072] like Figure 1-2 As shown, an autonomous decision-making support system for water environment cleaning based on machine intelligence according to an embodiment of the present invention includes:

[0073] Preliminary planning module: collect images of the bottom of the ship to obtain grid area images, divide the grid area images into several cleaning areas, process the attachment data in each cleaning area to obtain the attachment degree value, and complete the preliminary planning of the cleaning path according to the attachment degree value;

[0074] It should be noted that: the grid area image is divided in a manner of: dividing the grid area image in a manner of equal area to obtain a cleaning area;

[0075] In some embodiments, the attachment data in each cleaning area is collected to obtain the attachment data, wherein the attachment data includes the attachment area and the coordinates of the attachment, and the attachment distribution value and the attachment area ratio are obtained respectively. The attachment distribution value and the attachment area ratio are processed to obtain the attachment degree value. The specific process is as follows:

[0076] Obtain the area of ​​the attachments in the cleaning area, add up the areas of the attachments in the cleaning area, and obtain the total area of ​​the attachments;

[0077] Calculate the ratio of the total area of ​​the attachments to the area of ​​the cleaning area to obtain the attachment area ratio;

[0078] The intersection of the two intersecting boundary lines at the edge of the cleaning area is taken as the origin of the two-dimensional coordinate system. A straight line in the horizontal direction is drawn with the origin as the starting point as the X axis of the two-dimensional coordinate system. A straight line in the vertical direction is drawn with the origin as the starting point as the Y axis of the two-dimensional coordinate system to obtain the cleaning area coordinate system.

[0079] In the cleaning area coordinate system, obtain the midpoint coordinates in the cleaning area, and mark the midpoint coordinates in the cleaning area as the cleaning midpoint coordinates ( , ), where z represents the midpoint;

[0080] Get the coordinates of each attachment in the cleaning area coordinate system ( , ), where n represents the number of all attachments in the cleaning area coordinate system;

[0081] Get the coordinates of each attachment in the cleaning area coordinate system ( , ) and the cleaning midpoint coordinates ( , ), marked as the attachment distance value;

[0082] For example, the coordinates of each attachment in the cleaning area coordinate system ( , ) and the cleaning midpoint coordinates ( , ) into the formula: , calculate and obtain the floating distance value L;

[0083] All floating object distance values ​​L are added and averaged to obtain the floating distance mean, the perimeter of the cleaning area is measured, and the floating distance mean is compared with the perimeter of the cleaning area to obtain the attachment distribution value;

[0084] The attachment area ratio is calculated with the attachment distribution value to obtain the attachment degree value;

[0085] It can be understood that the meaning of the attachment degree value is: reflecting the proportion and distribution of attachments in each cleaning area, and comprehensively evaluating the area proportion and aggregation degree of attachments in the cleaning area. Specifically, the larger the value, the larger the proportion of attachments in the cleaning area, the more dispersed, and the greater the regional pollution degree. On the contrary, the smaller the value, the smaller the proportion of attachments in the cleaning area, the more concentrated, and the less regional pollution degree.

[0086] Based on the attachment degree value, the attachment degree value corresponding to each key cleaning area is obtained, the attachment degree values ​​corresponding to each key cleaning area are compared, and they are sorted and cleaned in order from large to small until the bottom attachments in the cleaning area are cleaned and a preliminary cleaning path is obtained;

[0087] Exemplarily, the cleaning area corresponding to the maximum attachment degree is selected as the first cleaning area, and the cleaning area corresponding to the minimum attachment degree is selected as the nth cleaning area;

[0088] The specific implementation scheme of the embodiment of the present invention is as follows: performing image acquisition on the bottom of the ship to obtain a grid area image, dividing the grid area image into a number of cleaning areas, processing the attachment data in each cleaning area, the attachment data including an attachment distribution value and an attachment area ratio, processing the attachment distribution value and the attachment area ratio to obtain an attachment degree value, and completing preliminary planning of the cleaning path according to the attachment degree value, thereby comprehensively evaluating the area proportion and aggregation degree of the attachments in the cleaning area by analyzing the proportion and distribution of the attachments in each cleaning area, and completing preliminary planning of the cleaning path according to the size of the attachment degree value corresponding to each cleaning area.

