CN117475526A - Method and system for inspecting plunge pool - Google Patents

Abstract

The invention discloses a method and a system for inspecting a plunge pool, which relate to the field of plunge pool robot inspection and comprise the following steps: firstly, establishing communication connection with an underwater robot, then acquiring map information of a target plunge pool, then determining a target inspection line, operating the underwater robot for inspection, and finally receiving an inspection result and analyzing the inspection result. By means of the design, the specific inspection depth is located through the historical plunge pool level diagram, inspection is conducted on a plurality of depths, and overall monitoring and management of the plunge pool are effectively achieved.

Description

Method and system for inspecting plunge pool

Technical Field

The invention relates to the field of pool robot inspection, in particular to a pool inspection method and system.

Background

The traditional plunge pool inspection method often needs manual work, and is low in efficiency and easy to cause safety accidents. Especially for deep inspection of plunge pool, due to lack of effective equipment and methods, it is often difficult to achieve accuracy and comprehensiveness.

In addition, conventional methods often fail to perform detailed inspection for every depth within the pond, resulting in some potential problems that may be ignored. Therefore, it is important to develop a novel pool inspection method capable of performing comprehensive inspection at a plurality of depths.

Disclosure of Invention

The invention aims to provide a method and a system for pool inspection.

In a first aspect, an embodiment of the present invention provides a method for inspecting a pool, where the method includes:

responding to a patrol task starting instruction aiming at a target plunge pool, and establishing communication connection with the underwater robot;

acquiring pool map information corresponding to a target pool, wherein the pool map information comprises historical pool level diagrams of a plurality of depths;

determining a target inspection line and operating the underwater robot to inspect according to a historical plunge pool level diagram corresponding to the target depth, wherein the target depth is any one depth of a plurality of depths;

receiving inspection results of the underwater robot for a plurality of depths;

and analyzing the inspection result to obtain an inspection analysis report of the inspection result.

In a second aspect, an embodiment of the present invention provides a server system, including a server, configured to perform the method of the first aspect.

Compared with the prior art, the invention has the following beneficial effects:

the embodiment of the invention provides a method and a system for inspecting a plunge pool, which are characterized in that communication connection is established between the plunge pool and an underwater robot, map information of a target plunge pool is acquired, a target inspection line is determined, the underwater robot is operated to inspect, and finally an inspection result is received and analyzed. By means of the design, the specific inspection depth is located through the historical plunge pool level diagram, inspection is conducted on a plurality of depths, and overall monitoring and management of the plunge pool are effectively achieved.

Drawings

In order to more clearly illustrate the technical solution of the embodiments of the present invention, the drawings that are required to be used in the embodiments will be briefly described. It is appreciated that the following drawings depict only certain embodiments of the invention and are therefore not to be considered limiting of its scope. Other relevant drawings may be made by those of ordinary skill in the art without undue burden from these drawings.

FIG. 1 is a schematic flow chart of steps of a pool inspection method according to an embodiment of the present invention;

fig. 2 is a schematic block diagram of a computer device according to an embodiment of the present invention.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the technical solutions in the embodiments of the present application will be described in more detail below with reference to the accompanying drawings in the preferred embodiments of the present application. In the drawings, the same or similar reference numerals refer to the same or similar components or components having the same or similar functions throughout. The described embodiments are some, but not all, of the embodiments of the present application. The embodiments described below by referring to the drawings are exemplary and intended for the purpose of explaining the present application and are not to be construed as limiting the present application. All other embodiments, which can be made by one of ordinary skill in the art based on the embodiments herein without making any inventive effort, are intended to be within the scope of the present application.

Embodiments of the present application are described in detail below with reference to the accompanying drawings.

In order to solve the technical problems in the prior art, fig. 1 is a schematic flow chart of a method for inspecting a plunge pool according to an embodiment of the present disclosure, and the method for inspecting a plunge pool is described in detail below.

Step S201, a communication connection with an underwater robot is established in response to a patrol task starting instruction for a target plunge pool;

step S202, obtaining pool map information corresponding to a target pool, wherein the pool map information comprises historical pool level diagrams of a plurality of depths;

step S203, determining a target inspection line and operating the underwater robot to inspect according to a historical plunge pool level diagram corresponding to the target depth, wherein the target depth is any one of a plurality of depths;

step S204, receiving inspection results of the underwater robot for a plurality of depths;

step S205, analyzing the inspection result to obtain an inspection analysis report of the inspection result.

In the embodiment of the invention, a pool manager designates a target pool to be inspected in an inspection plan, and sends an inspection task starting instruction through a control panel or a mobile application program. After receiving the instruction, the underwater robot establishes stable communication connection with the control center by using the wireless communication equipment. This may involve using bluetooth, wi-Fi, or proprietary communication protocols, among other techniques. Underwater robots are equipped with various sensors and scanning devices such as sonar, lidar, cameras, and the like. Before the start of the inspection, the underwater robot will activate these sensors and devices to scan and measure the plunge pool. The scanned data comprise the geometric structure of the underwater environment, the distribution of obstacles, the water quality condition and the like. These data are processed and integrated to form the pool map information, including histories at different depths, providing the pool hierarchy and trends. Based on the target depth, the underwater robot plans a patrol route using data of the historical pool level diagram. For example, if the target depth is 10 meters, the underwater robot will determine a patrol path to cover the area of this depth range based on the historical hierarchy map corresponding to the depth. The inspection path may traverse known underwater attractions, critical areas, or potentially problematic areas, ensuring that the target depth is fully and effectively inspected. The underwater robot starts to execute tasks according to the planned inspection line, and simultaneously monitors the plunge pool in real time by using the sensor and the equipment. In the inspection process, the underwater robot records and transmits inspection result data to the control center. The inspection results may include measured water quality parameters, detected obstacle locations, image or video recordings, and the like. For each depth, the robot sends the corresponding inspection results back for further analysis and processing. After the inspection result is received, the control center or the operator can deeply analyze and read the data. The analysis process may include the following aspects: and (3) water quality analysis: and (3) evaluating the water quality condition of the plunge pool according to measurement data collected by the underwater robot, such as pH value, dissolved oxygen, turbidity and the like. These data may be compared to preset criteria or historical data to determine if there is a water quality anomaly or a pollution problem. Obstacle detection: based on sensor data acquired by the underwater robot, obstructions in the plunge pool, such as aquatic weeds, floats, or other potential objects, are identified and analyzed. This helps to assess overall cleanliness and navigational safety of the plunge pool. Topography analysis: and generating a submarine topography map by using the topography data recorded by the underwater robot, and analyzing the fluctuation, gradient and unevenness of the bottom of the plunge pool. This can be used to evaluate the topographical features of the plunge pool and detect potential sediment or soil erosion conditions. The inspection analysis report will include the assessment results for each aspect, indications of abnormal conditions, and suggested improvements. The report may contain charts, data visualizations, and text descriptions so that the pool manager can fully understand the inspection results and make corresponding decisions. In addition, the report may also record historical inspection data, trend analysis, and long term monitoring recommendations to support maintenance and management work for the plunge pool.

