A path planning method, system, medium and device for a river traveling device

Abstract

The present invention discloses a path planning method for a river channel walking device, which realizes the accuracy, intelligence and adaptability of path planning in the process of river channel dredging through technical means such as multi-dimensional information acquisition, landform image similarity matching, multi-unit path planning and fusion, and real-time dynamic adjustment of water flow parameters. The method can effectively improve the dredging efficiency, reduce the operation risk, optimize the overall performance of river channel dredging operations, accurately match the river channel landform characteristics; dynamically adjust the dredging path; monitor water flow parameters in real time; intelligently plan the path; and improve the safety and accuracy of dredging operations.

Description

Technical Field

[0001] The present invention relates to the field of path planning, and in particular to a path planning method, system, medium and equipment for a river traveling device. Background Art

[0002] River dredging refers to the use of certain technical means and methods to remove silt and solid waste deposited in the river channel to maintain the smooth flow of the river channel, improve water quality and ecological environment, and ensure the normal flow of water. Existing river dredging generally uses dredging boats on the water surface to extend to the bottom of the riverbed through pipelines to carry out dredging operations. During the dredging process, the pipeline on the water surface cannot know the specific conditions of the bottom of the river channel during the extension process, resulting in a relatively blind dredging path, and in larger dredging areas, it takes a certain period of time to complete the dredging task. Due to the environmental differences between the river surface and the bottom of the river channel, it is impossible to fully grasp the areas that have been cleaned and the areas that have not been cleaned in the dredging area, resulting in low dredging efficiency and omissions in the dredging area. Summary of the invention

[0003] In view of this, the purpose of the present invention is to propose a path planning method, system, medium and equipment for a river walking device.

[0004] In order to achieve the above technical objectives, the technical solution adopted by the present invention is:

[0005] In a first aspect, the present invention provides a path planning method for a river channel running device, which is applicable to the river channel running device, and the river channel running device is used to perform a dredging task of the current river channel, and the method includes:

[0006] S101, obtaining first river channel information of the current river channel, where the first river channel information includes at least one of first river length information, first river channel name, first historical water flow parameter, first historical dredging parameter, and first surrounding environment information;

[0007] S102, selecting riverbed geomorphic images of rivers similar to the current first river channel information in a database according to the first river channel information to obtain a reference image data set, where the reference image data set includes a reference riverbed image;

[0008] S103, obtaining a regional geomorphic image of the current first river channel information, where the regional geomorphic image is a partial geomorphic image of the riverbed of the current river channel;

[0009] S104, performing similarity matching between the regional geomorphic image and the reference riverbed image in the reference image dataset, and obtaining a reference riverbed image having the highest similarity to the current regional geomorphic image, which is recorded as the reference riverbed image;

[0010] S105, constructing a riverbed geomorphology model of the current river channel according to the reference riverbed image, and generating a first planned path of the current river channel walking device according to the riverbed geomorphology model, wherein the first planned path is the overall walking direction information of the river channel walking device in the current river channel;

[0011] S106, dividing the riverbed geomorphology model along the extension direction of the current river channel into a plurality of dredging units according to preset conditions, each dredging unit having a regional geomorphology model;

[0012] S107, inputting the dredging units and the first historical water flow parameters into the path planning model one by one to obtain a second planned path for each dredging unit;

[0013] S108, arranging the plurality of second planned paths according to the arrangement order of the dredging units, and merging them with the first planned path to obtain a final planned path;

[0014] S109, controlling the river channel walking device to perform the dredging task according to the final planned path, and collecting the real-time water flow parameters of the current river channel at a preset frequency during the walking process;

[0015] S110, obtaining a water flow threshold of the current position information of the river channel traveling device according to the first historical water flow parameter;

[0016] S111, determining whether the real-time water flow parameter is within the range of the water flow threshold;

[0017] S112, if not, generating an adjustment strategy for the river channel traveling device in a preset area according to the real-time water flow parameters, the adjustment strategy including traveling speed adjustment and path adjustment of the river channel traveling device;

[0018] S113, recording the adjusted path as a first adjustment path, obtaining a partial path corresponding to the current preset area in the final planned path, and recording it as a first regression path;

[0019] S114, calculating a displacement deviation value between the first adjustment path and the first regression path;

[0020] S115, determining whether the displacement deviation value is within a preset deviation threshold value;

[0021] S116, if not, the remaining dredging units that have not yet performed the dredging task are recorded as the units to be dredged, and the dredging units and the real-time water flow parameters are re-input into the path planning model one by one to obtain the feedback planning path, and the feedback planning path is updated to the final planning path;

[0022] Repeat steps S109 to S116 until the dredging task of the entire river channel is completed.

[0023] In some embodiments, other rivers in the database have second river information, and the second river information includes at least one of second river length information, second river name, second historical water flow parameter, second historical dredging parameter, and second surrounding environment information;

[0024] According to the first river channel information, riverbed geomorphic images similar to the current first river channel information are screened in the database to obtain a reference image data set including:

[0025] Constructing a semantic dictionary, converting the first river channel information into a first semantic feature, and converting the second river channel information of other rivers in the database into a second semantic feature;

[0026] The semantic dictionary is expressed by formula (1), which is as follows:

[0027] ;

[0028] In formula (1), is a semantic dictionary, is the characteristic vector representation of river information, is the characteristic vector representation of the river length information, is the feature vector representation of the river name, is the characteristic vector representation of historical water flow parameters, is the characteristic vector representation of historical desilting parameters, is the feature vector representation of the surrounding environment information, for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameter of

[0029] The first semantic feature is described according to the semantic dictionary and is expressed by formula (2). Formula (2) is as follows:

[0030] ;

[0031] In formula (2), is the first semantic feature, is the feature vector representation of the first river channel information, is the feature vector representation of the first river length information, is the feature vector representation of the name of the first river, is the eigenvector representation of the first historical flow parameter, is the eigenvector representation of the first historical desilting parameter, is the feature vector representation of the first surrounding environment information, for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameter of

[0032] The second semantic feature is described according to the semantic dictionary and is expressed by formula (3). Formula (3) is as follows:

[0033] ;

[0034] In formula (3), is the second semantic feature, is the feature vector representation of the second river channel information, is the feature vector representation of the second river length information, is the feature vector representation of the second river name, is the eigenvector representation of the second historical flow parameter, is the eigenvector representation of the second historical desilting parameter, is the feature vector representation of the second surrounding environment information, for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameter of

[0035] The similarity between the second semantic feature and the first semantic feature is calculated one by one and recorded as the semantic similarity value. The similarity is calculated by formula (4), which is as follows:

[0036] ;

[0037] In formula (4), For the The semantic similarity value of the second river information, is the similarity function, For the The second semantic feature of the second river information, , N is a positive integer;

[0038] It is determined whether the semantic similarity values ​​exceed the range of the semantic similarity threshold one by one. If so, the riverbed topography image in the second river channel information associated with the semantic similarity value is screened and stored to obtain a reference image data set, which is expressed by formula (5). Formula (5) is as follows:

[0039] ;

[0040] In formula (5), is the semantic similarity threshold, is the reference image dataset, for Associated riverbed geomorphic image.

[0041] In some embodiments, the regional geomorphic image is matched with the reference riverbed image in the reference image dataset for similarity, and a reference riverbed image having the highest similarity to the current regional geomorphic image is obtained, which is recorded as the reference riverbed image, including:

[0042] Inputting the regional landform image into the texture recognition model to obtain the texture feature of the regional landform image, recorded as the first texture feature, and inputting the reference riverbed images into the texture recognition model one by one to obtain the texture feature of each reference riverbed image, recorded as the second texture feature;

[0043] Inputting the regional landform image into the edge recognition model to obtain the edge feature of the regional landform image, recorded as the first edge feature, and inputting the reference riverbed images into the edge recognition model one by one to obtain the edge feature of each reference riverbed image, recorded as the second edge feature;

[0044] Obtaining a first image feature vector according to the first texture feature and the first edge feature, and obtaining a second image feature vector according to the second texture feature and the second edge feature;

[0045] The mutual correlation coefficient between the second image feature vector and the first image feature vector is calculated one by one. The mutual correlation coefficient can be expressed by formula (6), which is as follows:

[0046] ;

[0047] In formula (6), is the first image feature vector, For the The second image feature vector of the reference riverbed image, For the The correlation coefficient between the reference riverbed image and the regional topographic image, is the first horizontal coordinate of the pixel point in the regional landform image, is the second horizontal coordinate of the pixel point in the regional landform image, is the height coordinate of the pixel point in the regional topographic image, is the first horizontal offset of the pixel in the reference riverbed image, is the second horizontal offset of the pixel in the reference riverbed image, is the height offset of the pixel in the reference riverbed image;

[0048] The reference riverbed image associated with the second image feature vector having the largest value of the mutual correlation coefficient is selected and recorded as the reference riverbed image.