[0089] Example 2

[0090] like Figure 1-2 As shown, based on Example 1, an autonomous decision-making support system for water environment cleaning based on machine intelligence according to an embodiment of the present invention includes:

[0091] Evaluation planning module: Based on the preliminary cleaning path, obtain the non-recent cleaning quantity ratio and the non-recent cleaning degree value, process the non-recent cleaning quantity ratio and the non-recent cleaning degree value, output the planning analysis value, and compare it with the threshold to generate a path analysis signal;

[0092] Wherein, the path analysis signal includes a path analysis signal or a path non-analysis signal;

[0093] In some embodiments, when the preliminary cleaning path is obtained, all cleaning regions are combined to obtain multiple cleaning region groups;

[0094] Exemplarily, the first cleaning area is combined with the second cleaning area, the second cleaning area is combined with the third cleaning area...the n-1th cleaning area is combined with the nth cleaning area;

[0095] Based on any cleaning area group;

[0096] In the cleaning area group, obtain the distance between two cleaning areas, marked as cleaning distance;

[0097] In the cleaning area group, the cleaning area with the highest ranking is marked as the front cleaning area, and the distances between the front cleaning area and all other cleaning areas are obtained and marked as other cleaning distances;

[0098] It should be noted that all other cleaning areas do not include cleaning areas that are in the same cleaning area group as the front cleaning area;

[0099] If all other cleaning distances are greater than the cleaning distance, the cleaning area group is marked as the nearest cleaning area group;

[0100] If all other cleaning distances are not greater than the cleaning distance, the cleaning area group is marked as a non-closest cleaning area group;

[0101] The number of non-recent cleaning area groups is counted and the ratio is processed with the total number of cleaning area groups to obtain the non-recent cleaning number ratio, which is marked as SL;

[0102] Based on any non-recently cleaned area group;

[0103] Obtain the closest cleaning area adjacent to the front cleaning area in the non-closest cleaning area group, and measure the distance between the front cleaning area and the adjacent closest cleaning area, marking it as the closest cleaning distance;

[0104] Perform difference processing on the cleaning distance of the non-closest cleaning area group and the closest cleaning distance, and take the absolute value of the difference to obtain the cleaning distance difference of the non-closest cleaning area group, and sum and average the cleaning distance differences of all non-closest cleaning area groups to obtain the mean cleaning distance difference;

[0105] Get the cleaning distances of all cleaning area groups and sum them up to get the total cleaning distance;

[0106] The average cleaning distance difference is processed by ratio processing with the total cleaning distance to obtain the non-closest cleaning degree value, which is marked as CD;

[0107] Substitute the non-recent cleaning quantity ratio and non-recent cleaning degree values ​​into the formula: , calculate the planning analysis value H, where It is expressed as the ratio of the number of non-recent cleanings, and L is expressed as the non-recent cleaning degree value. , Expressed as a preset scale factor, The value can be 0.4662. The value can be 0.5338;

[0108] It can be understood that the meaning of the planning analysis value is to reflect the rationality of the distance performance of the cleaning area sorting in the preliminary cleaning path. Specifically, the larger the value is, the more unreasonable the sorting of the cleaning areas in the preliminary cleaning path is in terms of distance performance. On the contrary, the smaller the value is, the more reasonable the sorting of the cleaning areas in the preliminary cleaning path is in terms of distance performance.

[0109] The planning analysis value is compared with the planning analysis threshold. The comparison process is as follows:

[0110] If the planning analysis value is greater than or equal to the planning analysis threshold, it means that the order of the cleaning areas in the preliminary cleaning path in terms of distance performance is relatively unreasonable, and a path analysis signal is generated;

[0111] If the planning analysis value is less than the planning analysis threshold, it means that the order of the cleaning areas in the preliminary cleaning path in terms of distance performance is reasonable, and the path is generated without analyzing the signal;

[0112] If the signal is not analyzed based on the path, the preliminary cleaning path is used as the final cleaning path;

[0113] The specific implementation scheme of the embodiment of the present invention is: based on the preliminary cleaning path, the non-recent cleaning quantity ratio and the non-recent cleaning degree value are obtained, the non-recent cleaning quantity ratio and the non-recent cleaning degree value are substituted into the formula, and the planning analysis value is output. If the planning analysis value is greater than or equal to the planning analysis threshold, it means that the order of the cleaning areas in the preliminary cleaning path in terms of distance performance is relatively unreasonable, and thus by analyzing the non-recent cleaning quantity ratio and the non-recent cleaning degree value, the rationality of the order of the cleaning areas in terms of distance performance in the preliminary cleaning path is reflected.