In the embodiment of the present invention, the aforementioned step S203 may be implemented by the following detailed implementation.

(1) Obtaining a pool level diagram to be inspected based on iteration of the historical pool level diagram, wherein the pool level diagram to be inspected is an iterated historical pool level diagram, and inspection lines of the historical pool level diagram are historical inspection lines;

(2) In the process of controlling the underwater robot to patrol on the hierarchical diagram of the plunge pool to be patrolled according to the historical patrol line, identifying the area which can not pass, and obtaining a target area which can not pass;

(3) Determining a pending non-passable range according to the detour originating location;

(4) Determining the regional unit bypass cost of each current regional unit in the undetermined non-passable range according to at least one of a first spatial span and a second spatial span, wherein the first spatial span is the spatial span between the current regional unit and a bypass starting position, and the second spatial span is the spatial span between the current regional unit and a target non-passable region;

(5) Acquiring a current area unit group on each candidate patrol strategy in a plurality of candidate patrol strategies in a pending non-passable range;

(6) Determining an area unit detour cost array corresponding to the current area unit group according to the area unit detour cost;

(7) Carrying out data normalization on the regional unit detour cost array to obtain detour cost corresponding to each candidate routing strategy;

(8) Selecting at least one detour strategy from a plurality of candidate patrol strategies according to the detour cost;

(9) Determining a current candidate range in the undetermined non-passable range according to at least one detour strategy;

(10) Traversing a passable area unit group adjacent to a current originating location area unit in the current candidate range, wherein the current originating location area unit is a current area unit corresponding to a bypass originating location;

(11) Acquiring the routing inspection cost corresponding to each current area unit in the passable area unit group, and determining the current area unit which is traversed latest and corresponds to the acquired routing inspection cost as a previous-level area unit, wherein the routing inspection cost is the cost from the current initial position area unit to the current leaving position area unit through the corresponding current area unit;

(12) Traversing a next passable area unit group adjacent to the target current area unit, stopping traversing until the traversed passable area unit group comprises a current departure position area unit, and backtracking a previous level area unit from the current departure position area unit to a current starting position area unit to obtain a target bypass line between a bypass starting position and a bypass departure position, wherein the target current area unit is a current area unit which is not traversed and has optimal routing inspection cost, the target bypass line is used for sending a bypass target non-passable area from the bypass starting position to reach a bypass departure position, the bypass starting position is a bypass starting position on a historical routing inspection line, and the bypass departure position is a bypass ending position on the historical inspection line;

(13) According to the target detour line, adjusting the historical inspection line to obtain a target inspection line;

(14) According to the target inspection line, the underwater robot is controlled to carry out inspection on the hierarchical diagram of the plunge pool to be inspected.

In the embodiment of the invention, it is assumed that in a plunge pool, an iterative process is performed using a historical plunge pool hierarchy chart as an initial chart. By analyzing the information such as the shape, depth, barriers and the like of the current plunge pool, the plunge pool level diagram to be patrolled and examined is updated, so that the patrolling path can be planned more accurately. When an underwater robot is patrolled in a plunge pool, it will use sensors to detect the underwater environment. During the detection process, if the robot detects rocks, branches or other obstructions, these detected areas will be marked as target non-passable areas. The detour origination location may be a robot origination location or a starting point for a particular task requirement. A boundary is disposed around the target non-passable region, the boundary forming a pending non-passable range. The spatial span may be defined in terms of physical characteristics in the pond, such as the distance or path length between two area units. Other factors such as water flow rate, energy consumption, etc. may also be considered in calculating the area unit bypass costs. The candidate inspection strategy may be a predefined set of rules for selecting the next movement of the robot during inspection. The current regional unit group is a set of regional units which are adjacent to the current position of the robot and can pass through and are determined by a candidate inspection strategy. For the region units in each current region unit group, its detour costs are calculated and these costs are stored in an array. The detour cost array reflects the detour cost of each regional unit and helps to evaluate the merits of different regional unit sets. Based on the regional unit detour cost array, the regional unit group with the lowest detour cost is selected as one of the detour strategies. This ensures that the robot selects the most economical and efficient path during the inspection process.

Suppose that bypass policy a is selected as the current bypass policy. The detour strategy a requires the robot to select a specific regional unit group adjacent to the current location and accessible. Based on this strategy, a current candidate range meeting the conditions within the pending non-passable range is determined. Within the current candidate range, there are multiple adjacent passable zone cell groups. The underwater robot traverses the adjacent regional unit groups in turn to evaluate the inspection cost of each regional unit group and continues the next exploration. For the current zone unit in each adjacent passable zone unit group, the underwater robot will acquire the inspection cost to reach that zone unit. The patrol cost is derived by calculating the time, energy consumption or other metrics required to reach the current area unit from the current originating location area unit. After the inspection cost is obtained, the robot updates the current area unit into the upper-level area unit and continues to trace back the next step.

