CN116108601A - Power cable depth geometric information supplementing method, detector, equipment and medium - Google Patents
Power cable depth geometric information supplementing method, detector, equipment and medium Download PDFInfo
- Publication number
- CN116108601A CN116108601A CN202310147150.4A CN202310147150A CN116108601A CN 116108601 A CN116108601 A CN 116108601A CN 202310147150 A CN202310147150 A CN 202310147150A CN 116108601 A CN116108601 A CN 116108601A
- Authority
- CN
- China
- Prior art keywords
- information
- point cloud
- data
- power cable
- cloud data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 57
- 230000001502 supplementing effect Effects 0.000 title claims abstract description 33
- 239000011159 matrix material Substances 0.000 claims description 35
- 230000006870 function Effects 0.000 claims description 31
- 230000000295 complement effect Effects 0.000 claims description 16
- 238000012549 training Methods 0.000 claims description 16
- 239000013598 vector Substances 0.000 claims description 10
- 238000012545 processing Methods 0.000 claims description 7
- 238000011176 pooling Methods 0.000 claims description 6
- 238000006243 chemical reaction Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 230000004913 activation Effects 0.000 claims description 3
- 230000009469 supplementation Effects 0.000 claims description 2
- 238000004378 air conditioning Methods 0.000 claims 1
- 238000003384 imaging method Methods 0.000 abstract description 7
- 230000000694 effects Effects 0.000 abstract description 5
- 238000004590 computer program Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 3
- 239000000523 sample Substances 0.000 description 3
- 238000013459 approach Methods 0.000 description 2
- 238000003169 complementation method Methods 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 239000004973 liquid crystal related substance Substances 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 235000019800 disodium phosphate Nutrition 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000013074 reference sample Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 239000011800 void material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/10—Geometric CAD
- G06F30/18—Network design, e.g. design based on topological or interconnect aspects of utility systems, piping, heating ventilation air conditioning [HVAC] or cabling
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/06—Energy or water supply
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/005—General purpose rendering architectures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y04—INFORMATION OR COMMUNICATION TECHNOLOGIES HAVING AN IMPACT ON OTHER TECHNOLOGY AREAS
- Y04S—SYSTEMS INTEGRATING TECHNOLOGIES RELATED TO POWER NETWORK OPERATION, COMMUNICATION OR INFORMATION TECHNOLOGIES FOR IMPROVING THE ELECTRICAL POWER GENERATION, TRANSMISSION, DISTRIBUTION, MANAGEMENT OR USAGE, i.e. SMART GRIDS
- Y04S10/00—Systems supporting electrical power generation, transmission or distribution
- Y04S10/50—Systems or methods supporting the power network operation or management, involving a certain degree of interaction with the load-side end user applications
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- Economics (AREA)
- Health & Medical Sciences (AREA)
- Geometry (AREA)
- Computer Graphics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Pure & Applied Mathematics (AREA)
- Multimedia (AREA)
- Mathematical Optimization (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Computer Hardware Design (AREA)
- Public Health (AREA)
- Water Supply & Treatment (AREA)
- General Health & Medical Sciences (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Primary Health Care (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- General Business, Economics & Management (AREA)
- Image Processing (AREA)
Abstract
The embodiment of the application discloses a method, a detector, equipment and a medium for supplementing depth geometric information of a power cable, wherein the method comprises the following steps: acquiring point cloud data of a power cable to be complemented as data to be processed, and acquiring a pre-trained information complementing network; inputting the data to be processed into a coding layer of the information complementing network for characteristic coding to obtain coding characteristics; and inputting the coding features to a decoding layer of the information complementing network for feature decoding to obtain complementing data after the information complementing of the data to be processed. According to the method, point cloud data are used as input of a constructed information complementing network, geometric information with complete shapes is saved, the operation efficiency of a memory is improved, the point cloud data are encoded and decoded through the information complementing network, dense and complete point cloud is generated in a thick-to-thin mode, the effect of complementing depth information of an electric power cable is good, and the situation of incomplete imaging of the indoor and outdoor cables can be dealt with.