[0114] Example 3

[0115] like Figure 1-2 As shown, based on Example 1 and Example 2, an autonomous decision-making support system for water environment cleaning based on machine intelligence according to an embodiment of the present invention includes:

[0116] Path analysis module: based on the path analysis signal, re-plan the cleaning path to obtain a reference cleaning path, perform deviation analysis on the reference cleaning path and the preliminary cleaning path to obtain a path deviation performance value, and determine the final cleaning path based on the path deviation performance value;

[0117] Specifically, the first cleaning area is combined with other cleaning areas respectively to obtain a cleaning area unit group;

[0118] Exemplarily, the first cleaning area is combined with the second cleaning area, the first cleaning area is combined with the third cleaning area, ... the first cleaning area is combined with the nth cleaning area;

[0119] Based on any cleaning area unit group;

[0120] In the cleaning area unit group, the adhesion degree values ​​corresponding to the two cleaning areas are subjected to difference processing, and the absolute value of the difference is taken to obtain the adhesion difference corresponding to the cleaning area unit group;

[0121] The front cleaning area in the cleaning area unit group is marked as the front cleaning area, and the adhesion difference corresponding to the cleaning area unit group is processed with the adhesion degree value of the front cleaning area to obtain the adhesion proximity value;

[0122] In the cleaning area unit group, the distance between two cleaning areas is obtained to obtain the cleaning distance of the cleaning area unit group, the cleaning distances of all cleaning area unit groups are obtained and summed to obtain the total cleaning distance, and the cleaning distance of the cleaning area unit group is processed by ratio with the total cleaning distance to obtain the cleaning distance ratio;

[0123] The attachment proximity value and the cleaning distance ratio are summed to obtain the priority value of the cleaning area unit group;

[0124] Compare the priority values ​​of all cleaning area unit groups, and select the cleaning area unit group corresponding to the minimum priority value as the target unit group;

[0125] Two cleaning areas in the target unit group are regarded as adjacent cleaning areas, wherein the positions of the adjacent cleaning areas in the reference cleaning path are adjacent;

[0126] When the cleaning area adjacent to the first cleaning area is determined, the cleaning area adjacent to the first cleaning area is used again as the second cleaning area;

[0127] Processing the second cleaning area, wherein the processing method of the second cleaning area is the same as that of the first cleaning area, and determining the cleaning area adjacent to the second cleaning area until all cleaning areas are determined and completed, and finally obtaining a reference cleaning path;

[0128] The deviation analysis between the reference cleaning path and the preliminary cleaning path is as follows:

[0129] In the reference cleaning path and the preliminary cleaning path, the cleaning areas with the same ranking are integrated into cleaning groups with the same ranking;

[0130] Exemplarily, the cleaning region ranked first in the reference cleaning path and the cleaning region ranked first in the preliminary cleaning path are integrated into a cleaning group with the same ranking;

[0131] In the same ranking cleaning group, if the positions of the two cleaning areas in the grid area image are different (i.e., they are not the same area), the same ranking cleaning group is marked as a non-overlapping group; if the positions of the two cleaning areas in the grid area image are the same (i.e., they are the same area), the same ranking cleaning group is marked as an overlapping group;

[0132] Count the number of non-overlapping groups and compare them with the number of cleaned groups with the same ranking to obtain the path deviation performance value;

[0133] In some embodiments, the path deviation performance value is compared to a path deviation performance threshold;

[0134] If the path deviation performance value is greater than the path deviation performance threshold, it means that the deviation between the reference cleaning path and the preliminary cleaning path is large, and the reference cleaning path is used as the final cleaning path;

[0135] If the path deviation performance value is less than or equal to the path deviation performance threshold, it means that the deviation between the reference cleaning path and the preliminary cleaning path is small, and the reference cleaning path or the preliminary cleaning path is used as the final cleaning path.

[0136] The specific implementation scheme of the embodiment of the present invention is: based on the unreasonable path signal, the preliminary cleaning path is optimized to obtain the optimized cleaning path, the optimized reasonable value is obtained, the path deviation value corresponding to the optimized cleaning path is compared with the path deviation threshold, and the final cleaning path is determined, thereby determining the final cleaning path by comparing the size of the optimized reasonable value and the path deviation value, thereby realizing decision assistance in determining the final path.