[0049] In some embodiments, recording the second image feature vector of the reference riverbed image as the reference image feature vector, and constructing the riverbed geomorphology model of the current river channel according to the reference riverbed image includes:

[0050] The first image feature vector is normalized to obtain a first normalized feature vector. The first normalized feature vector is expressed by formula (7). Formula (7) is as follows:

[0051] ;

[0052] In formula (7), is the first normalized eigenvector, is the first lower limit, is the first upper limit value;

[0053] The first standardized feature vector and the reference image feature vector are used to construct a feature association matrix, which is expressed by formula (8). Formula (8) is as follows:

[0054] ;

[0055] In formula (8), is the feature correlation matrix, is the first normalized eigenvector, is the reference image feature vector, for The variance of for The variance of for and The covariance of for and The covariance of and same;

[0056] Perform principal component analysis based on the feature correlation matrix to obtain the first difference eigenvector;

[0057] According to the first difference eigenvector, a riverbed trend change function is constructed in the DEM algorithm model. The riverbed trend change function is expressed by formula (9), which is as follows:

[0058] ;

[0059] In formula (9), is the riverbed trend change function, For time, is the first difference eigenvector, is the model parameter value corresponding to the first difference eigenvector, is the feature influence coefficient corresponding to the first difference eigenvector, is the model parameter value of the characteristic influence coefficient, is the first surrounding environment information, is the model parameter value of the first surrounding environment information;

[0060] The preset landform model in the DEM algorithm model is updated according to the riverbed trend change function to generate the riverbed landform model corresponding to the current river channel. The riverbed landform model is expressed by formula (10), which is as follows:

[0061] ;

[0062] In formula (10), is the riverbed geomorphic model of the current river channel, It is the preset landform model in the DEM algorithm model.

[0063] In some embodiments, the preset condition is a dredging rate matching coefficient, and dividing a plurality of dredging units according to the preset condition along the extension direction of the current river channel in the riverbed topography model includes:

[0064] Obtaining first historical dredging parameters, the first historical dredging parameters including the first historical total amount of silt, the first historical dredging cycle, and the first historical daily dredging mileage corresponding to each batch of dredging tasks in the current river channel;

[0065] Generate a first historical dredging rate according to the first historical total silt volume, the first historical dredging cycle, and the first historical daily dredging mileage;

[0066] Obtain the standard dredging rate of the current river travel device and the initial dredging unit length;

[0067] A rate matching coefficient is obtained according to the standard dredging rate and the first historical dredging rate;

[0068] The desilting unit division length is calculated based on the rate matching coefficient and the initial desilting unit length;

[0069] The predicted silt volume and predicted dredging time of each dredging unit are calculated based on the length of the dredging unit and are organized into a list for display.

[0070] In some embodiments, the first historical water flow parameter includes a first historical flow velocity distribution, a first historical water flow direction, a first historical water flow velocity, and a first seasonal variation parameter. The dredging units and the first historical water flow parameters are input into the path planning model one by one, and the second planned path for each dredging unit is obtained, including:

[0071] For each desilting unit, perform the following steps:

[0072] The regional geomorphic model corresponding to the current dredging unit is converted into a cost matrix, which includes a travel cost and a movement cost. The travel cost is the terrain complexity of the current regional geomorphic model, and the movement cost is the first historical flow velocity distribution. The first historical water flow direction is used as the path selection weight.

[0073] and, generating a terrain constraint condition based on the regional geomorphic model and generating a water flow constraint condition based on the first historical water flow parameter;

[0074] The cost matrix, path selection weight, terrain constraints and water flow constraints are used as input parameters of the path planning model to obtain the optimal path of the current dredging unit, which is also the second planning path;

[0075] Arrange multiple second planned paths according to the arrangement order of the dredging units, and merge them with the first planned path to obtain the final planned path expressed by formula (11). Formula (11) is as follows:

[0076] ;

[0077] In formula (11), For the final planning path after fusion, For the first planning path, For the The second planning path, is the number of the second planned paths, is the fusion coefficient, which is obtained by fitting the path complexity and the degree of local change of the path.

[0078] In some embodiments, the walking speed adjustment of the river walking device includes a speed adjustment value, and the speed adjustment value is obtained by the following steps:

[0079] Obtain the basic walking speed and real-time walking direction of the river walking device;

[0080] Obtaining real-time water flow parameters, including real-time water flow direction and real-time water flow velocity;

[0081] Generate a first influence coefficient according to the real-time walking direction and the real-time water flow direction;

[0082] Generate a second influence coefficient according to the real-time water flow velocity and the basic walking speed;

[0083] Obtaining a speed adjustment value according to the first influence coefficient, the second influence coefficient and the basic walking speed;

[0084] The first regulation path is obtained by the following steps:

[0085] The real-time water flow parameters and the current dredging unit are input into the path planning model to obtain the first adjustment path.

[0086] In a second aspect, the present invention provides a path planning system for a river channel traveling device, comprising a river channel traveling device, a sensor component, a dredging component and a control unit, wherein the river channel traveling device comprises a supporting platform, a driving unit and traveling wheels, wherein the traveling wheels are arranged below the supporting platform, and the driving unit is arranged on the supporting platform; the sensor component is arranged on the supporting platform, and the sensor component comprises at least one of a sonar sensor, a laser sensor, a water flow sensor and a silt sensor; the dredging component comprises a dredging pipeline, which is arranged on the supporting platform, and the dredging pipeline is used to extract silt from the bottom of the river channel; the control unit is arranged on the supporting platform, and the control unit is electrically connected to the driving unit, the sensor component and the dredging component, respectively, and the control unit is used to execute the method described in the first aspect.

[0087] In a third aspect, the present invention further provides a computer-readable storage medium storing computer program instructions, which implement the method described in the first aspect when executed by a processor.

[0088] In a fourth aspect, the present invention further provides an electronic device, comprising a memory and a processor, wherein the memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method described in the first aspect.

[0089] By adopting the above technical solution, the present invention has the following beneficial effects compared with the prior art:

[0090] Different from the existing technology, the above technical solution realizes the accuracy, intelligence and adaptability of path planning in the process of river dredging through technical means such as multi-dimensional information acquisition, similarity matching of landform images, multi-unit path planning and fusion, and real-time dynamic adjustment of water flow parameters. This method can effectively improve the efficiency of dredging, reduce operational risks, and optimize the overall performance of river dredging operations. It has significant technical innovation value and practical application prospects, accurately matches river geomorphic characteristics; dynamically adjusts dredging paths; monitors water flow parameters in real time; intelligently plans paths; and improves the safety and accuracy of dredging operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0091] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0092] Figure 1 is a flowchart of steps S101 to S116 of the path planning method described in the specific implementation method;

[0093] Figure 2 It is a schematic diagram of the path planning method described in the specific implementation method performing the dredging task in the current river channel.

[0094] Reference numerals:

[0095] 1. River channel;

[0096] 11. Dredging unit;

[0097] 2. The first planning path;

[0098] 3. Second planning path;

[0099] 4. River traveling device. DETAILED DESCRIPTION

[0100] The present invention will be further described in detail below in conjunction with the accompanying drawings and examples. It is particularly noted that the following examples are only used to illustrate the present invention, but are not intended to limit the scope of the present invention. Similarly, the following examples are only partial embodiments of the present invention rather than all embodiments, and all other embodiments obtained by those of ordinary skill in the art without creative work are within the scope of protection of the present invention.

[0101] See also Figure 1 and Figure 2In a first aspect, this embodiment provides a path planning method for a river channel walking device, which is applicable to the river channel walking device, and the river channel walking device is used to perform a dredging task of the current river channel, and the method includes:

[0102] S101, obtaining first river channel information of the current river channel, where the first river channel information includes at least one of first river length information, first river channel name, first historical water flow parameter, first historical dredging parameter, and first surrounding environment information;

[0103] S102, selecting riverbed geomorphic images of rivers similar to the current first river channel information in a database according to the first river channel information to obtain a reference image data set, where the reference image data set includes a reference riverbed image;

[0104] S103, obtaining a regional geomorphic image of the current first river channel information, where the regional geomorphic image is a partial geomorphic image of the riverbed of the current river channel;

[0105] S104, performing similarity matching between the regional geomorphic image and the reference riverbed image in the reference image dataset, and obtaining a reference riverbed image having the highest similarity to the current regional geomorphic image, which is recorded as the reference riverbed image;

[0106] S105, constructing a riverbed geomorphology model of the current river channel according to the reference riverbed image, and generating a first planned path of the current river channel walking device according to the riverbed geomorphology model, wherein the first planned path is the overall walking direction information of the river channel walking device in the current river channel;

[0107] S106, dividing the riverbed geomorphology model along the extension direction of the current river channel into a plurality of dredging units according to preset conditions, each dredging unit having a regional geomorphology model;

[0108] S107, inputting the dredging units and the first historical water flow parameters into the path planning model one by one to obtain a second planned path for each dredging unit;

[0109] S108, arranging the plurality of second planned paths according to the arrangement order of the dredging units, and merging them with the first planned path to obtain a final planned path;

[0110] S109, controlling the river channel walking device to perform the dredging task according to the final planned path, and collecting the real-time water flow parameters of the current river channel at a preset frequency during the walking process;

[0111] S110, obtaining a water flow threshold of the current position information of the river channel traveling device according to the first historical water flow parameter;

[0112] S111, determining whether the real-time water flow parameter is within the range of the water flow threshold;

[0113] S112, if not, generating an adjustment strategy for the river channel traveling device in a preset area according to the real-time water flow parameters, the adjustment strategy including traveling speed adjustment and path adjustment of the river channel traveling device;

[0114] S113, recording the adjusted path as a first adjustment path, obtaining a partial path corresponding to the current preset area in the final planned path, and recording it as a first regression path;

[0115] S114, calculating a displacement deviation value between the first adjustment path and the first regression path;

[0116] S115, determining whether the displacement deviation value is within a preset deviation threshold value;

[0117] S116, if not, the remaining dredging units that have not yet performed the dredging task are recorded as the units to be dredged, and the dredging units and the real-time water flow parameters are re-input into the path planning model one by one to obtain the feedback planning path, and the feedback planning path is updated to the final planning path;

[0118] Repeat steps S109 to S116 until the dredging task of the entire river channel is completed.