In the traversal process, when the current departure position area unit is included in the passable area unit group, the robot terminates the traversal operation. At this time, the robot starts the backtracking operation of the previous stage area unit from the current departure position area unit to the current starting position area unit to form a complete detour line. By way of example, it is assumed that there is a historical inspection line consisting of the following area units: a→b→c→d. At the time of inspection, the robot detects that there is an obstacle between the area units B and C, which is marked as a target non-passable area. And selecting A as a detour starting position and D as a detour leaving position according to the historical patrol line. And selecting a regional unit group which is adjacent to the current position and can pass through according to the detour strategy. Assuming that the current robot is located in region unit B, the current candidate range includes neighboring region units A, C and D. The robot traverses the adjacent passable zone unit group from zone unit B: A. c and D. The robot calculates the patrol cost from the origin position (a) to the current area unit (A, C or D) and selects the optimal current area unit. Suppose the patrol cost is as follows: a is 10 minutes; c, 15 minutes; d, 12 minutes. According to the lowest inspection cost, the robot selects the area unit A as the target current area unit. Since the target current region unit (a) already contains the detour exit position (D) during the traversal process, the traversal operation is terminated. Then, the robot performs backtracking of the upper level area unit from the detour leaving position (D) to the detour originating position (a). Through backtracking operation, the robot gets the target detour line: the bypass starting position is A, the bypass leaving position is D, and the target bypass line is A, B, C and D. The robot forms a target detour line between a detour originating location and a detour exiting location by traversing adjacent groups of navigable area units and selecting a current area unit for which the cost of the patrol is lowest. The detour line can help the robot avoid the target non-passable area and perform tasks according to the requirements of the historical patrol line.

And inserting the target detour line into the historical inspection line based on the detour line and the backtracking process calculated by the robot. This results in a new target patrol route that combines historical data with an optimized detour path. And finally, according to the target inspection line, the robot executes a specific inspection task on the hierarchical diagram of the plunge pool to be inspected. The robot moves according to the sequence indicated by the target inspection line, performs a corresponding inspection operation on each target area unit, collects data or performs a specific task.

In the embodiment of the present invention, when the at least one detour policy is a plurality of detour policies, the step of determining the current candidate range in the pending non-passable range according to the at least one detour policy may be implemented by the following example.

Acquiring a plurality of local candidate ranges corresponding to a plurality of detour strategies, and combining the plurality of local candidate ranges to form a current candidate range in the undetermined non-passable range; or determining a target detour strategy from a plurality of detour strategies according to the current departure position area unit, and determining a candidate range on the target detour strategy as a current candidate range in the undetermined non-passable range, wherein the current departure position area unit is a current area unit corresponding to the detour departure position.

In the embodiment of the invention, various bypass strategies can be selected in the inspection task. These policies may vary depending on the particular needs and scenario. For example: shortest path strategy: and selecting the passable area closest to the current position as a next target. Optimal cost strategy: and selecting the passable area with the lowest cost as a next target according to factors such as inspection time or energy consumption. Random strategy: a passable area is randomly selected as the next target. The current candidate range is determined according to a detour strategy, wherein the current candidate range comprises a local candidate range and a pending unvented range.

a) Obtaining local candidate ranges corresponding to a plurality of bypass strategies: two bypass strategies, a shortest path strategy and an optimal cost strategy are assumed. For the shortest path strategy, the local candidate ranges are region units A, C and D (neighboring region units). For the optimal cost strategy, the local candidate ranges are region units B, C and D. The two local candidate ranges are combined to form the current candidate range in the pending non-passable range, A, B, C and D.

b) Determining a target detour strategy according to the current departure location area unit: let the recorded historical patrol route be a→b→c→d, where D is the current departure location area unit. From the historical patrol route, the target detour strategy can be determined as the shortest path strategy, since to go back from D to a, the path with the shortest distance needs to be selected. Then, a candidate range (A, B, C) on the target detour policy is determined as the current candidate range in the pending non-passable range.

If multiple bypass policies are available, each policy may be used to obtain its corresponding local candidate range and combine the local candidate ranges together to form the current candidate range. In addition, the target detour strategy may be selected from a plurality of detour strategies according to the current departure location area unit, and the candidate range thereof may be used as the current candidate range. Therefore, the robot can conduct next path planning and decision according to the current candidate range so as to bypass the target non-passable area and complete the inspection task.

In an embodiment of the present invention, before performing the foregoing step of obtaining the pool level diagram to be patrolled and examined based on the iteration of the historical pool level diagram, the following implementation manner may be further included.

(1) Performing area unit segmentation on the historical plunge pool hierarchical graph to obtain an area unit set, wherein the area unit set comprises a plurality of historical area units;

(2) In the area unit set, taking a historical initial position area unit as a starting point, and determining a plurality of sensing paths corresponding to a plurality of preset deflection amounts according to preset sensing intervals;

(3) Extracting a reliable sensing path array from a plurality of sensing paths according to preset rules, wherein the preset rules are determined according to at least one of sensing distance and preset commands of a target underwater robot;

(4) Acquiring a preset number of reliable sensing paths with preset deflection quantity sequence adjacent to each other from the reliable sensing path array, and acquiring sensing path ranges corresponding to the preset number of reliable sensing paths;

(5) Combining the history area units in the sensing path range which is larger than the preset range to form a first history area unit group connected with the history starting position area unit;

(6) When the first history area unit group does not include the history departure position area unit, the following processing is continuously performed:

performing distance sensing of a target sensing interval by taking a current history area unit group as a starting point to obtain a next history area unit group connected with the current history area unit group, wherein the positive feedback relation exists between the target sensing interval and the connection coincidence rate of the distance sensing;

(7) Ending the distance sensing when the next history area unit group comprises the history leaving position area unit, and combining the history starting position area unit and all the history area unit groups to form a connected preset number of history area units;

(8) Determining a preset number of history area units as history candidate ranges;

(9) And carrying out bypass line layout on the historical starting position area unit and the historical leaving position area unit according to the historical candidate range to obtain a historical inspection line, wherein the historical starting position area unit is a historical area unit corresponding to the inspection starting position of the historical pond hierarchical diagram, the historical leaving position area unit is a historical area unit corresponding to the inspection leaving position of the historical pond hierarchical diagram, and the preset number of historical area units at least comprises the historical starting position area unit and the historical leaving position area unit.

In an embodiment of the present invention, the preparation before the iteration of the historic pond-level graph may include:

a) Region unit segmentation: and carrying out regional unit segmentation on the historical plunge pool hierarchical graph to obtain a group of regional unit sets. For example, the pond is divided into regular grids, each grid cell representing an area cell.

b) Presetting a sensing path: and in the area unit set, taking the area unit at the original historical position as a starting point, and determining a plurality of sensing paths corresponding to the preset deflection according to the preset sensing interval. These sense paths may be used to detect information of different area units. For example, from the start point region unit, the region units of the next sensing path are sequentially selected along the preset deflection amount at the set sensing interval.

c) Extracting a reliable sense path array: and extracting a reliable sensing path array from the plurality of sensing paths according to a preset rule. The preset rules may be determined based on factors such as a sensing distance or a preset command to the target underwater robot. For example, a path with good communication quality with the target robot in the sensing path array is selected according to the set minimum sensing distance and the communication capability requirement with the target robot.

d) Acquiring a sensing path range: and acquiring a preset number of reliable sensing paths with preset deflection quantity sequence adjacent to each other from the reliable sensing path array, and acquiring sensing path ranges corresponding to the sensing paths. For example, if the preset number is 3 and the reliable sense path array is [ A, B, C, D, E ], the sense path ranges may be [ A, B, C ], [ B, C, D ] and [ C, D, E ].