[0137] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A water environment cleaning autonomous decision-making support system based on machine intelligence, characterized by: include: Preliminary planning module: collect images of the bottom of the ship to obtain grid area images, divide the grid area images into several cleaning areas, process the attachment data in each cleaning area, where the attachment data includes the attachment area and the coordinates of the attachment, obtain the attachment distribution value and the attachment area ratio respectively, process the attachment distribution value and the attachment area ratio to obtain the attachment degree value, and complete the preliminary planning of the cleaning path according to the attachment degree value; Evaluation planning module: Based on the preliminary cleaning path, obtain the non-recent cleaning quantity ratio and the non-recent cleaning degree value, process the non-recent cleaning quantity ratio and the non-recent cleaning degree value, output the planning analysis value, and compare it with the threshold to generate a path analysis signal; Wherein, the path analysis signal includes a path analysis signal or a path non-analysis signal; The planning analysis value is obtained as follows: Substitute the non-recent cleaning quantity ratio and non-recent cleaning degree values ​​into the formula: , calculate the planning analysis value H, where It is expressed as the ratio of the number of non-recent cleanings, and L is expressed as the non-recent cleaning degree value. , Expressed as a preset scaling factor, The value is 0.4662. The value is 0.5338; The non-recent cleaning quantity ratio is obtained as follows: When the preliminary cleaning path is obtained, all cleaning areas are combined to obtain multiple cleaning area groups; Based on any cleaning area group; In the cleaning area group, obtain the distance between two cleaning areas, marked as cleaning distance; In the cleaning area group, the cleaning area with the highest ranking is marked as the front cleaning area, and the distances between the front cleaning area and all other cleaning areas are obtained and marked as other cleaning distances; If all other cleaning distances are greater than the cleaning distance, the cleaning area group is marked as the nearest cleaning area group; If all other cleaning distances are not greater than the cleaning distance, the cleaning area group is marked as a non-closest cleaning area group; The number of non-recent cleaning area groups is counted and the ratio is processed with the total number of cleaning area groups to obtain the non-recent cleaning number ratio, which is marked as SL; The method for obtaining the non-recent cleaning degree value is as follows: Based on any non-recently cleaned area group; Obtain the closest cleaning area adjacent to the front cleaning area in the non-closest cleaning area group, and measure the distance between the front cleaning area and the adjacent closest cleaning area, marking it as the closest cleaning distance; Perform difference processing on the cleaning distance of the non-closest cleaning area group and the closest cleaning distance, and take the absolute value of the difference to obtain the cleaning distance difference of the non-closest cleaning area group, and sum and average the cleaning distance differences of all non-closest cleaning area groups to obtain the mean cleaning distance difference; Get the cleaning distances of all cleaning area groups and sum them up to get the total cleaning distance; The average cleaning distance difference is processed by ratio processing with the total cleaning distance to obtain the non-closest cleaning degree value, which is marked as CD; Path analysis module: based on the path analysis signal, re-plan the cleaning path to obtain a reference cleaning path, perform deviation analysis on the reference cleaning path and the preliminary cleaning path to obtain a path deviation performance value, and determine the final cleaning path based on the path deviation performance value; The reference cleaning path is obtained as follows: Combining the first cleaning area with other cleaning areas respectively to obtain a cleaning area unit group; Based on any cleaning area unit group; In the cleaning area unit group, the adhesion degree values ​​corresponding to the two cleaning areas are subjected to difference processing, and the absolute value of the difference is taken to obtain the adhesion difference corresponding to the cleaning area unit group; The front cleaning area in the cleaning area unit group is marked as the front cleaning area, and the adhesion difference corresponding to the cleaning area unit group is processed with the adhesion degree value of the front cleaning area to obtain the adhesion proximity value; In the cleaning area unit group, the distance between two cleaning areas is obtained to obtain the cleaning distance of the cleaning area unit group, the cleaning distances of all cleaning area unit groups are obtained and summed to obtain the total cleaning distance, and the cleaning distance of the cleaning area unit group is processed by ratio with the total cleaning distance to obtain the cleaning distance ratio; The attachment proximity value and the cleaning distance ratio are summed to obtain the priority value of the cleaning area unit group; Compare the priority values ​​of all cleaning area unit groups, and select the cleaning area unit group corresponding to the minimum priority value as the target unit group; Two cleaning areas in the target unit group are regarded as adjacent cleaning areas, wherein the positions of the adjacent cleaning areas in the reference cleaning path are adjacent; When the cleaning area adjacent to the first cleaning area is determined, the cleaning area adjacent to the first cleaning area is used again as the second cleaning area; Processing the second cleaning area, wherein the processing method of the second cleaning area is the same as that of the first cleaning area, and determining the cleaning area adjacent to the second cleaning area until all cleaning areas are determined and completed, and finally obtaining a reference cleaning path; The specific process of performing deviation analysis between the reference cleaning path and the preliminary cleaning path and obtaining the optimized cleaning path is as follows: In the reference cleaning path and the preliminary cleaning path, the cleaning areas with the same ranking are integrated into cleaning groups with the same ranking; In the same ranking cleaning group, if the positions of the two cleaning areas in the grid area image are different, the same ranking cleaning group is marked as a non-overlapping group; if the positions of the two cleaning areas in the grid area image are the same, the same ranking cleaning group is marked as an overlapping group; Count the number of non-overlapping groups and compare them with the number of cleaned groups with the same ranking to obtain the path deviation performance value; comparing the path deviation performance value to a path deviation performance threshold; If the path deviation performance value is greater than the path deviation performance threshold, the reference cleaning path is used as the final cleaning path; If the path deviation performance value is less than or equal to the path deviation performance threshold, the reference cleaning path or the preliminary cleaning path is used as the final cleaning path.