[0119] In step S101, the current river channel has the first river channel information, specifically, the first river channel information includes the first river length information, the first river channel name, the first historical water flow parameter, the first historical dredging parameter, and the first surrounding environment information. Among them, the first river length information is the river length value of the current river channel, the first river channel name is the name of the current river channel, the first historical water flow parameter is the water flow data recorded in the current river channel during the previous multiple surveys, including the water flow rate, water flow direction and water flow change cycle information of each small section, etc. The first historical dredging parameter can be understood as the dredging parameter measured in each previous batch of dredging operations, including the silt removal amount, silt removal time, and the number of dredging equipment invested in each previous batch of dredging, etc.; the first surrounding environment information includes the environmental information within the preset width on both sides of the river channel with the current river channel as the central axis, including terrain, landform, residential area, etc. The preset width can be set according to actual needs. The first river channel information is used to support subsequent path planning and riverbed landform model construction.

[0120] In step S102, the database can be a national hydrological database, or a database module can be constructed separately according to actual needs, and the river channel information of each river in the national area is summarized in the form of network collection or docking with an external database. The database stores the river channel information of each river channel. For the convenience of distinction, the river channel information of the remaining rivers except the current river channel is recorded as the second river channel information, which is consistent with the following description and has the same interpretation. According to the first river channel information, the riverbed geomorphic image of the second river channel information similar to the current first river channel information is screened in the database. It should be noted that the semantic similarity between the first river channel information and the second river channel information is used for screening here, and the riverbed geomorphic image in the second river channel information of the river channel that meets the similarity index is screened, and it is recorded as a reference riverbed image. Multiple reference riverbed images are stored in the form of a collection to obtain a reference image data set. The reference riverbed image obtained by screening in the form of semantic similarity has environmental data similar to the current river channel. It should be noted that the riverbed geomorphic image in the second river channel information can be a partial riverbed geomorphic image of other rivers, or a complete riverbed geomorphic image of the entire river channel.

[0121] In step S103, a regional geomorphic image of the current first river channel information is obtained. It should be noted that the regional geomorphic image can be acquired in real time by a river channel walking device, and the regional geomorphic image can also be a regional geomorphic image stored in a historical survey process.

[0122] In step S104, the regional landform image is similarly matched with the reference riverbed image in the reference image data set. Different from the similarity matching in step S102, the similarity matching in the current step is based on image to image, that is, based on the similarity matching between images to further screen the reference riverbed image, and obtain the reference riverbed image with the highest similarity to the current regional landform image, which is recorded as the reference riverbed image.

[0123] In step S105, a riverbed geomorphology model of the current river channel is constructed based on the reference riverbed image. It should be noted that the riverbed geomorphology model of the current river channel is a complete riverbed geomorphology model of the entire river channel, and a first planned path is generated based on this riverbed geomorphology model. It should be noted that the first planned path can be obtained through the path planning model described later, or it can be directly obtained using feature extraction. The main information of the first planned path is the overall walking direction information of the current river channel, which can also be understood as the winding direction of the current river channel. Specifically, when using the path planning model, the first path planning model can be obtained by modifying the input parameters and constraints. The input parameters can be the winding characteristics of the river channel in the riverbed geomorphology model, and the constraints can be modified to the main water flow constraints of the river channel, so as to obtain a rough path, which is recorded as the first planned path, such as Figure 2 shown.

[0124] In step S106, in the riverbed geomorphology model, the riverbed geomorphology model is divided according to preset conditions to obtain a plurality of dredging units, each of which has a regional geomorphology model, that is, the regional geomorphology model is obtained by dividing the riverbed geomorphology model. Figure 2 As shown, multiple dredging units are spliced ​​together to form a complete river channel.

[0125] In step S107, the dredging unit and the first historical water flow parameter are input into the path planning model to obtain a second planned path corresponding to each dredging unit. Specifically, the path planning model can be an algorithm model generated based on the A path planning algorithm, or an algorithm model generated based on the D* path planning algorithm, or a path planning model obtained by training a deep learning model.

[0126] In step S108, the plurality of second planned paths are arranged in the order of arrangement of the dredging units and merged with the first planned path to obtain a final planned path, such as Figure 2 As shown, after dividing into multiple dredging units, the second planned paths are connected one by one in the direction shown by the first planned path. This can keep the overall path planning trend consistent with the river channel winding trend, and refine the second planned path according to the regional geomorphic model corresponding to each dredging unit.

[0127] In step S109, the river channel walking device is controlled to perform the dredging task according to the final planned path. During this process, the river channel walking device can walk in the river channel along the final planned path, while absorbing silt information on the walking area, and collect real-time water flow parameters of the current river channel at a preset frequency during the walking process as environmental feedback data.

[0128] In step S110, the water flow threshold of the position information of the current river walking device is obtained according to the first historical water flow parameter. The specific process may be: obtaining the first historical water flow velocity in the first historical water flow parameter, constructing a two-dimensional coordinate system, and plotting the first historical water flow velocity in the two-dimensional coordinate system according to the survey time point to obtain a discrete point distribution diagram of the first historical water flow velocity, and performing linear fitting on the first historical water flow velocity, such as using the linear fitting function in Matlab, to obtain the water flow threshold.

[0129] In step S111, it is determined whether the real-time water flow parameters are within the range of the water flow threshold. If so, it means that the current river channel walking device is in normal use during the dredging task, and there is no need to adjust the current walking speed of the river channel walking device, and there is no need to adjust the final planned path.

[0130] In step S112, if the real-time water flow parameters are not within the range of the water flow threshold, it means that the current water flow is abnormal, which will affect the dredging task of the river channel walking device, and then generate an adjustment strategy for the river channel walking device in the preset area based on the actual water flow parameters. The adjustment strategy may specifically include the adjustment of the walking speed of the river channel walking device, or the adjustment strategy may also include the path adjustment of the river channel walking device. It should be noted that the generation method of the adjustment strategy can be obtained through a neural network model, and the neural network model is trained to enable it to have the ability to generate corresponding adjustment strategies according to actual conditions. For example, when the water flow rate is too large, the walking instruction of the current final planned path will cause the river walking device to roll over. In this case, the impact of the water flow resistance caused by the large water flow rate can be reduced by adjusting the path. For another example, in the walking instruction of the current final planned path, the water flow rate is too small, which causes the current river walking device to be unable to complete the preset journey at the current walking speed. In this case, the walking speed can be increased so that the current river walking device can complete the preset journey within the specified time. For another example, when the real-time water flow parameters change due to temporary obstacles in the river, the specific factors may be that when encountering mudslides and landslides that have caused a certain amount of silt accumulation in the current river, resulting in lower real-time water flow parameters, the current abnormal area can be bypassed by adjusting the walking speed and adjusting the final planned path. It should be noted that the preset area can be a pre-determined adjustment range or an adjustment range calculated based on real-time water flow parameters. The adjustment range calculated based on real-time water flow parameters will be more in line with the actual scenario than the pre-selected adjustment range. The specific steps may be: calculate the difference between the first real-time water flow velocity in the real-time water flow parameters and the first historical water flow velocity in the first historical water flow parameters. If the difference is positive, it means that there is a turbulent watershed at present, and the influence range of the turbulent water flow in the current riverbed landform model is predicted by the riverbed landform model, that is, the adjustment range; if the difference is negative, it means that there is a slow-flowing watershed at present, and the influence range of the slow-flowing water flow in the current riverbed landform model is predicted by the riverbed landform model, that is, the adjustment range.

[0131] In step S113, if the path adjustment is involved, the adjusted path in the preset area is recorded as the first adjustment path, and the part of the planned path corresponding to the current preset area in the final planned path is obtained and recorded as the first regression path.

[0132] In step S114, the displacement deviation value between the first adjustment path and the first regression path is calculated. The calculation method of this displacement deviation value can be obtained through Euclidean distance calculation. The specific steps may be: in the first adjustment path, multiple first calculation points are obtained according to the latest walking speed step position of the river channel walking device, and in the first regression path, multiple second calculation points are obtained according to the latest walking speed step position of the river channel walking device, the first calculation point is matched with the second calculation point, and the first calculation point and the second calculation point of the same step point are grouped together, and the Euclidean distance between the first calculation point and the second calculation point is calculated to obtain multiple distance deviation values, that is, displacement deviation values.