The iterative process of the historical pond-level graph may include:

a) Target sensing separation distance sensing: and taking the current historical area unit group as a starting point, and performing distance sensing according to the target sensing interval. This results in the next set of history area units being connected to the current set of history area units. The result of the distance sensing will instruct the robot to select the next zone unit group. For example, according to the set distance sensing interval, starting from the current zone unit group, and sensing the next zone unit group connected thereto within a certain range.

b) Judging whether the next history area unit group includes a history departure position area unit: if the next history area unit group includes a history exit position area unit, the distance sensing is ended. At this time, the history originating location area unit and all the history area unit groups are combined to form a connected preset number of history area units.

The determining of the history candidate range may include determining a preset number of history area units as the history candidate range. These history area units will be candidate paths for inspection by the robot. For example, if the preset number is 5, then a number of history area units in front of and behind the history originating location and the history leaving location may be selected as history candidate ranges.

The layout of the historical patrol route may include performing a detour route layout on the historical origination location area unit and the historical departure location area unit according to the historical candidate range, resulting in a historical patrol route. The historical origination location area units correspond to patrol origination locations of the historical pond-level graph, and the historical departure location area units correspond to patrol departure locations of the historical pond-level graph. The predetermined number of history area units includes at least a history originating location area unit and a history leaving location area unit. For example, a path is planned between the area units according to the selection of the history candidate range, forming a history patrol route. Illustratively, assuming a pool inspection task, the pool is divided into the following area units (denoted by letters): A. b, C, D, E, F, G, H, I. The history originating location area element is a and the history leaving location area element is I.

The historical patrol route is now planned according to the historical candidate range. Assuming that the preset number is 5, the two front and rear history area units related to the history originating position and the history leaving position are selected as history candidate ranges, i.e., [ a, B, C, D, E, I ]. And then, carrying out bypass line layout according to the history candidate range to form a history inspection line. The method comprises the following specific steps: starting from the historical starting position area unit A, the area unit B adjacent to the historical starting position area unit A is selected as the next inspection point. Continuing from B, selecting the adjacent area unit C as the next inspection point. Then, turn from C to D, and select D as the next inspection point. Continuing from D, E is then selected as the next inspection point. Finally, turn from E to I, and select I as the last point of the patrol. According to the steps, the obtained historical inspection line is A, B, C, D, E and I.

In this example, the front and rear two area units associated with the history originating location and the history leaving location are selected as the defined range of the patrol route based on the history candidate range. Then, according to the connection relation between the area units, the next inspection point is gradually selected from the starting position until the departure position is reached. By the mode, a historical inspection line can be planned, so that the requirements of inspection tasks are met.

In the embodiment of the present invention, before the step of performing the foregoing distance sensing of the target sensing interval with the current history area unit group as the starting point to obtain the next history area unit group connected to the current history area unit group, the following example may be further included.

(1) Acquiring the number of the identified area units and the number of the connected area units, wherein the number of the identified area units is the number of the identified historical area units before the distance sensing of the current round, and the number of the connected area units is the number of the identified connected historical area units before the distance sensing of the current round;

(2) And acquiring a target sensing interval which has a negative feedback relation with the number of the identified area units and has a positive feedback relation with the number of the connected area units.

In the embodiment of the present invention, it is assumed that the robot starts the inspection from the area unit a. During the inspection process, the robot uses the sensor to perform target sensing and records the next historical region unit set (e.g., B, C, D) connected to the current historical region unit set (i.e., a). Through distance sensing, the robot can determine the distance between each connected regional unit group and the current regional unit group. Prior to the target sensing interval, the robot has identified some historical area units. It is assumed that the robot has identified the history area units A, B and C before the current round. At this time, the number of recognized area units is 3. Meanwhile, during the target sensing interval, the robot also finds the region units B, C and D connected to the current history region unit group according to the distance sensing result. Thus, the number of connected area units is 3. For the determination of the target sensing interval, there are two key factors: the number of identified area units and the number of connected area units. The robot adjusts the frequency of target sensing based on these factors.

If the robot observes that the number of recognized area units increases, the probability of finding the actual target is low, and a strategy to reduce the target sensing frequency may be adopted. For example, when a robot has identified multiple history area units, there may not be a great deal of benefit in continuing object sensing because there is a higher likelihood of detecting more duplicate objects. (negative feedback)

However, when the robot finds a greater number of newly added connected area units, it indicates that there may be more unexplored targets. In this case, the robot may take a strategy of increasing the target sensing frequency. By shortening the target sensing interval, the robot can discover potential targets faster and take corresponding actions in time. (Positive feedback)

In the inspection task of the plunge pool, the robot flexibly adjusts the interval time of target sensing according to the feedback relation between the number of the identified area units and the number of the connected area units. Therefore, the efficiency of the inspection task can be optimized, the target discovery rate is improved, and the robot can better adapt to the differences among different area units and the target distribution condition.

In the embodiment of the present invention, the following examples are also provided before the step of obtaining the pool level diagram to be patrolled and examined by performing the iteration based on the historical pool level diagram.

(1) According to the historical inspection line, the underwater robot is controlled to carry out inspection on the historical plunge pool level diagram;

(2) In the process of carrying out inspection on the historical plunge pool level diagram, identifying the environmental transition of the historical plunge pool level diagram;

(3) When the environmental transition of the historical plunge pool level diagram is identified, and the underwater robot does not reach the patrol leaving position, determining that the historical plunge pool level diagram is adjusted; or, based on the proactive adjustment command, determining that an adjustment occurs to the historical pond hierarchy.