Description
Technical Field
The application relates to the technical field of machine vision, in particular to a method, a detector, equipment and a medium for supplementing depth geometric information of a power cable.
Background
In three-dimensional vision, the depth map records geometric information of a scene, is used for representing the distance between an object in the scene and the optical center of a camera, and can further form point cloud data. The advantages and disadvantages of the depth map have great influence on the 3D recognition, positioning and other works.
When power operation is performed, the object of the depth imaging is a black (light absorbing) cylindrical (reflecting) power cable, and this object has a great challenge to the depth camera, especially in an outdoor complex light environment, the depth of the power cable is obtained only depending on the hardware characteristics of the camera itself, and in many cases, broken depth data can be obtained, so it is particularly important to complement the hole break of the depth map.
Currently, the depth complement method is roughly divided into three types: geometry-based, alignment-based, and learning-based depth complement approaches. Wherein the geometry-based approach refers to complementing the shape without any external data using geometric cues from local inputs, which requires that the geometric information of the missing region in a moderately complete input is directly derivable from the observed region, but which assumption is not applicable to incomplete data acquired by the power cable; the alignment-based method is based on template shape from a large shape database, and the shape is complemented by matching local input, so that the method requires higher optimized resources in practical application and is sensitive to noise; the learning-based method can provide quick reasoning and better generalization by mapping the local input into a complete shape, but the current learning-based method uses voxels to represent the shape, which is more conveniently applied to a convolutional neural network, but also obscures the structural characteristics and geometric information of the point cloud itself.
Disclosure of Invention
The embodiment of the application provides a method, a detector, equipment and a medium for supplementing depth geometric information of a power cable, which can realize the supplementation of the depth geometric information of the power cable by taking point cloud data as the input of a network.
In a first aspect, an embodiment of the present application provides a method for supplementing depth geometric information of a power cable, where the method includes:
acquiring point cloud data of a power cable to be complemented as data to be processed, and acquiring a pre-trained information complementing network;
inputting the data to be processed into a coding layer of the information complementing network for characteristic coding to obtain coding characteristics;
and inputting the coding features to a decoding layer of the information complementing network for feature decoding to obtain complementing data after the information complementing of the data to be processed.
In a second aspect, an embodiment of the present application further provides a power cable depth geometric information complement device, where the power cable depth geometric information complement device includes:
the acquisition unit is used for acquiring the point cloud data of the electric power cable to be complemented as the data to be processed and acquiring the information complementing network trained in advance;
the coding unit is used for inputting the data to be processed into a coding layer of the information complement network to perform feature coding to obtain coding features;
and the decoding unit is used for inputting the coding characteristics to a decoding layer of the information complementing network to perform characteristic decoding to obtain the complementing data after the information complementing of the data to be processed.
In a third aspect, an embodiment of the present application further provides a detector, where the detector uses the method to complement the depth geometric information of the power cable.
In a fourth aspect, embodiments of the present application also provide an electronic device, including a memory storing at least one instruction; and a processor executing the instructions stored in the memory to implement the above method.
In a fifth aspect, embodiments of the present application also provide a computer-readable storage medium having at least one instruction stored therein, the at least one instruction being executed by a processor in an electronic device to implement the above-described method.
The embodiment of the application provides a power cable depth geometric information supplementing method, a detector, equipment and a medium. Has the following beneficial effects: the point cloud data is used as the input of the constructed information complementing network, the geometric information with complete shape is saved, the operation efficiency of the memory is improved, the point cloud data is encoded and decoded through the information complementing network, dense and complete point cloud is generated in a mode from thick to thin, the depth information complementing of the power cable has a better effect, and the condition of incomplete imaging of the indoor and outdoor cables can be dealt with.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a schematic view of an application environment of a method for supplementing geometric information of depth of a power cable according to an embodiment of the present application;
fig. 2 is a schematic flow chart of a method for supplementing geometric information of depth of a power cable according to an embodiment of the present application;
fig. 3 is a schematic diagram of a functional module of a device for supplementing geometric information of depth of a power cable according to an embodiment of the present application;
fig. 4 is a schematic block diagram of an electronic device according to an embodiment of the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all, of the embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.