[0133] In step S115, it is determined one by one whether each displacement deviation value is within the range of the preset deviation threshold. In some embodiments, the first calculation points within the range of the preset deviation threshold are counted. When the number of first calculation points within the preset deviation threshold accounts for 75% of the total number of first calculation points, it means that the current first adjustment path is not much different from the first regression path or is basically the same, and the river walking device will eventually reintegrate into the path area specified by the final planned path after moving according to the first adjustment path; when it is less than 75%, it means that the first adjustment path is more different from the first regression path. The specific determination percentage can be adjusted according to actual needs, and this embodiment does not limit this.

[0134] In step S116, when the first adjustment path deviates greatly from the first regression path, it means that the current river channel walking device cannot return to the final planned path after walking according to the first adjustment path, and it is necessary to readjust the part of the final planned path that has not yet been traveled. Specifically, the remaining dredging units that have not yet performed the dredging task are recorded as the units to be dredged, and the units to be dredged and the real-time water flow parameters are re-input into the path planning model one by one to obtain the feedback planning path, and the feedback planning path is used to correct and update the final planned path in the form of fusion. The specific fusion step can use the path fusion function, or it can replace the part of the final planned path that has not yet traveled in the form of splicing.

[0135] In this embodiment, by repeating steps S109 to S116, the final planned path can be regularly updated with the help of the change of real-time water flow parameters. It should be noted that the preset frequency can be set according to actual needs, and this embodiment does not limit this.

[0136] See also Figure 2, which is a schematic diagram of the application of the above-mentioned logical steps in the current river channel, divides the river channel 1 into multiple dredging units 11, and takes the water flow direction as the walking direction of the river channel walking device 4. The first planned path 2 is the main walking path of the river channel walking device 4 for performing the dredging task in the river channel 1. The second planned path 3 is generated based on the regional landform model associated with the dredging unit 11, which can perform independent path planning for a single dredging unit 11 in a more detailed manner, so that the river channel walking device 4 can achieve more comprehensive silt removal from the dredging unit 11 during the walking process. By constraining the river channel walking device in a general direction through the first planned path 2, it can be ensured that the river channel walking device 4 can be consistent with the extension direction of the river channel 1 during the execution of the dredging task.

[0137] The method steps shown in this embodiment obtain multiple similar reference riverbed images by performing semantic similarity screening on the first river channel information, and then obtain the reference riverbed image with the highest similarity to the current river channel by matching the similarity between the regional landform image and the reference riverbed image on this basis, and then construct a riverbed landform model of the current river channel based on the reference riverbed image, and obtain the first planned path and the second planned path in sequence based on the constructed riverbed landform model, and finally obtain the final planned path, and on this basis, the final planned path is regularly updated and corrected with the help of environmental feedback data of real-time water flow parameters, so as to realize the improvement and correction of the river channel path during the execution of the dredging task, so as to improve the accuracy of the path planning of the entire river channel walking device, and more accurately control the environment at the bottom of the river channel, so as to make the dredging area more perfect, thereby improving the dredging efficiency.

[0138] Correspondingly, in the second aspect, the present embodiment provides a path planning system for a river traveling device, including a river traveling device, a sensor component, a dredging component and a control unit, the river traveling device including a supporting platform, a driving unit and traveling wheels, the traveling wheels are arranged below the supporting platform, and the driving unit is arranged on the supporting platform; the sensor component is arranged on the supporting platform, and the sensor component includes at least one of a sonar sensor, a laser sensor, a water flow sensor and a silt sensor; the dredging component includes a dredging pipeline, the dredging pipeline is arranged on the supporting platform, and the dredging pipeline is used to extract silt at the bottom of the river channel; the control unit is arranged on the supporting platform, and the control unit is electrically connected to the driving unit, the sensor component and the dredging component, respectively, and the control unit is used to execute the method described in the first aspect.

[0139] In this embodiment, the river channel walking device has a support platform, a drive unit and a walking wheel, wherein the drive unit can be a servo motor, and the support platform is the main structure of the entire river channel walking device, and a dredging component, a sensor component and a control unit can be fixed on the support platform. The walking wheel can be selected from tracks, fan-blade walking wheels and tires according to actual needs, preferably tracks, which can have better gripping ability and the advantage of smooth travel in complex terrain. It can be known that the path planning system shown in this embodiment also includes other detailed components to form a complete whole, that is, the river channel walking device can also be provided with telescopic rods, ball joints, bearings, steering gears, protective covers and other components according to actual needs to make the river channel walking device adapt to the ability to walk in the water flow.

[0140] The sensor component is set on the supporting platform, and the sensor component includes sonar sensor, laser sensor, water flow sensor and silt sensor. Among them, the sonar sensor and laser sensor are used to obtain regional landform images to assist in the construction of riverbed landform model. The water flow sensor can obtain real-time water flow parameters. The silt sensor can be used to detect the depth of silt to facilitate the adjustment of various parameters in the dredging component.

[0141] The dredging assembly includes a dredging pipeline, which is arranged on a supporting platform. The other end of the dredging pipeline can be connected to a sludge storage device according to actual needs. The sludge storage device can be a sludge storage tank, a river bank, and a ship hull. Different sludge storage devices are selected according to actual needs. A crusher can be provided at the input port of the dredging pipeline to crush the sludge, stones, garbage or other obstacles at the bottom of the river. It is known that the dredging assembly is also equipped with a pump to pump the sludge at the bottom into the sludge storage device above.

[0142] The control unit is arranged on the supporting platform. It can be known that the control unit is a general term for electrical components that perform logical operations, which may include multiple chips such as communication modules, operation modules, and image processing modules. The control unit may be a microcomputer chip, MUC, etc., which is obtained by connecting multiple components on a PCB circuit board. This embodiment does not limit this. The control unit can execute the steps described in the first aspect according to the information collected by the sensor component, and then control the river travel device to move along the final planned path, and at the same time control the dredging component to perform dredging during the moving process, and collect real-time water flow parameters in real time with the help of the sensor component to update and correct the final planned path.

[0143] The path planning system of the river walking device shown in this embodiment can be applied to most river dredging operations, and the specific structure can be adjusted and modified accordingly according to different river environments. The path planning system shown in this embodiment realizes the accuracy, intelligence and adaptability of path planning during river dredging through technical means such as multi-dimensional information acquisition, landform image similarity matching, multi-unit path planning and fusion, and real-time dynamic adjustment of water flow parameters. This method can effectively improve dredging efficiency, reduce operation risks, and optimize the overall performance of river dredging operations. It has significant technical innovation value and practical application prospects, accurately matches river geomorphic features; dynamically adjusts dredging paths; monitors water flow parameters in real time; intelligent path planning; and improves the safety and accuracy of dredging operations.

[0144] Further, in some embodiments, other river channels in the database have second river channel information, and the second river channel information includes at least one of second river length information, second river channel name, second historical water flow parameter, second historical dredging parameter, and second surrounding environment information;

[0145] According to the first river channel information, riverbed geomorphic images similar to the current first river channel information are screened in the database to obtain a reference image data set including:

[0146] Constructing a semantic dictionary, converting the first river channel information into a first semantic feature, and converting the second river channel information of other rivers in the database into a second semantic feature;

[0147] The semantic dictionary is expressed by formula (1), which is as follows:

[0148] ;

[0149] In formula (1), is a semantic dictionary, is the characteristic vector representation of river information, is the characteristic vector representation of the river length information, is the feature vector representation of the river name, is the characteristic vector representation of historical water flow parameters, is the characteristic vector representation of historical desilting parameters, is the feature vector representation of the surrounding environment information, for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameter of

[0150] The first semantic feature is described according to the semantic dictionary and is expressed by formula (2). Formula (2) is as follows:

[0151] ;

[0152] In formula (2), is the first semantic feature, is the feature vector representation of the first river channel information, is the feature vector representation of the first river length information, is the feature vector representation of the name of the first river, is the eigenvector representation of the first historical flow parameter, is the eigenvector representation of the first historical desilting parameter, is the feature vector representation of the first surrounding environment information, for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameter of

[0153] The second semantic feature is described according to the semantic dictionary and is expressed by formula (3). Formula (3) is as follows:

[0154] ;

[0155] In formula (3), is the second semantic feature, is the feature vector representation of the second river channel information, is the feature vector representation of the second river length information, is the feature vector representation of the second river name, is the eigenvector representation of the second historical flow parameter, is the eigenvector representation of the second historical desilting parameter, is the feature vector representation of the second surrounding environment information, for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameter of

[0156] The similarity between the second semantic feature and the first semantic feature is calculated one by one and recorded as the semantic similarity value. The similarity is calculated by formula (4), which is as follows:

[0157] ;

[0158] In formula (4), For the The semantic similarity value of the second river information, is the similarity function, For the The second semantic feature of the second river information, , N is a positive integer;

[0159] It is determined whether the semantic similarity values ​​exceed the range of the semantic similarity threshold one by one. If so, the riverbed topography image in the second river channel information associated with the semantic similarity value is screened and stored to obtain a reference image data set, which is expressed by formula (5). Formula (5) is as follows:

[0160] ;

[0161] In formula (5), is the semantic similarity threshold, is the reference image dataset, for Associated riverbed geomorphic image.