In the implementation of the invention, it is assumed that the historical inspection lines are performed in the order of A, B, C, D, E and I. The underwater robot first reaches the area unit a, and sequentially patrols the area units B, C, D and E along the planned path, and finally reaches the patrol departure position I. The robot completes the inspection task along the path by adjusting the motion of the robot according to a preset movement and navigation strategy. During inspection, the underwater robot uses sensors and vision systems to sense changes in the surrounding environment. For example, the robot may detect a change in water quality, such as an increase in turbidity or temperature, through a water quality sensor; or the camera is used for identifying that new damage, cracks or barriers appear on the plunge pool surface. The identified environmental transitions can be used as a basis for a robot to determine whether the historical pond-level diagram needs to be adjusted. It is assumed that during the inspection the underwater robot detects the presence of a new area unit F and that the robot has not yet reached the predetermined inspection exit position I. In this case, it may be determined that the historical pond hierarchy chart needs to be adjusted, the newly added area unit F is added to the hierarchy chart, and the connection relationship with other area units is updated. Therefore, the robot can be ensured to cover the newly added area unit F in the next inspection task. Another scenario is that the commander may send commands to the underwater robot asking it to adjust the historical plunge pool hierarchy chart according to specific needs. For example, if a report is received regarding a security issue for a particular area, the command may issue a command that the robot increase coverage of that area in the next inspection mission and modify the historical pond level graph accordingly. Therefore, the robot can be ensured to carry out inspection according to the new instruction so as to meet the changing requirements of the task. Through the steps, the underwater robot can be controlled according to the historical inspection line, and the environmental transition is identified and the historical plunge pool level diagram is adjusted in the inspection process. This enables the robot to adapt to changes in the plunge pool and to flexibly cope with changes in task requirements.

In the embodiment of the invention, after the step of operating the underwater robot to perform inspection on the pool level diagram to be inspected according to the target inspection line is performed, the following implementation manner is further provided.

(1) In the process of controlling the underwater robot to patrol on the pool level diagram to be patrol according to the target patrol line, a new pool level diagram is obtained based on iteration of the pool level diagram to be patrol, wherein the new pool level diagram is the pool level diagram to be patrol after iteration;

(2) In the process of controlling the underwater robot to patrol on the new plunge pool level diagram according to the target patrol line, updating the target patrol line according to the bypass line of the identified non-passable area to obtain a new patrol line;

(3) And operating the underwater robot to patrol on the new plunge pool level diagram according to the new patrol line.

The underwater robot is assumed to carry out inspection according to a preset target inspection line A, B, C, D and E, and the inspection of the plunge pool to be inspected is successfully completed. At this point, the robot has collected data and information about the individual area units (mat surfaces) of the pond. In the inspection process, the underwater robot acquires data about the state of the mat surface, the water quality, the temperature and the like by using a sensor and a vision system. With these new data, iterations of the pool-level diagram to be inspected, i.e. updating the state or properties of the mat surface, can be performed. For example, if the robot detects that the mat surface of a certain area unit is severely worn, that area unit may be marked as an area requiring repair or replacement in the new pond level graph. During the inspection of the new pool level diagram, the underwater robot may encounter some non-trafficable areas, such as obstacles or damaged mats. By means of sensors and vision systems, the robot is able to detect these non-passable areas and recognize detours to avoid these obstacles. Based on the information of the identified non-passable area, the target inspection line can be updated to generate a new inspection line, so that the robot can bypass the non-passable area and continue inspecting other area units. Based on the updated inspection line, the underwater robot performs inspection on the new plunge pool level diagram according to new guidance. For example, the robot may sequentially patrol the area units a→c→d→b→e in the new line order. By performing inspection on the new plunge pool level diagram, the robot can collect updated data and information, providing more accurate references for maintenance, management or next inspection tasks.

Through the steps, the underwater robot can control and self-carry out inspection on the pool level diagram to be inspected according to the target inspection line, iterate in the inspection process, and obtain a new pool level diagram. Meanwhile, according to the information of the identified non-passable area, the target inspection line is updated, so that the robot can avoid the obstacle and smoothly complete the inspection task. For example, in an actual scenario, it is assumed that the pool-level diagram to be patrolled includes area units A, B, C and D. The initial target inspection line is A, B, C and D. During inspection, the underwater robot finds that the area unit C has a damaged mat surface, which results in no passage. The robot detects this by means of sensors and makes a path planning to bypass the area unit C. The robot may update the target patrol route according to the detouring route of the identified non-passable area. In this example, the updated inspection line may become a→b→d to avoid the damaged area unit C. And then, the underwater robot performs inspection on the new plunge pool level diagram according to the new inspection line operation, and sequentially inspects corresponding area units according to the sequence of A, B and D.

In the embodiment of the present invention, the aforementioned step S205 may be implemented by performing the following manner.

(1) Performing feature learning operation on the inspection result to obtain a feature matrix of the inspection result, wherein the feature matrix is used for indicating the density of the inspection video content in each region in the inspection result;

(2) Determining sample selection frequency of a corresponding region according to the density of the inspection video content in each region in the feature matrix;

(3) According to the sample selection frequency, sample selection is carried out on the inspection result to obtain a plurality of first inspection frequency bands, the inspection frequency content is determined according to image elements in the inspection result, and the density of the inspection frequency content has positive feedback relation with the sample selection frequency;

(4) For any first inspection frequency band, performing a cutting operation on the first inspection frequency band to obtain a plurality of second inspection frequency bands;

(5) The method comprises the steps of obtaining feature vectors of a plurality of second inspection frequency bands, wherein the feature vectors of the plurality of second inspection frequency bands are determined by combining video attributes of the plurality of second inspection frequency bands and space attributes of the plurality of second inspection frequency bands;

(6) According to the first standardization module, carrying out standardization on the feature vectors of the plurality of second inspection frequency bands;

(7) Processing the feature vectors of the standardized multiple second inspection frequency bands according to a long-period memory network provided with a preset static filter to obtain a first transition feature vector;

(8) Determining a second transition feature vector according to the feature vectors of the plurality of second inspection frequency bands and the first transition feature vector;

(9) Normalizing the second transition feature vector according to a second normalization module;

(10) Processing the second transition feature vector according to the deep full-connection network to obtain intermediate inspection video vectors of a plurality of second inspection frequency bands;

(11) According to a long-term memory network, a standardized module and a deep full-connection network which are provided with a preset dynamic filter, extracting features of intermediate inspection video vectors of a plurality of second inspection frequency bands to obtain inspection video vectors of a first inspection frequency band, and carrying out feature reconstruction on the inspection video vectors of the first inspection frequency band for any first inspection frequency band to obtain abnormal grades corresponding to a plurality of image elements in the first inspection frequency band;

(12) Generating an inspection video frame of the first inspection frequency band according to the abnormal grades corresponding to the plurality of image elements in the first inspection frequency band, wherein element characteristics of the image elements in the inspection video frame are used for representing the corresponding abnormal grades, and each inspection video frame is configured with the abnormal grade;

(13) Performing identification operation according to the inspection video frames of the plurality of first inspection frequency bands to obtain the danger level of each first inspection frequency band;

(14) And evaluating the dangerous grade of the inspection result according to the dangerous grade of each first inspection frequency band to obtain an inspection analysis report of the inspection result.