It should be understood that the terms "comprises" and "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
It is also to be understood that the terminology used in the description of the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
It should be further understood that the term "and/or" as used in this specification and the appended claims refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations.
The embodiment of the application provides a power cable depth geometric information supplementing method, a detector, equipment and a medium, wherein the method can be applied to a power cable depth geometric information supplementing device.
The execution main body of the power cable depth geometric information supplementing method may be the power cable depth geometric information supplementing device provided in the embodiment of the present application, or an electronic device integrated with the power cable depth geometric information supplementing device, where the power cable depth geometric information supplementing device may be implemented in a hardware or software manner, and the electronic device may be a detector, or specifically may be a control device in the detector, or a terminal or a server that has a communication connection with the control device.
Referring to fig. 1, fig. 1 is a schematic view of an application environment of a method for supplementing geometric information of a depth of a power cable according to an embodiment of the present application. The detector 20 communicates with the distribution network robot 10, the distribution network robot 10 includes a mechanical arm, and the distribution network robot 10 operates the mechanical arm to carry the working tool 30 to move to a target working point according to a detection result of the detector 20 to perform a work. The detector 20 acquires the image of the distribution network robot 10 in real time through a connected depth image acquisition device 40, and the depth image acquisition device 40 may also be connected to the distribution network robot 10, for example: the depth image capture device 40 may be mounted at the end of a robotic arm of the distribution network robot 10.
Referring to fig. 2, fig. 2 is a flow chart of a method for supplementing geometric information of depth of a power cable according to an embodiment of the present application. The power cable depth geometric information supplementing method provided by the embodiment of the application can be applied to a power cable depth geometric information supplementing device. Of course, in other embodiments, the method for supplementing the depth geometric information of the power cable can be applied to other types of detection devices, which is not limited in the application.
Specifically, the power cable depth geometric information complementing method includes the following steps S110 to S130.
S110, acquiring point cloud data of the electric power cable to be complemented as data to be processed, and acquiring a pre-trained information complementing network.
In this embodiment, before the obtaining the to-be-completed power cable point cloud data as the to-be-processed data, the method further includes:
acquiring a depth image acquired by a depth image acquisition device and acquiring a camera internal reference of the depth image acquisition device;
and converting the depth map into three-dimensional coordinates under a world coordinate system by taking the camera internal parameters as constraints to obtain the power cable point cloud data.
Wherein, the depth image acquisition device can comprise a depth camera and the like.
The camera internal parameters can comprise the distance from an imaging plane to a lens and the projection of a center point of the lens on the imaging plane.
Through the embodiment, the collected depth map of the power cable can be converted into point cloud data.
The obtained point cloud data may have a fracture, a void, or the like due to the influence of weather, the material of the power cable itself, the color, or the like, and therefore, it is necessary to perform depth geometric information completion on the point cloud data having information loss.
In this embodiment, before the obtaining the pre-trained information complements the network, the method further includes:
acquiring an initial network, training samples and a pre-constructed loss function; the training samples comprise point cloud data with missing information and point cloud data with complete information corresponding to the point cloud data with missing information;
determining the point cloud data with the missing information as input data, determining the point cloud data with the complete information as a training target, and performing back propagation training on an initial network by using the loss function;
and stopping training when the value of the loss function reaches convergence, and obtaining the information completion network.
Wherein the initial network includes a codec structure.