[0162] In this embodiment, the channel information of other channels in the database is recorded as the second channel information. Specifically, by constructing a semantic dictionary, the first channel information is converted into a first semantic feature, and the plurality of second channel information are correspondingly converted into second semantic features.

[0163] The semantic dictionary can be understood as a semantic space composed of multiple attribute features. Each attribute feature radiates in the form of a vector. The mapping parameter corresponding to the vector of each attribute feature is the vector length of the current attribute feature. By comparing the vector lengths of multiple attribute features of the two semantic spaces one by one, the similarity between the two semantic spaces can be obtained. Under this principle, the semantic dictionary is represented by formula (1), where the mapping parameter is the vector length value of each attribute feature, and the attribute features are river length information, river name, historical water flow parameters, historical dredging parameters, and surrounding environment information. The first semantic feature is displayed in the form of a semantic dictionary, as represented by formula (2). Correspondingly, the second semantic feature is displayed in the form of a semantic dictionary, as represented by formula (3). When there are multiple second semantic features, in The subscripts correspond to increasing the count letters .

[0164] Then the similarity calculation between the first semantic feature and the second semantic feature can be converted into and The vector calculation of is shown in formula (4). In formula (4), is a similarity function, which can be a cosine similarity function or a Jaccard similarity function. In the specific calculation process, by calculating The attribute characteristics and The similarity of the same attribute features in the second channel information can be calculated by calculating the similarities of multiple attribute features to obtain the semantic similarity value corresponding to each second river channel information.

[0165] Further, a semantic similarity threshold is set. The semantic similarity threshold can be obtained by using big data calculation or can be set manually, and this embodiment does not limit this. Optionally, the semantic similarity threshold can be obtained in the form of manual setting. After the semantic similarity threshold is known, the semantic similarity values ​​are judged one by one, and when the semantic similarity threshold is exceeded, the riverbed landform image associated with the second river channel information is obtained and filtered into the reference image data set to wait for the next operation.

[0166] This embodiment can more effectively capture and describe the multi-dimensional attributes of the river by converting the river information into semantic features, and integrate the various attribute features of the river (such as river length, river name, historical water flow parameters, etc.), so as to form a more comprehensive semantic space of river information, which is helpful for comprehensive analysis of the state and changes of the river. By calculating the similarity between the first river information and the second river information, and by comparing the vector lengths of multiple attribute features one by one, the similarity between the river information can be more accurately reflected. This refined analysis improves the reliability and accuracy of the results; effectively screens out riverbed geomorphological images similar to the current river; this correlation enhances the efficiency of data utilization and helps to better understand the characteristics and changes of the river.

[0167] In some embodiments, the regional geomorphic image is matched with the reference riverbed image in the reference image dataset for similarity, and a reference riverbed image having the highest similarity to the current regional geomorphic image is obtained, which is recorded as the reference riverbed image, including:

[0168] Inputting the regional landform image into the texture recognition model to obtain the texture feature of the regional landform image, recorded as the first texture feature, and inputting the reference riverbed images into the texture recognition model one by one to obtain the texture feature of each reference riverbed image, recorded as the second texture feature;

[0169] Inputting the regional landform image into the edge recognition model to obtain the edge feature of the regional landform image, recorded as the first edge feature, and inputting the reference riverbed images into the edge recognition model one by one to obtain the edge feature of each reference riverbed image, recorded as the second edge feature;

[0170] Obtaining a first image feature vector according to the first texture feature and the first edge feature, and obtaining a second image feature vector according to the second texture feature and the second edge feature;

[0171] The mutual correlation coefficient between the second image feature vector and the first image feature vector is calculated one by one. The mutual correlation coefficient can be expressed by formula (6), which is as follows:

[0172] ;

[0173] In formula (6), is the first image feature vector, For the The second image feature vector of the reference riverbed image, For the The correlation coefficient between the reference riverbed image and the regional topographic image, is the first horizontal coordinate of the pixel point in the regional landform image, is the second horizontal coordinate of the pixel point in the regional landform image, is the height coordinate of the pixel point in the regional topographic image, is the first horizontal offset of the pixel in the reference riverbed image, is the second horizontal offset of the pixel in the reference riverbed image, is the height offset of the pixel in the reference riverbed image;

[0174] The reference riverbed image associated with the second image feature vector having the largest value of the mutual correlation coefficient is selected and recorded as the reference riverbed image.

[0175] In this embodiment, the texture recognition model can be generated based on the grayscale co-occurrence matrix algorithm. Specifically, taking the regional landform image as an example, the grayscale co-occurrence matrix of the regional landform image is calculated, recorded as the first grayscale co-occurrence matrix, the first grayscale co-occurrence matrix is ​​normalized, and then the texture features of the first grayscale co-occurrence matrix are extracted. The texture features include correlation, contrast, energy, homogeneity and entropy. The texture features extracted from the first grayscale co-occurrence matrix are recorded as the first texture features. Similarly, the reference riverbed image is converted into a second grayscale co-occurrence matrix in turn, and normalized, and then the texture features of the second grayscale co-occurrence matrix are extracted to obtain the second texture features corresponding to each second grayscale co-occurrence matrix.

[0176] In this embodiment, the edge recognition model can be generated based on the Canny edge detection algorithm. Taking the regional landform image as an example, Gaussian filtering is used on the regional landform image, and then the gradient amplitude and gradient direction are calculated, and the local maximum gradient point is retained to achieve non-maximum suppression. At the same time, a double threshold detection principle is set. Based on the above steps, the edge feature extraction of the regional landform image can be obtained. The edge features include edge pixel density, total number of edge pixels, edge direction histogram, edge length, etc. The edge features extracted from the regional landform image are recorded as the first edge features; similarly, for the reference riverbed image, Gaussian filtering and gradient calculation are used in turn, and double threshold detection is performed to finally obtain the edge features of the reference riverbed image, which are recorded as the second edge features.

[0177] A first image feature vector is obtained according to the first texture feature and the first edge feature, and a second image feature vector is obtained according to the second texture feature and the second edge feature; the generation of the first image feature vector can be achieved by fusing the first texture feature and the first edge feature through feature splicing, feature weighted fusion, PCA fusion, and a fusion method based on a neural network model. Optionally, taking feature splicing as an example, the first texture feature and the first edge feature can be directly spliced ​​to form a first image feature vector. In some preferred embodiments, the PCA analysis method can be used to retain and fuse the key feature information in the first image feature and the first edge feature to reduce the computational redundancy in the subsequent cross-correlation coefficient calculation. Similarly, the generation of the second image feature vector can be achieved by fusing the second texture feature and the second edge feature through feature splicing, feature weighted fusion, PCA fusion, and a fusion method based on a neural network model. It can be understood that the generation method of the first image feature vector is preferably consistent with the generation method of the second image feature vector, so that the subsequent cross-correlation coefficient calculation process is more accurate.

[0178] In this embodiment, the mutual correlation coefficient is used to measure the similarity between the first image feature vector and the second image feature vector. Specifically, the first image feature vector is defined as , define the second image feature vector as ;in, and It represents the coordinates in two directions on the horizontal plane. is the vertical coordinate.

[0179] In some preferred embodiments, the first image feature vector can be represented in the form of a first matrix, wherein the first matrix has a plurality of first image attribute features, and the second image feature vector can be represented in the form of a second matrix, wherein the second matrix has a plurality of second image attribute features. The mutual correlation coefficients of the first image attribute features and the second image attribute features are calculated in sequence, and the calculated mutual correlation coefficients are averaged to finally obtain the final mutual correlation coefficient corresponding to the second matrix. This process is repeated until all second image feature vectors are traversed.

[0180] The final mutual correlation coefficients are arranged in order to obtain a reference riverbed image associated with the second image feature vector having the largest final mutual correlation coefficient.

[0181] This embodiment realizes multi-dimensional feature extraction of regional landform images and reference riverbed images by comprehensively using texture recognition models and edge recognition models, greatly improving the accuracy and reliability of image matching. The fusion of texture features and edge features not only retains the detailed information of the image, but also fully reflects the structural characteristics of the image. The use of mutual correlation coefficients as a similarity measurement method can accurately calculate the degree of correlation between image feature vectors, effectively overcoming the limitations of traditional single similarity calculation methods. By comparing and calculating one by one, the reference riverbed image that is most similar to the current regional landform image can be accurately screened out, which is convenient for improving the accuracy of subsequent riverbed landform model construction.