In the embodiment of the invention, the video data of the pool inspection can be analyzed by using a computer vision technology. For example, an image processing algorithm is used to detect defects, cracks, or corrosion, and to extract characteristic information such as color, texture, shape, etc. of these areas. According to the feature matrix in the inspection result, the abnormal density of each region can be determined. For example, the anomaly density (i.e., the problem severity) is determined by calculating the number of anomaly pixels or the average of anomaly levels in each region. A higher density indicates that the region may have a more serious problem and therefore requires more frequent selection of the region as a sample for further analysis and evaluation. Based on the sample selection frequency, a plurality of first inspection frequency bands are selected from the inspection result to serve as samples. For example, assuming that 10 areas with higher anomaly densities are found in the inspection result, one first inspection frequency band in each area may be selected as a sample, and 10 first inspection frequency bands in total are selected. For each selected first inspection frequency band, it is cut into a plurality of second inspection frequency bands. For example, if the first inspection band is 10 minutes long, it may be selected to be divided into 10 second inspection bands lasting 1 minute. This ensures that each second inspection frequency segment contains a particular event or anomaly. And extracting the characteristic vector of each second inspection frequency band. For example, for image features, color distribution information may be extracted using a color histogram algorithm; for texture features, local Binary Pattern (LBP) may be used to extract texture information; for shape features, edge information may be extracted using an edge detection algorithm. Thus, the characteristic information of each second inspection frequency band can be captured from different angles. And carrying out standardization processing on the feature vectors extracted from the plurality of second inspection frequency bands. This may include normalizing feature vectors to similar scales and distributions to ensure that they are comparable and to reduce the bias between different features. For example, the feature vector may be converted to a value with zero mean and unit variance using Z-score normalization. And processing the normalized feature vector by using a preconfigured Long Short Term Memory (LSTM) network, a normalization module and a deep full connection (DNN) network. For example, long-term dependencies in a feature sequence may be captured through an LSTM network and a first transitional feature vector generated for further feature extraction and analysis. And performing a feature reconstruction operation for each first inspection frequency band to restore the feature representation of the original image. May be implemented using a self-encoder or other reconstruction model. The anomaly level of each image element is calculated by comparison with the original image. For example, by calculating pixel level errors or reconstruction losses, it is determined whether the element is abnormal and assigned a corresponding level of abnormality. And generating a video inspection frame based on the abnormal grade result of the first video inspection frequency band. In these frames, each image element will be marked as having a corresponding level of abnormality to reflect its degree of abnormality or classification. The anomaly level may be added to the inspection video frame in the form of a color map or a symbol mark to visualize anomalies at different locations in the pipeline. And using a proper algorithm or model to perform identification operation on the generated inspection video frame. By analyzing the abnormal level and distribution of the image elements in each first inspection frequency band, a corresponding risk level can be obtained. For example, the inspection results are classified into different hazard levels, such as low, medium, high, or by comparison with a predefined threshold, depending on the number, density, and distribution of anomaly levels. And comprehensively considering the rating results of all the inspection frequency bands according to the risk level of each first inspection frequency band, and evaluating the risk level of the whole inspection result. The inspection results of different grades can be weighted or summarized to determine the overall dangerous degree of the inspection results, and corresponding inspection analysis reports are generated. In addition, detailed inspection reports can be generated according to the risk level evaluation and the analysis of inspection results. The report should contain an abnormality, risk level, and possibly cause analysis for each area. In addition, suggested maintenance plans, including repair measures, precautions, and optimization schemes, may also be provided to reduce the occurrence of potential problems with the pipeline. In order to ensure the safety and reliability of the plunge pool, repeated inspection and monitoring is required periodically. This can be achieved by setting up a patrol plan, and re-patrol the pipeline at certain time intervals. The new inspection data will be compared with the historical data to detect any new anomalies or changes and take corresponding action to process. By continuously collecting and analyzing inspection data, monitoring data, and implemented maintenance plans, potential problems and opportunities for improvement in the system can be identified. The feedback information can be used for improving the inspection flow, the optimization algorithm and the model, and improving the accuracy and the efficiency of the pool inspection.

In the embodiment of the present invention, the step of performing the identification operation according to the inspection video frames of the plurality of first inspection frequency bands to obtain the risk level of each first inspection frequency band may be implemented by the following example.

(1) For the inspection video frames of any first inspection frequency band, acquiring the number of image elements corresponding to each abnormal level in the inspection video frames of the first inspection frequency band;

(2) Determining the dangerous level of the first inspection frequency band according to the number of the image elements corresponding to each abnormal level;

the step of performing the identifying operation according to the inspection video frames of the plurality of first inspection frequency bands to obtain the risk level of each first inspection frequency band may also be performed by the following example.

(1) Correcting abnormal grades corresponding to a plurality of image elements in the inspection video frame according to the relative layout among the plurality of image elements in the inspection video frame for the inspection video frame of any first inspection frequency band;

(2) And obtaining the dangerous grade of the first inspection frequency band according to the number of the image elements corresponding to each abnormal grade in the corrected inspection video frame.

In the embodiment of the present invention, it is assumed that a patrol video frame of a certain first patrol frequency band includes three abnormal-level image elements: high, medium and low. By analyzing the image elements corresponding to each anomaly level in the inspection video frame, a specific number can be obtained. For example, there are 30 image elements for high level anomalies, 50 for medium level anomalies and 100 for low level anomalies. Based on the above scenario, the risk level of the first inspection frequency band may be calculated according to the number of image elements of different anomaly levels. For example, if the number of specified high-level abnormal image elements exceeds 50, it is determined that the risk level is high, if the number of intermediate-level abnormal image elements is between 20 and 50, it is determined that the risk level is medium, and if the number of low-level abnormal image elements is less than 20, it is determined that the risk level is low. According to the rules, the risk level of the first inspection frequency band can be determined as a medium risk level.