2. According to claim 1, a water environment cleaning autonomous decision-making support system based on machine intelligence is characterized by: The attachment area ratio is obtained as follows: Obtain the area of ​​the attachments in the cleaning area, add up the areas of the attachments in the cleaning area, and obtain the total area of ​​the attachments; The ratio of the total area of ​​the attachments to the area of ​​the cleaning region is calculated to obtain the attachment area ratio.

3. The water environment cleaning autonomous decision-making support system based on machine intelligence according to claim 1 is characterized by: The attachment distribution value is obtained as follows: The intersection of the two intersecting boundary lines at the edge of the cleaning area is taken as the origin of the two-dimensional coordinate system. A straight line in the horizontal direction is drawn with the origin as the starting point as the X axis of the two-dimensional coordinate system. A straight line in the vertical direction is drawn with the origin as the starting point as the Y axis of the two-dimensional coordinate system to obtain the cleaning area coordinate system. In the cleaning area coordinate system, obtain the midpoint coordinates in the cleaning area, and mark the midpoint coordinates in the cleaning area as the cleaning midpoint coordinates ( , ), where z represents the midpoint; Get the coordinates of each attachment in the cleaning area coordinate system ( , ), where n represents the number of all attachments in the cleaning area coordinate system; Get the coordinates of each attachment in the cleaning area coordinate system ( , ) and the cleaning midpoint coordinates ( , ), marked as the attachment distance value; The coordinates of each attachment in the cleaning area coordinate system ( , ) and the cleaning midpoint coordinates ( , ) into the formula: , calculate the floating object distance value L; All floating object distance values ​​L are added and averaged to obtain the mean floating distance, the perimeter of the cleaning area is measured, and the mean floating distance is ratioed with the perimeter of the cleaning area to obtain the attachment distribution value.

4. The water environment cleaning autonomous decision-making support system based on machine intelligence according to claim 1 is characterized by: The initial cleaning path is obtained as follows: The attachment area ratio is calculated with the attachment distribution value to obtain the attachment degree value; Based on the adhesion degree value, the adhesion degree value corresponding to each key cleaning area is obtained, the adhesion degree value corresponding to each key cleaning area is compared, and the cleaning is carried out in order from large to small until the bottom attachments in the cleaning area are cleaned and a preliminary cleaning path is obtained.

5. The water environment cleaning autonomous decision-making support system based on machine intelligence according to claim 1 is characterized by: The planning analysis value is compared with the planning analysis threshold. The comparison process is as follows: If the planning analysis value is greater than or equal to the planning analysis threshold, a path analysis signal is generated; If the planning analysis value is less than the planning analysis threshold, a path is generated without analyzing the signal; If the signal is not analyzed based on the path, the preliminary cleaning path is used as the final cleaning path.

Images