[0182] In some embodiments, recording the second image feature vector of the reference riverbed image as the reference image feature vector, and constructing the riverbed geomorphology model of the current river channel according to the reference riverbed image includes:

[0183] The first image feature vector is normalized to obtain a first normalized feature vector. The first normalized feature vector is expressed by formula (7). Formula (7) is as follows:

[0184] ;

[0185] In formula (7), is the first normalized eigenvector, is the first lower limit, is the first upper limit value;

[0186] The first standardized feature vector and the reference image feature vector are used to construct a feature association matrix, which is expressed by formula (8). Formula (8) is as follows:

[0187] ;

[0188] In formula (8), is the feature correlation matrix, is the first normalized eigenvector, is the reference image feature vector, for The variance of for The variance of for and The covariance of for and The covariance of and same;

[0189] Perform principal component analysis based on the feature correlation matrix to obtain the first difference eigenvector;

[0190] According to the first difference eigenvector, a riverbed trend change function is constructed in the DEM algorithm model. The riverbed trend change function is expressed by formula (9), which is as follows:

[0191] ;

[0192] In formula (9), is the riverbed trend change function, For time, is the first difference eigenvector, is the model parameter value corresponding to the first difference eigenvector, is the feature influence coefficient corresponding to the first difference eigenvector, is the model parameter value of the characteristic influence coefficient, is the first surrounding environment information, is the model parameter value of the first surrounding environment information;

[0193] The preset landform model in the DEM algorithm model is updated according to the riverbed trend change function to generate the riverbed landform model corresponding to the current river channel. The riverbed landform model is expressed by formula (10), which is as follows:

[0194] ;

[0195] In formula (10), is the riverbed geomorphic model of the current river channel, It is the preset landform model in the DEM algorithm model.

[0196] In this embodiment, the second image feature vector corresponding to the reference riverbed image is recorded as the reference image feature vector. Furthermore, the DEM algorithm model is used to construct the riverbed landform model. The first image feature vector is standardized to remove missing values ​​or outliers in the first image feature vector, and a feature association matrix is ​​constructed. The feature association matrix can be a covariance matrix. The difference value between each first image attribute feature and the second image attribute feature is calculated to refine the distinguishing features between the reference riverbed image and the regional landform image, and then the distinguishing features are subjected to principal component analysis to further optimize the number of distinguishing features, reduce redundancy, and obtain the first difference feature vector.

[0197] According to the first difference feature vector, a riverbed trend change function can be obtained. The riverbed trend change function can be understood as the change trend of the difference feature between the current river channel and the reference riverbed image over time. It should be noted that the model parameter values ​​in formula (9) can be set manually or the estimated values ​​in the DEM algorithm model can be used, and this embodiment does not limit this.

[0198] In the DEM algorithm model, the riverbed trend change function is spliced ​​with the preset landform model to generate the riverbed landform model after the current riverbed changes with time. You can choose according to actual needs. For example, one feasible way is to obtain the generation timestamp of the reference riverbed image, calculate the time interval between the current real-time timestamp and the generation timestamp of the reference riverbed image, and use this time interval as Value input is used to obtain the riverbed geomorphological model of the current river channel in its current state.

[0199] This embodiment effectively removes missing values ​​and outliers in the feature data of the first image feature vector by standardizing the first image feature vector, thereby improving the accuracy and reliability of subsequent analysis. This standardization process ensures that the feature vectors are compared on the same scale and enhances the comparability between different image features. Constructing a feature association matrix and performing principal component analysis can deeply explore the relationship between the first image feature and the reference image feature and identify the main difference features. This process not only simplifies the dimension of the feature and reduces redundant information, but also provides an important basis for the subsequent construction of the riverbed trend change function. Through the construction of the riverbed trend change function, the changing trend of the regional river channel during the time evolution process can be dynamically reflected; this method enables the riverbed geomorphic model to be updated in time to reflect the current geomorphic state, and has strong adaptability. The construction process of the riverbed geomorphic model shown in this embodiment has good practicality and promotion value, and can provide reliable technical support for river management and ecological protection.

[0200] In some embodiments, the preset condition is a dredging rate matching coefficient, and dividing a plurality of dredging units according to the preset condition along the extension direction of the current river channel in the riverbed topography model includes:

[0201] Obtaining first historical dredging parameters, the first historical dredging parameters including the first historical total amount of silt, the first historical dredging cycle, and the first historical daily dredging mileage corresponding to each batch of dredging tasks in the current river channel;

[0202] Generate a first historical dredging rate according to the first historical total silt volume, the first historical dredging cycle, and the first historical daily dredging mileage;

[0203] Obtain the standard dredging rate of the current river travel device and the initial dredging unit length;

[0204] A rate matching coefficient is obtained according to the standard dredging rate and the first historical dredging rate;

[0205] The desilting unit division length is calculated based on the rate matching coefficient and the initial desilting unit length;

[0206] The predicted silt volume and predicted dredging time of each dredging unit are calculated based on the length of the dredging unit and are organized into a list for display.

[0207] In this embodiment, the rate matching coefficient can be obtained by the ratio of the standard dredging rate to the first historical dredging rate. The standard dredging rate may vary depending on the equipment model of the dredging component on the river traveling device. The standard dredging rates of the dredging components on different river traveling devices may be different.

[0208] This embodiment obtains and analyzes the first historical dredging parameters, including the total amount of silt, dredging cycle and daily dredging mileage, to establish a benchmark for the historical dredging rate. Combined with the standard dredging rate of the current river running device, the length of the dredging unit can be dynamically adjusted, thereby achieving a more flexible division of dredging tasks, making the division of dredging units more scientific and reasonable, and maximizing the efficiency of dredging operations. The amount of silt and dredging time of each dredging unit are predicted, and the results are organized into a list, which is convenient for on-site management personnel to carry out effective scheduling and resource allocation, reducing uncertainty in the operation process. The differences between different dredging components and equipment models are taken into account, making the calculation of the rate matching coefficient more accurate, thereby improving the adaptability and actual effect of the entire dredging operation.

[0209] In some embodiments, the first historical water flow parameter includes a first historical flow velocity distribution, a first historical water flow direction, a first historical water flow velocity, and a first seasonal variation parameter. The dredging units and the first historical water flow parameters are input into the path planning model one by one, and the second planned path for each dredging unit is obtained, including:

[0210] For each desilting unit, perform the following steps:

[0211] The regional geomorphic model corresponding to the current dredging unit is converted into a cost matrix, which includes a travel cost and a movement cost. The travel cost is the terrain complexity of the current regional geomorphic model, and the movement cost is the first historical flow velocity distribution. The first historical water flow direction is used as the path selection weight.

[0212] and, generating a terrain constraint condition based on the regional geomorphic model and generating a water flow constraint condition based on the first historical water flow parameter;

[0213] The cost matrix, path selection weight, terrain constraints and water flow constraints are used as input parameters of the path planning model to obtain the optimal path of the current dredging unit, which is also the second planning path;

[0214] Arrange multiple second planned paths according to the arrangement order of the dredging units, and merge them with the first planned path to obtain the final planned path expressed by formula (11). Formula (11) is as follows:

[0215] ;

[0216] In formula (11), For the final planning path after fusion, For the first planning path, For the The second planning path, is the number of the second planned paths, is the fusion coefficient, which is obtained by fitting the path complexity and the degree of local change of the path.

[0217] In this embodiment, the first historical flow velocity distribution can be understood as the distribution schematic data of different water flow velocities in the current river channel, the first historical water flow direction can be understood as the water flow direction of multiple areas in the current river channel, the first historical water flow velocity can differ according to the first historical flow velocity distribution, and the first seasonal change parameter can be understood as the adjustment weight of the first historical water flow velocity that changes with the season. For example, the current river channel has a flood season and a dry season. During the dry season, the water volume decreases and the water flow decreases. During the flood season, the water volume increases and the water flow increases, and the water flow velocity will show dynamic changes.

[0218] Inputting the divided dredging units and the first historical water flow parameters into the path planning model can be understood as inputting the regional geomorphic model corresponding to the dredging units and the data of the first historical water flow parameters in this regional geomorphic model into the path planning model, thereby obtaining the second planned path.

[0219] In this embodiment, the path planning model used can be the D* path planning algorithm. The regional geomorphic model corresponding to the current dredging unit is converted into a cost matrix, the terrain complexity of the current regional geomorphic model is used as the travel cost, and the first historical flow velocity distribution is used as the movement cost. It should be noted that the terrain complexity can be obtained by measuring in the regional geomorphic model, and specifically can be obtained by calculating the slope, curvature, and undulation in the current regional geomorphic model. The first historical water flow direction is used as the path selection weight. At the same time, the regional geomorphic model generates terrain constraints, and the first historical water flow parameters generate water flow constraints.

[0220] After obtaining multiple constraints and the corresponding cost matrix and path selection weights, they are input into the path planning model to obtain the optimal path to the current dredging unit, which is the second planned path.

[0221] Furthermore, the plurality of second planned paths are arranged according to the arrangement order of the dredging units, and are merged with the first planned path to obtain a final planned path.

[0222] In formula (11), the fusion coefficient needs to be calculated according to the degree of local change of the path and the complexity of the path. Specifically, the complexity of the path can be understood as the degree of tortuosity and frequency of change of the path. Further refinement can be used to quantify the path complexity by the curvature, turning angle and number of path segments. For example, the curvature quantification process can be: calculating the radius of the arc formed by three consecutive points on the path, analyzing the angle change between adjacent path segments, and counting the average curvature per unit path length. The quantification of the turning angle can be: calculating the angle between adjacent path segments, counting the distribution of turning angles, and calculating the standard deviation of turning angles. The quantification of the number of path segments can be: counting the number of path turning points, calculating the number of path segments, and analyzing the length distribution of each path segment; the quantification of the path smoothness can be: calculating the change rate of the angle between adjacent path segments, analyzing the continuity index of the path, and counting the smoothness coefficient of the path. The quantification method of the path direction entropy can be: counting the probability distribution of the path direction, calculating the information entropy of the direction distribution, and analyzing the randomness of the path direction. The change rate of the path length can be: calculating the ratio of the lengths of adjacent path segments, counting the standard deviation of the length change, and analyzing the discrete degree of the path length distribution. After calculating the aforementioned one or more quantitative features, the results of each quantitative feature may be weighted averaged to obtain a final path complexity coefficient.