In another embodiment of the present invention, in the video frame of a first video segment, there is a centrally distributed anomaly region, wherein a portion of the image elements are incorrectly classified as low-level anomalies. By analyzing the relative layout among a plurality of image elements in the patrol video frame, the abnormal region which is intensively distributed can be detected, and the abnormal level of the related image elements is corrected, so that the level of the abnormal region is improved. For example, an image element that was misclassified as low-level may be reclassified as medium-level anomaly after correction. Based on the scene, after the inspection video frame of the first inspection frequency band is corrected, the number of image elements corresponding to each abnormal level is recalculated. For example, in the modified video frame, there are 25 image elements with high level anomalies, 40 image elements with medium level anomalies, and 90 image elements with low level anomalies. Based on the new number, the risk level of the first inspection frequency band may be redetermined. If the previous decision rule is followed, the risk level of the first inspection frequency band may be modified to a low risk level.

The embodiment of the invention provides a computer device 100, wherein the computer device 100 comprises a processor and a nonvolatile memory storing computer instructions, and when the computer instructions are executed by the processor, the computer device 100 executes the plunge pool inspection method. As shown in fig. 2, fig. 2 is a block diagram of a computer device 100 according to an embodiment of the present invention. The computer device 100 comprises a memory 111, a processor 112 and a communication unit 113. For data transmission or interaction, the memory 111, the processor 112 and the communication unit 113 are electrically connected to each other directly or indirectly. For example, the elements may be electrically connected to each other via one or more communication buses or signal lines.

The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, and alternatives falling within the spirit and principles of the invention.

Claims (10)

1. A method for inspecting a plunge pool, the method comprising:

responding to a patrol task starting instruction aiming at a target plunge pool, and establishing communication connection with the underwater robot;

Acquiring pool map information corresponding to the target pool, wherein the pool map information at least comprises historical pool level diagrams of a plurality of depths;

determining a target inspection line and operating the underwater robot to inspect according to a historical plunge pool level diagram corresponding to a target depth, wherein the target depth is any depth of the plurality of depths;

receiving inspection results of the underwater robot for the plurality of depths;

and analyzing the inspection result to obtain an inspection analysis report of the inspection result.

2. The method of claim 1, wherein determining a target inspection line and maneuvering the underwater robot for inspection based on the historical pool level diagram corresponding to the target depth comprises:

obtaining a pool level diagram to be inspected based on iteration of the historical pool level diagram, wherein the pool level diagram to be inspected is the historical pool level diagram after iteration, and an inspection line of the historical pool level diagram is a historical inspection line;

in the process of controlling the underwater robot to patrol on the plunge pool level diagram to be patrolled according to the historical patrol line, identifying an unvented area to obtain a target unvented area;

Determining a pending non-passable range according to the detour originating location;

determining the area unit bypass cost of each current area unit in the undetermined non-passable range according to at least one of a first space span and a second space span, wherein the first space span is the space span between the current area unit and the bypass starting position, and the second space span is the space span between the current area unit and the target non-passable area;

acquiring a current area unit group on each candidate patrol strategy in a plurality of candidate patrol strategies in the undetermined non-passable range;

determining an array of the regional unit detour cost corresponding to the current regional unit group according to the regional unit detour cost;

performing data normalization on the regional unit detour cost array to obtain the detour cost corresponding to each candidate routing strategy;

selecting at least one detour strategy from a plurality of candidate patrol strategies according to the detour cost;

determining a current candidate range in the undetermined non-passable range according to at least one detour strategy;

traversing a passable area unit group adjacent to a current originating location area unit in the current candidate range, wherein the current originating location area unit is a current area unit corresponding to the bypass originating location;

Acquiring the routing inspection cost corresponding to each current area unit in the passable area unit group, and determining the current area unit which is traversed most recently and corresponds to the routing inspection cost as a previous-level area unit, wherein the routing inspection cost is the cost from the current originating position area unit to the current leaving position area unit through the corresponding current area unit;

traversing a next passable area unit group adjacent to a target current area unit until the traversed passable area unit group comprises the current departure position area unit, terminating the traversing, and carrying out backtracking of a previous-stage area unit from the current departure position area unit to the current starting position area unit to obtain a target bypass line between the bypass starting position and the bypass departure position, wherein the target current area unit is the current area unit which is not traversed and has the optimal routing inspection cost, the target bypass line is used for bypassing the target non-passable area from the bypass starting position to reach the bypass departure position, the bypass starting position is a bypass starting position on the historical routing inspection line, and the bypass departure position is a bypass ending position on the historical routing inspection line;

Adjusting the historical inspection line according to the target detour line to obtain the target inspection line;

and according to the target inspection line, the underwater robot is controlled to inspect on the hierarchical diagram of the plunge pool to be inspected.

3. The method of claim 2, wherein when at least one of the detour policies is a plurality of the detour policies, the determining the current candidate range of the pending non-passable range according to at least one of the detour policies comprises:

obtaining a plurality of local candidate ranges corresponding to the detour strategies, and combining the local candidate ranges to form the current candidate range in the undetermined non-passable range; or determining a target detour strategy from a plurality of detour strategies according to a current departure position area unit, and determining a candidate range on the target detour strategy as the current candidate range in the undetermined non-passable range, wherein the current departure position area unit is a current area unit corresponding to the detour departure position.