Through the information complement network, corresponding point cloud data after information complement can be effectively predicted under the condition of given input point cloud data. In addition, the point cloud data is easy to obtain, so that a large-scale training sample can be used for training, and the training effect of the network is better. The point cloud data stores geometric information with complete shape, and the operation efficiency of the memory is improved.
In this embodiment, it is further required to detect whether the value of the loss function reaches convergence in the training process.
Specifically, the method further comprises:
in the training process, performing reference sampling on the point cloud data with complete information to obtain a sub-sample, acquiring rough point cloud data corresponding to the sub-sample in the output data of the current model, acquiring fine point cloud data corresponding to the point cloud data with missing information in the output data of the current model, and acquiring configuration weights;
calculating a Distance between the subsamples and corresponding coarse point cloud data to be used as a first Distance by using a Chamfer Distance function (CD) and an Earth motion Distance function (EMD), and calculating a Distance between the point cloud data with complete information and corresponding fine point cloud data to be used as a second Distance by using the Chamfer Distance function;
calculating a weighted sum of the first distance and the second distance according to the configuration weight to obtain the value of the loss function;
and when the loss function value is detected not to be reduced, determining that the loss function value reaches convergence.
The configuration weights may be configured in a self-defined manner, or may be obtained according to a large number of experiments, which is not limited in this application.
For example: the loss function may be expressed as:
L(Y c ,Y d ,Y gt )＝d 1 (Y c ,Y′ gt )+α*d 2 (Y d ,Y gt )
wherein Y is c Representing coarse point cloud data, Y d Representing fine point cloud data, Y gt Point cloud data representing complete information, Y' gt Represents subsamples, α represents the configuration weights, L (Y c ,Y d ,Y gt ) Representing the loss function, d 1 (Y c ,Y′ gt ) Representing the first distance, d, which is the distance between the rough point cloud data and the subsamples of the corresponding reference sample 2 (Y d ,Y gt ) And representing the second distance as the distance between the fine point cloud data and the point cloud data with complete information. And Y' gt And Y is equal to c Is the same size.
When calculating the first distance and the second distance, since both point clouds are disordered, the arrangement of the points is required to be unchanged by the loss function, and therefore a chamfer distance function and a ground movement distance function are introduced.
Specifically, the chamfer distance function may be expressed as:
wherein S is 1 ,S 2 Representing two point cloud sets, CD (S 1 ,S 2 ) Represent S 1 And S is equal to 2 The chamfer distance between the two is x is S 1 In (2), y is S 2 Is a point in (a).
Specifically, the ground movement distance function may be expressed as:
wherein, the liquid crystal display device comprises a liquid crystal display device,
for minimizing S 1 ，EMD(S 1 ,S 2 ) Represent S 1 And S is equal to 2 Ground movement distance between them.
And S120, inputting the data to be processed into a coding layer of the information complement network for feature coding, and obtaining coding features.
In this embodiment, inputting the data to be processed to the coding layer of the information complementing network to perform feature coding, and obtaining the coding feature includes:
acquiring the number of points input in the data to be processed;
converting the data to be processed into a matrix according to the number of the input points to obtain a first matrix;
converting each point in the first matrix into a feature vector by using a first multi-layer perceptron, and forming a second matrix by using the feature vectors obtained by conversion;
performing maximum pooling operation on the second matrix point by point to obtain a third matrix;
splicing the second matrix and the third matrix to obtain an enhancement matrix;
converting each point in the enhancement matrix into a feature vector by using a second multi-layer perceptron, and forming a fourth matrix by using the feature vectors obtained by conversion;
and carrying out maximum pooling operation on the fourth matrix point by point to obtain the coding characteristic.
Wherein the first multi-layer perceptron consists of two linear layers with ReLu activation functions.
Likewise, the second multi-layer perceptron is also composed of two linear layers with ReLu activation functions.