[0223] Furthermore, the local change degree of the path can be used to describe the change of the path in the local area, for example, the height difference of the path, the change of the width of the path, etc. The height difference change of the path can be obtained by Euclidean distance calculation, and the width change can be obtained by sampling. Based on the height difference of the path and the change of the width of the path, the quantitative value of the local change degree of the path can be obtained, which is recorded as the local change degree coefficient of the path.

[0224] Preferably, a regression analysis method is used to fit the path complexity coefficient, the path local change degree coefficient and the fusion coefficient, and the specific value of the current fusion coefficient is obtained by solving the regression model.

[0225] This embodiment significantly improves the path planning accuracy and efficiency of river dredging operations by comprehensively analyzing the first historical water flow parameters, including the first historical flow velocity distribution, the first historical water flow direction, the first historical water flow velocity, and the first historical seasonal variation parameters. By inputting the dredging units and the first historical water flow parameters into the path planning model one by one, the D* path planning algorithm is used to generate the optimal path for each dredging unit, ensuring efficient operation under complex terrain and dynamic water flow conditions. The construction of the cost matrix takes into account the complexity of the terrain and the flow velocity distribution, optimizes the path selection, and enables the dredging operation to better adapt to the actual situation of the river. In addition, the detailed quantification of the path complexity and the degree of local variation provides a scientific basis for the smoothness and feasibility of the path, ensuring the safety and effectiveness of the path. The final planning path after fusion makes full use of the advantages of the first planning path and the second planning path, and realizes the optimal resource allocation and operation scheduling. This method not only improves the flexibility and adaptability of the dredging task, but also effectively reduces the uncertainty in the execution process, optimizes resource utilization, and improves the overall dredging efficiency.

[0226] In some embodiments, the walking speed adjustment of the river walking device includes a speed adjustment value, and the speed adjustment value is obtained by the following steps:

[0227] Obtain the basic walking speed and real-time walking direction of the river walking device;

[0228] Obtaining real-time water flow parameters, including real-time water flow direction and real-time water flow velocity;

[0229] Generate a first influence coefficient according to the real-time walking direction and the real-time water flow direction;

[0230] Generate a second influence coefficient according to the real-time water flow velocity and the basic walking speed;

[0231] Obtaining a speed adjustment value according to the first influence coefficient, the second influence coefficient and the basic walking speed;

[0232] The first regulation path is obtained by the following steps:

[0233] The real-time water flow parameters and the current dredging unit are input into the path planning model to obtain the first adjustment path.

[0234] In this embodiment, the walking speed adjustment of the river walking device is achieved through a series of precise steps to ensure the safety and efficiency of the dredging operation. First, the basic walking speed of the river walking device is obtained, and this basic walking speed is determined according to the design of the device. At the same time, the real-time walking direction is also collected, which helps to evaluate the actual movement of the device in the river. Real-time monitoring of water flow parameters, including real-time water flow direction and real-time water flow velocity, by comparing the real-time walking direction with the real-time water flow direction, a first influence coefficient is generated, which reflects the degree of influence of the water flow on the walking direction. The comparison of the real-time water flow velocity and the basic walking speed generates a second influence coefficient to evaluate the influence of the current water flow speed on the actual walking speed of the device. The rate adjustment value is finally calculated, which will be used as the basis for the dynamic adjustment of the device during the dredging operation. After obtaining the rate adjustment value, the real-time water flow parameters and the regional landform model corresponding to the current dredging unit are input into the path planning model together to determine the first adjustment path. It can be understood that the generation process and generation method of the first adjustment path can be further understood by referring to the generation process of the second planning path.

[0235] This embodiment significantly improves the safety and efficiency of river dredging operations through precise walking speed adjustment and path planning. Real-time monitoring and analysis of water flow parameters enable the device to flexibly adjust the walking speed according to the actual water flow conditions to avoid safety hazards under strong water flow conditions. Through dynamically generated speed adjustment values, the device can achieve higher stability in different operating environments. Combined with the geomorphic characteristics of the dredging unit and water flow parameters for path planning, the generated first adjustment path is more scientific and reasonable, which maximizes the dredging efficiency and resource utilization.

[0236] In a third aspect, this embodiment further provides a computer-readable storage medium on which computer program instructions are stored. When the computer program instructions are executed by a processor, the method described in the first aspect is implemented.

[0237] In a fourth aspect, this embodiment further provides an electronic device, comprising a memory and a processor, wherein the memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method described in the first aspect.

[0238] The storage medium / memory includes but is not limited to: RAM, ROM, disk, tape, CD, flash memory, USB flash drive, mobile hard disk, memory card, memory stick, network server storage, network cloud storage, etc. The processor includes but is not limited to CPU (central processing unit), GPU (graphics processing unit), MCU (microprocessor), etc.

[0239] Different from the existing technology, the above technical solution realizes the accuracy, intelligence and adaptability of path planning in the process of river dredging through technical means such as multi-dimensional information acquisition, similarity matching of landform images, multi-unit path planning and fusion, and real-time dynamic adjustment of water flow parameters. This method can effectively improve the efficiency of dredging, reduce operational risks, and optimize the overall performance of river dredging operations. It has significant technical innovation value and practical application prospects, accurately matches river geomorphic characteristics; dynamically adjusts dredging paths; monitors water flow parameters in real time; intelligently plans paths; and improves the safety and accuracy of dredging operations.

[0240] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0241] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) or a processor (processor) to perform all or part of the steps of each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0242] The above descriptions are only some embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any equivalent device or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A path planning method for a river traveling device, characterized in that: Applicable to a river channel traveling device, the river channel traveling device is used to perform the dredging task of the current river channel, and the method includes: S101, obtaining first river channel information of a current river channel, wherein the first river channel information includes at least one of first river length information, first river channel name, first historical water flow parameter, first historical dredging parameter, and first surrounding environment information; S102, selecting riverbed geomorphic images of rivers similar to the current first river channel information in a database according to the first river channel information to obtain a reference image data set, wherein the reference image data set includes a reference riverbed image; S103, obtaining a regional geomorphic image of the current first river channel information, wherein the regional geomorphic image is a partial geomorphic image of the riverbed of the current river channel; S104, performing similarity matching between the regional geomorphic image and the reference riverbed image in the reference image dataset, and obtaining a reference riverbed image having the highest similarity to the current regional geomorphic image, which is recorded as the reference riverbed image; S105, constructing a riverbed geomorphology model of the current river channel according to the reference riverbed image, and generating a first planned path of the current river channel walking device according to the riverbed geomorphology model, wherein the first planned path is the overall walking direction information of the river channel walking device in the current river channel; S106, dividing the riverbed landform model into a plurality of dredging units along the extension direction of the current river channel according to preset conditions, each dredging unit having a regional landform model; S107, inputting the dredging units and the first historical water flow parameters into the path planning model one by one to obtain a second planned path for each dredging unit; S108, arranging the plurality of second planned paths according to the arrangement order of the dredging units, and merging them with the first planned path to obtain a final planned path; S109, controlling the river channel walking device to perform the dredging task according to the final planned path, and collecting real-time water flow parameters of the current river channel at a preset frequency during the walking process; S110, obtaining a water flow threshold of the current position information of the river channel traveling device according to the first historical water flow parameter; S111, determining whether the real-time water flow parameter is within the range of the water flow threshold; S112, if not, generating an adjustment strategy for the river channel traveling device in a preset area according to the real-time water flow parameters, the adjustment strategy including traveling speed adjustment and path adjustment of the river channel traveling device; S113, recording the adjusted path as a first adjustment path, obtaining a partial path corresponding to the current preset area in the final planned path, and recording it as a first regression path; S114, calculating a displacement deviation value between the first adjustment path and the first regression path; S115, determining whether the displacement deviation value is within a preset deviation threshold value; S116, if not, the remaining dredging units that have not yet performed the dredging task are recorded as the units to be dredged, and the units to be dredged and the real-time water flow parameters are re-input into the path planning model one by one to obtain a feedback planning path, and the feedback planning path is updated to the final planning path; Repeat steps S109 to S116 until the dredging task of the entire river channel is completed.