4. The method of claim 2, wherein prior to obtaining the pool level diagram to be patrolled based on the iteration of the historical pool level diagram, the method further comprises:

Performing region unit segmentation on the historical plunge pool hierarchical graph to obtain a region unit set, wherein the region unit set comprises a plurality of historical region units;

in the area unit set, taking a historical initial position area unit as a starting point, and determining a plurality of sensing paths corresponding to a plurality of preset deflection amounts according to preset sensing intervals;

extracting a reliable sensing path array from a plurality of sensing paths according to preset rules, wherein the preset rules are determined according to at least one of sensing distance and preset commands of a target underwater robot;

acquiring a preset number of reliable sensing paths with the preset deflection quantity being adjacent in sequence from the reliable sensing path array, and acquiring a sensing path range corresponding to the preset number of reliable sensing paths;

combining the history area units in the sensing path range which is larger than a preset range to form a first history area unit group connected with the history starting position area unit;

when the first history area unit group does not include the history departure location area unit, continuing the following processing:

performing distance sensing of a target sensing interval by taking a current history area unit group as a starting point to obtain a next history area unit group connected with the current history area unit group, wherein the connecting coincidence rate of the target sensing interval and the distance sensing has a positive feedback relationship;

When the next history area unit group comprises the history leaving position area units, ending the distance sensing, and combining the history starting position area units and all the history area unit groups to form a connected preset number of history area units;

determining a preset number of history area units as history candidate ranges;

and carrying out bypass line layout on a history starting position area unit and a history leaving position area unit according to the history candidate range to obtain the history inspection line, wherein the history starting position area unit is the history area unit corresponding to the inspection starting position of the history plunge pool level diagram, the history leaving position area unit is the history area unit corresponding to the inspection leaving position of the history plunge pool level diagram, and the preset number of history area units at least comprises the history starting position area unit and the history leaving position area unit.

5. The method of claim 4, wherein the method further comprises, before the distance sensing of the target sensing interval starting from the current history area unit group to obtain the next history area unit group connected to the current history area unit group:

Acquiring the number of identified area units and the number of connected area units, wherein the number of identified area units is the number of the history area units identified before the distance sensing of the current round, and the number of connected area units is the number of the history area units identified before the distance sensing of the current round;

and acquiring the target sensing interval which has a negative feedback relation with the number of the identified area units and has a positive feedback relation with the number of the connected area units.

6. The method of claim 2, wherein prior to obtaining the pool level diagram to be patrolled based on the iteration of the historical pool level diagram, the method further comprises:

according to the historical inspection line, the underwater robot is controlled to conduct inspection on the historical plunge pool level diagram;

during the process of carrying out inspection on the historical plunge pool level diagram, identifying the environmental transition of the historical plunge pool level diagram;

when the environmental transition of the historical plunge pool level diagram is identified, and the underwater robot does not reach the patrol leaving position, determining that the historical plunge pool level diagram is adjusted; or, based on the active adjustment command, determining that the historical pond hierarchy map is adjusted.

7. The method of claim 2, wherein the maneuvering the underwater robot following the inspection on the pool-to-be-inspected hierarchical map according to the target inspection line, the method further comprises:

in the process of controlling the underwater robot to patrol on the pool level diagram to be patrol according to the target patrol line, obtaining a new pool level diagram based on iteration of the pool level diagram to be patrol, wherein the new pool level diagram is the pool level diagram to be patrol after iteration;

in the process of controlling the underwater robot to patrol on the new plunge pool level diagram according to the target patrol line, updating the target patrol line according to the bypass line of the identified non-passable area to obtain a new patrol line;

and operating the underwater robot to patrol on the new plunge pool level diagram according to the new patrol line.

8. The method of claim 1, wherein analyzing the inspection result to obtain an inspection analysis report of the inspection result comprises:

performing feature learning operation on the inspection result to obtain a feature matrix of the inspection result, wherein the feature matrix is used for indicating the density of inspection video content in each region in the inspection result;

Determining sample selection frequency of a corresponding region according to the density of the inspection video content in each region in the feature matrix;

according to the sample selection frequency, sample selection is carried out on the inspection result to obtain a plurality of first inspection frequency bands, the inspection frequency content is determined according to image elements in the inspection result, and the density of the inspection frequency content has a positive feedback relation with the sample selection frequency;

for any first inspection frequency band, performing a cutting operation on the first inspection frequency band to obtain a plurality of second inspection frequency bands;

the feature vectors of the plurality of second inspection frequency bands are obtained and are determined by combining the video attributes of the plurality of second inspection frequency bands and the spatial attributes of the plurality of second inspection frequency bands;

according to the first standardization module, the feature vectors of the plurality of second inspection frequency bands are standardized;

processing the normalized feature vectors of the plurality of second inspection frequency bands according to a long-period memory network provided with a preset static filter to obtain a first transition feature vector;

determining a second transition feature vector according to the feature vectors of the plurality of second inspection frequency bands and the first transition feature vector;

Normalizing the second transition feature vector according to a second normalization module;

processing the second transition feature vector according to the deep full-connection network to obtain intermediate inspection video vectors of the plurality of second inspection frequency bands;

according to a long-period memory network, a standardization module and a deep full-connection network which are configured with a preset dynamic filter, extracting features of intermediate inspection video vectors of a plurality of second inspection frequency bands to obtain inspection video vectors of the first inspection frequency bands, and carrying out feature reconstruction on the inspection video vectors of the first inspection frequency bands for any one first inspection frequency band to obtain abnormal grades corresponding to a plurality of image elements in the first inspection frequency bands;

generating an inspection video frame of the first inspection frequency band according to the abnormal grades corresponding to a plurality of image elements in the first inspection frequency band, wherein element characteristics of each image element in the inspection video frame are used for representing the corresponding abnormal grades, and each inspection video frame is configured with the abnormal grade;

performing identification operation according to the inspection video frames of the plurality of first inspection frequency bands to obtain the danger level of each first inspection frequency band;

And evaluating the dangerous grade of the inspection result according to the dangerous grade of each first inspection frequency band to obtain an inspection analysis report of the inspection result.

9. The method of claim 8, wherein the performing an identification operation based on the inspection video frames of the plurality of first inspection frequency bands to obtain a risk level for each of the first inspection frequency bands comprises:

for a patrol video frame of any first patrol video frequency band, acquiring the number of image elements corresponding to each abnormal level in the patrol video frame of the first patrol video frequency band;

determining the dangerous level of the first inspection frequency band according to the number of the image elements corresponding to each abnormal level;

the identifying operation is performed according to the inspection video frames of the plurality of first inspection frequency bands to obtain the danger level of each first inspection frequency band, and the method further comprises the following steps:

correcting abnormal grades corresponding to a plurality of image elements in the inspection video frame according to the relative layout among the plurality of image elements in the inspection video frame for the inspection video frame of any first inspection frequency band;

and obtaining the dangerous grade of the first inspection frequency band according to the number of the image elements corresponding to each abnormal grade in the corrected inspection video frame.

10. A server system comprising a server for performing the method of any of claims 1-9.