For example: the data to be processed is X, is power cable point cloud data, has m points, is input into a coding layer of the information complement network, is converted into a first matrix P with m X3, each row of P is a 3D coordinate of a point p_i= (X, y, z), each point p_i in P is converted into a feature vector f_i by utilizing a first multi-layer perceptron, a second matrix F is formed, and a point-by-point maximum pooling operation is carried out to obtain a global feature third matrix g with k dimensions. Further, stitching g with each f_i to obtain [ f_i, g ]]To obtain an enhanced point feature matrix, i.e. an enhanced matrix
Then->
And outputting a final feature vector, namely the coding feature v, through another shared second multi-layer perceptron and a point-by-point maximization pooling operation.
And S130, inputting the coding features to a decoding layer of the information complementing network for feature decoding to obtain complementing data after the information complementing of the data to be processed.
In this embodiment, the inputting the coding feature to the decoding layer of the information complementing network to perform feature decoding, and obtaining the complementing data after the information complementing of the data to be processed includes:
inputting the coding features to a full-connection layer in the decoding layer for processing to obtain target rough point cloud data;
and taking the points in the target rough point cloud data and the coding features as inputs, and carrying out up-sampling processing on the points in the target rough point cloud data through a pre-configured 2D grid to obtain fine point cloud data serving as the complement data.
For example: the method comprises the steps of transmitting the coding characteristic v through a fully-connected network with 3s output units to generate rough output, namely target rough point cloud data Y_coarse, changing the output into a matrix with s x 3 again, specifically taking a point q_i in the Y_coarse and the coding characteristic v as input through folding operation, generating a patch of local coordinates t=u 2 points taking q_i as a center through a 2D grid of deformation (form) u x u, namely upsampling a point q_i to u 2 points through the 2D grid of deformation. The specific method comprises the following steps:
firstly taking points on a u grid with the side length of r and 0 as the center, rearranging and flattening the coordinates of the u grid, such as a matrix with the original u x 2, wherein 2 represents that the dimension of coordinate points in the matrix is two-dimensional, the formed matrix is G after flattening, then conducting a jointing operation on each row of G and the coordinates and v of the center point q_i, and conducting a shared MLP sharing multi-layer perceptron on the jointed matrix to obtain a final t 3 matrix Q, namely a final patch output, wherein the whole process can expand a single point q_i (1*3) to t 3, further, the operation is conducted on each point in Y_source, and the point number of the final obtained fine point cloud data Y_detail is s t points, namely the final complement data.
Unlike the conventional encoder which uses three-dimensional codes for input, the embodiment uses point cloud data as input, and utilizes the folding decoder to realize the output of multi-stage point cloud data, so that not only the global geometric features, but also the local finer geometric features can be identified, and the information is more complete.
According to the technical scheme, the point cloud data is used as the input of the constructed information complementing network, the geometric information with complete shape is saved, the operation efficiency of the memory is improved, the point cloud data is encoded and decoded through the information complementing network, dense and complete point cloud is generated in a mode from thick to thin, the depth information complementing of the power cable has a better effect, and the situation of incomplete imaging of the indoor and outdoor power cables can be dealt with.
Fig. 3 is a schematic block diagram of a depth geometric information supplementing device for an electric power cable according to an embodiment of the present application. As shown in fig. 3, corresponding to the above method for supplementing the depth geometric information of the power cable, the present application further provides a device for supplementing the depth geometric information of the power cable. The power cable depth geometric information complementing device comprises a unit for executing the power cable depth geometric information complementing method, and the device can be configured on a detector and the like. Specifically, referring to fig. 3, the power cable depth geometric information complement apparatus 300 includes an obtaining unit 301, an encoding unit 302, and a decoding unit 303, wherein:
the acquiring unit 301 is configured to acquire to-be-completed power cable point cloud data as to-be-processed data, and acquire a pre-trained information completion network;
the encoding unit 302 is configured to input the data to be processed to an encoding layer of the information complementing network for feature encoding, so as to obtain encoding features;
the decoding unit 303 is configured to input the coding feature to a decoding layer of the information complementing network to perform feature decoding, so as to obtain complementing data after the information complementing is performed on the data to be processed.