2. The path planning method for a river running device according to claim 1, characterized in that: The other rivers in the database have second river channel information, and the second river channel information includes at least one of second river length information, second river channel name, second historical water flow parameters, second historical dredging parameters, and second surrounding environment information; According to the first river channel information, riverbed geomorphic images similar to the current first river channel information are screened in the database to obtain a reference image data set including: Constructing a semantic dictionary, converting the first river channel information into a first semantic feature, and converting the second river channel information of other rivers in the database into a second semantic feature; The semantic dictionary is represented by formula (1), which is as follows: ； In formula (1), is a semantic dictionary, is the feature vector representation of river information, is the characteristic vector representation of the river length information, is the feature vector representation of the river name, is the characteristic vector representation of historical water flow parameters, is the characteristic vector representation of historical desilting parameters, is the feature vector representation of the surrounding environment information, for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameter of The first semantic feature is described according to the semantic dictionary, and the first semantic feature is expressed by formula (2), and the formula (2) is as follows: ； In formula (2), is the first semantic feature, is the feature vector representation of the first river channel information, is the feature vector representation of the first river length information, is the feature vector representation of the name of the first river, is the eigenvector representation of the first historical flow parameter, is the eigenvector representation of the first historical desilting parameter, is the feature vector representation of the first surrounding environment information, for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameter of The second semantic feature is described according to the semantic dictionary, and the second semantic feature is expressed by formula (3), and the formula (3) is as follows: ； In formula (3), is the second semantic feature, is the feature vector representation of the second river channel information, is the feature vector representation of the second river length information, is the feature vector representation of the second river name, is the eigenvector representation of the second historical flow parameter, is the eigenvector representation of the second historical desilting parameter, is the feature vector representation of the second surrounding environment information, for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameters of for The mapping parameter of The similarity between the second semantic feature and the first semantic feature is calculated one by one and recorded as a semantic similarity value. The similarity is calculated by formula (4), and the formula (4) is as follows: ； In formula (4), For the The semantic similarity value of the second river information, is the similarity function, For the The second semantic feature of the second river information, , N is a positive integer; It is determined whether the semantic similarity values ​​exceed the range of the semantic similarity threshold value one by one. If so, the riverbed topography image in the second river channel information associated with the semantic similarity value is screened and stored to obtain the reference image data set, which is expressed by formula (5). Formula (5) is as follows: ； In formula (5), is the semantic similarity threshold, is the reference image dataset, for Associated riverbed geomorphic image.

3. The path planning method for a river running device according to claim 1, characterized in that: The regional geomorphic image is matched with the reference riverbed image in the reference image dataset to obtain a reference riverbed image with the highest similarity to the current regional geomorphic image, which is recorded as the reference riverbed image. Inputting the regional landform image into a texture recognition model to obtain a texture feature of the regional landform image, recorded as a first texture feature, and inputting the reference riverbed images into the texture recognition model one by one to obtain a texture feature of each reference riverbed image, recorded as a second texture feature; Inputting the regional landform image into the edge recognition model to obtain edge features of the regional landform image, recorded as first edge features, and inputting the reference riverbed images into the edge recognition model one by one to obtain edge features of each reference riverbed image, recorded as second edge features; Obtaining a first image feature vector according to the first texture feature and the first edge feature, and obtaining a second image feature vector according to the second texture feature and the second edge feature; The mutual correlation coefficient between the second image feature vector and the first image feature vector is calculated one by one. The mutual correlation coefficient can be expressed by formula (6), which is as follows: ; In formula (6), is the first image feature vector, For the The second image feature vector of the reference riverbed image, For the The correlation coefficient between the reference riverbed image and the regional topographic image, is the first horizontal coordinate of the pixel point in the regional landform image, is the second horizontal coordinate of the pixel point in the regional landform image, is the height coordinate of the pixel point in the regional topographic image, is the first horizontal offset of the pixel in the reference riverbed image, is the second horizontal offset of the pixel in the reference riverbed image, is the height offset of the pixel in the reference riverbed image; The reference riverbed image associated with the second image feature vector having the largest value of the mutual correlation coefficient is selected and recorded as the reference riverbed image.

4. The path planning method for a river running device according to claim 3, characterized in that: Recording the second image feature vector of the reference riverbed image as the reference image feature vector, and constructing a riverbed geomorphic model of the current river channel according to the reference riverbed image includes: The first image feature vector is normalized to obtain a first normalized feature vector, and the first normalized feature vector is expressed by formula (7), which is as follows: ； In formula (7), is the first normalized eigenvector, is the first lower limit, is the first upper limit value; The first standardized feature vector and the reference image feature vector are used to construct a feature association matrix, and the feature association matrix is ​​expressed by formula (8), which is as follows: ； In formula (8), is the feature correlation matrix, is the first normalized eigenvector, is the reference image feature vector, for The variance of for The variance of for and The covariance of for and The covariance of and same; Performing principal component analysis according to the feature association matrix to obtain a first difference feature vector; The riverbed trend change function is constructed in the DEM algorithm model according to the first difference feature vector. The riverbed trend change function is expressed by formula (9), which is as follows: ； In formula (9), is the riverbed trend change function, For time, is the first difference eigenvector, is the model parameter value corresponding to the first difference eigenvector, is the feature influence coefficient corresponding to the first difference feature vector, is the model parameter value of the characteristic influence coefficient, is the first surrounding environment information, is the model parameter value of the first surrounding environment information; The preset landform model in the DEM algorithm model is updated according to the riverbed trend change function to generate the riverbed landform model corresponding to the current river channel. The riverbed landform model is expressed by formula (10), which is as follows: ； In formula (10), is the riverbed geomorphic model of the current river channel, It is the preset landform model in the DEM algorithm model.

5. The path planning method for a river running device according to claim 1, characterized in that: The preset condition is a dredging rate matching coefficient. In the riverbed topography model, a plurality of dredging units are divided according to the preset condition along the extension direction of the current river channel, including: Acquire first historical dredging parameters, where the first historical dredging parameters include a first historical total amount of silt, a first historical dredging cycle, and a first historical daily dredging mileage corresponding to each batch of dredging tasks in the current river channel; Generate a first historical dredging rate according to the first historical total sludge volume, the first historical dredging cycle, and the first historical daily dredging mileage; Obtain the standard dredging rate of the current river travel device and the initial dredging unit length; Obtaining a rate matching coefficient according to the standard dredging rate and the first historical dredging rate; The desilting unit division length is calculated according to the rate matching coefficient and the initial desilting unit length; The predicted silt volume and predicted silt removal duration of each silt removal unit are calculated based on the length of the silt removal unit divisions and are sorted into a list for display.

6. The path planning method for a river running device according to claim 1, characterized in that: The first historical water flow parameters include a first historical flow velocity distribution, a first historical water flow direction, a first historical water flow velocity, and a first seasonal variation parameter. The dredging units and the first historical water flow parameters are input into the path planning model one by one, and the second planning path of each dredging unit is obtained, which includes: The following steps are performed for each of the desilting units: The regional geomorphic model corresponding to the current dredging unit is converted into a cost matrix, wherein the cost matrix includes a passage cost and a movement cost, wherein the passage cost is the terrain complexity of the current regional geomorphic model, the movement cost is the first historical flow velocity distribution, and the first historical water flow direction is used as a path selection weight; and generating terrain constraints according to the regional geomorphic model and generating water flow constraints according to the first historical water flow parameters; The cost matrix, path selection weight, terrain constraint condition and water flow constraint condition are used as input parameters of the path planning model to obtain the optimal path of the current dredging unit, that is, the second planning path; Arrange the plurality of second planned paths according to the arrangement order of the dredging units, and merge them with the first planned path to obtain a final planned path represented by formula (11), which is as follows: ； In formula (11), The final planning path after fusion. For the first planning path, For the The second planning path, is the number of the second planned paths, is the fusion coefficient, which is obtained by fitting the path complexity and the local change degree of the path.

7. The path planning method for a river running device according to claim 1, characterized in that: The walking speed adjustment of the river walking device includes a speed adjustment value, and the speed adjustment value is obtained by the following steps: Obtaining the basic walking speed and real-time walking direction of the river channel walking device; Acquire real-time water flow parameters, wherein the real-time water flow parameters include real-time water flow direction and real-time water flow velocity; Generate a first influence coefficient according to the real-time walking direction and the real-time water flow direction; Generate a second influence coefficient according to the real-time water flow velocity and the basic walking speed; Obtaining a speed adjustment value according to the first influence coefficient, the second influence coefficient and the basic walking speed; The first regulation path is obtained by the following steps: The real-time water flow parameters and the current dredging unit are input into a path planning model to obtain the first adjustment path.

8. A path planning system for a river travel device, characterized in that: include: A river channel walking device comprises a supporting platform, a driving unit and walking wheels, wherein the walking wheels are arranged below the supporting platform, and the driving unit is arranged on the supporting platform; A sensor assembly is arranged on the support platform, and the sensor assembly includes at least one of a sonar sensor, a laser sensor, a water flow sensor, and a silt sensor; A dredging assembly, comprising a dredging pipeline, wherein the dredging pipeline is arranged on the supporting platform and is used to extract silt from the bottom of the river channel; A control unit is arranged on the support platform, and the control unit is electrically connected to the driving unit, the sensor component and the dredging component respectively, and the control unit is used to execute the method described in any one of claims 1 to 7.

9. A computer-readable storage medium storing computer program instructions, characterized in that: The computer program instructions implement the method according to any one of claims 1 to 7 when executed by a processor.

10. An electronic device comprising a memory and a processor, characterized in that: The memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method according to any one of claims 1 to 7.

Images