According to the technical scheme, the point cloud data is used as the input of the constructed information complementing network, the geometric information with complete shape is saved, the operation efficiency of the memory is improved, the point cloud data is encoded and decoded through the information complementing network, dense and complete point cloud is generated in a mode from thick to thin, the depth information complementing of the power cable has a better effect, and the situation of incomplete imaging of the indoor and outdoor power cables can be dealt with.
The above-described power cable depth geometry information supplementing device may be implemented in the form of a computer program that can be run on an electronic device as shown in fig. 4.
Referring to fig. 4, fig. 4 is a schematic block diagram of an electronic device according to an embodiment of the present application. The electronic device 400 may be a device such as a detector, in particular a manipulation device in the detector, or a terminal or server having a communication connection with the manipulation device.
Referring to fig. 4, the electronic device 400 includes a processor 402, a memory, and a network interface 405, which are connected by a system bus 401, wherein the memory may include a non-volatile storage medium 403 and an internal memory 404.
The non-volatile storage medium 403 may store an operating system 4031 and a computer program 4032. The computer program 4032 includes program instructions that, when executed, cause the processor 402 to perform a power cable depth geometry information complementation method.
The processor 402 is used to provide computing and control capabilities to support the operation of the overall electronic device 400.
The internal memory 404 provides an environment for the execution of a computer program 4032 in the non-volatile storage medium 403, which computer program 4032, when executed by the processor 402, causes the processor 402 to perform a power cable depth geometry information complementation method.
The network interface 405 is used for network communication with other devices. Those skilled in the art will appreciate that the structure shown in fig. 4 is merely a block diagram of a portion of the structure associated with the present application and is not limiting of the electronic device 400 to which the present application is applied, and that a particular electronic device 400 may include more or fewer components than shown, or may combine certain components, or have a different arrangement of components.
Wherein the processor 402 is configured to execute a computer program 4032 stored in the memory to implement the steps of:
acquiring point cloud data of a power cable to be complemented as data to be processed, and acquiring a pre-trained information complementing network;
inputting the data to be processed into a coding layer of the information complementing network for characteristic coding to obtain coding characteristics;
and inputting the coding features to a decoding layer of the information complementing network for feature decoding to obtain complementing data after the information complementing of the data to be processed.
It should be appreciated that in embodiments of the present application, the processor 402 may be a central processing unit (Central Processing Unit, CPU), the processor 402 may also be other general purpose processors, digital signal processors (Digital Signal Processor, DSPs), application specific integrated circuits (Application Specific Integrated Circuit, ASICs), off-the-shelf programmable gate arrays (Field-Programmable Gate Array, FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. Wherein the general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
Those skilled in the art will appreciate that all or part of the flow in a method embodying the above described embodiments may be accomplished by computer programs instructing the relevant hardware. The computer program comprises program instructions, and the computer program can be stored in a storage medium, which is a computer readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the flow steps of the embodiments of the method described above.
Accordingly, the present application also provides a storage medium. The storage medium may be a computer readable storage medium. The storage medium stores a computer program, wherein the computer program includes program instructions. The program instructions, when executed by the processor, cause the processor to perform the steps of:
acquiring point cloud data of a power cable to be complemented as data to be processed, and acquiring a pre-trained information complementing network;
inputting the data to be processed into a coding layer of the information complementing network for characteristic coding to obtain coding characteristics;
and inputting the coding features to a decoding layer of the information complementing network for feature decoding to obtain complementing data after the information complementing of the data to be processed.
The storage medium may be a U-disk, a removable hard disk, a Read-Only Memory (ROM), a magnetic disk, or an optical disk, or other various computer-readable storage media that can store program codes.
Those of ordinary skill in the art will appreciate that the elements and algorithm steps described in connection with the embodiments disclosed herein may be embodied in electronic hardware, in computer software, or in a combination of the two, and that the elements and steps of the examples have been generally described in terms of function in the foregoing description to clearly illustrate the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present application.
In the several embodiments provided in this application, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of each unit is only one logic function division, and there may be another division manner in actual implementation. For example, multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed.
The steps in the method of the embodiment of the application can be sequentially adjusted, combined and deleted according to actual needs. The units in the device of the embodiment of the application can be combined, divided and deleted according to actual needs. In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit.
The integrated unit may be stored in a storage medium if implemented in the form of a software functional unit and sold or used as a stand-alone product. Based on such understanding, the technical solution of the present application is essentially or a part contributing to the prior art, or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium, comprising several instructions for causing an electronic device (which may be a personal computer, a terminal, a network device, or the like) to perform all or part of the steps of the method described in the embodiments of the present application.
While the invention has been described with reference to certain preferred embodiments, it will be understood by those skilled in the art that various changes and substitutions of equivalents may be made and equivalents will be apparent to those skilled in the art without departing from the scope of the invention. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (10)
1. The power cable depth geometric information complementing method is characterized by comprising the following steps of:
acquiring point cloud data of a power cable to be complemented as data to be processed, and acquiring a pre-trained information complementing network;
inputting the data to be processed into a coding layer of the information complementing network for characteristic coding to obtain coding characteristics;
and inputting the coding features to a decoding layer of the information complementing network for feature decoding to obtain complementing data after the information complementing of the data to be processed.
2. The method for supplementing depth geometric information of a power cable according to claim 1, wherein before the obtaining of the point cloud data of the power cable to be supplemented as the data to be processed, the method further comprises:
acquiring a depth image acquired by a depth image acquisition device and acquiring a camera internal reference of the depth image acquisition device;
and converting the depth map into three-dimensional coordinates under a world coordinate system by taking the camera internal parameters as constraints to obtain the power cable point cloud data.
3. The power cable depth geometry information completion method of claim 1, wherein prior to the obtaining the pre-trained information completion network, the method further comprises:
acquiring an initial network, training samples and a pre-constructed loss function; the training samples comprise point cloud data with missing information and point cloud data with complete information corresponding to the point cloud data with missing information;
determining the point cloud data with the missing information as input data, determining the point cloud data with the complete information as a training target, and performing back propagation training on an initial network by using the loss function;
and stopping training when the value of the loss function reaches convergence, and obtaining the information completion network.
4. A method of supplementing depth-geometric information for a power cable as in claim 3, wherein said method further comprises:
in the training process, performing reference sampling on the point cloud data with complete information to obtain a sub-sample, acquiring rough point cloud data corresponding to the sub-sample in the output data of the current model, acquiring fine point cloud data corresponding to the point cloud data with missing information in the output data of the current model, and acquiring configuration weights;
calculating the distance between the subsamples and the corresponding rough point cloud data to be used as a first distance by using a chamfer distance function and a ground movement distance function, and calculating the distance between the point cloud data with complete information and the corresponding fine point cloud data to be used as a second distance by using the chamfer distance function;
calculating a weighted sum of the first distance and the second distance according to the configuration weight to obtain the value of the loss function;
and when the loss function value is detected not to be reduced, determining that the loss function value reaches convergence.
5. The method for supplementing depth geometric information of a power cable according to claim 1, wherein inputting the data to be processed into an encoding layer of the information supplementing network for feature encoding, and obtaining the encoded features includes:
acquiring the number of points input in the data to be processed;
converting the data to be processed into a matrix according to the number of the input points to obtain a first matrix;
converting each point in the first matrix into a feature vector by using a first multi-layer perceptron, and forming a second matrix by using the feature vectors obtained by conversion;
performing maximum pooling operation on the second matrix point by point to obtain a third matrix;
splicing the second matrix and the third matrix to obtain an enhancement matrix;
converting each point in the enhancement matrix into a feature vector by using a second multi-layer perceptron, and forming a fourth matrix by using the feature vectors obtained by conversion;
and carrying out maximum pooling operation on the fourth matrix point by point to obtain the coding characteristic.
6. The power cable depth geometry information completion method of claim 5, wherein the first multi-layer perceptron is comprised of two linear layers with a ReLu activation function.
7. The method for supplementing depth geometric information of a power cable according to claim 1, wherein inputting the coding feature to a decoding layer of the information supplementing network for feature decoding, obtaining the supplementing data of the data to be processed after information supplementation comprises:
inputting the coding features to a full-connection layer in the decoding layer for processing to obtain target rough point cloud data;
and taking the points in the target rough point cloud data and the coding features as inputs, and carrying out up-sampling processing on the points in the target rough point cloud data through a pre-configured 2D grid to obtain fine point cloud data serving as the complement data.
8. A detector, characterized in that it complements the power cable depth geometry information with a method according to any of claims 1-7.
9. An electronic device, the electronic device comprising:
a memory storing at least one instruction; a kind of electronic device with high-pressure air-conditioning system
A processor executing instructions stored in the memory to implement the power cable depth geometry information supplementing method of any one of claims 1-7.
10. A computer-readable storage medium, characterized by: the computer-readable storage medium has stored therein at least one instruction that is executed by a processor in an electronic device to implement the power cable depth geometric information complement method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310147150.4A CN116108601B (en) | 2023-02-21 | 2023-02-21 | Power cable depth geometric information supplementing method, detector, equipment and medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310147150.4A CN116108601B (en) | 2023-02-21 | 2023-02-21 | Power cable depth geometric information supplementing method, detector, equipment and medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116108601A true CN116108601A (en) | 2023-05-12 |
| CN116108601B CN116108601B (en) | 2023-11-14 |
Family
ID=86256026
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310147150.4A Active CN116108601B (en) | 2023-02-21 | 2023-02-21 | Power cable depth geometric information supplementing method, detector, equipment and medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116108601B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140207385A1 (en) * | 2011-08-26 | 2014-07-24 | Philip Morris Products Sa | Systems and methods for characterizing topological network perturbations |
| US20200082561A1 (en) * | 2018-09-10 | 2020-03-12 | Mapbox, Inc. | Mapping objects detected in images to geographic positions |
| CN110933031A (en) * | 2019-10-25 | 2020-03-27 | 国网吉林省电力有限公司电力科学研究院 | Intelligent power grid power distribution terminal unit intrusion detection method based on LSTM |
| CN111627054A (en) * | 2019-06-24 | 2020-09-04 | 长城汽车股份有限公司 | Method and device for predicting depth completion error map of high-confidence dense point cloud |
| CN113205104A (en) * | 2021-04-23 | 2021-08-03 | 广西大学 | Point cloud completion method based on deep learning |
| CN114659463A (en) * | 2022-03-14 | 2022-06-24 | 华南农业大学 | Plant phenotype acquisition device and acquisition method thereof |
| WO2023280558A1 (en) * | 2021-07-06 | 2023-01-12 | Nokia Technologies Oy | Performance improvements of machine vision tasks via learned neural network based filter |
-
2023
- 2023-02-21 CN CN202310147150.4A patent/CN116108601B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20140207385A1 (en) * | 2011-08-26 | 2014-07-24 | Philip Morris Products Sa | Systems and methods for characterizing topological network perturbations |
| US20200082561A1 (en) * | 2018-09-10 | 2020-03-12 | Mapbox, Inc. | Mapping objects detected in images to geographic positions |
| CN111627054A (en) * | 2019-06-24 | 2020-09-04 | 长城汽车股份有限公司 | Method and device for predicting depth completion error map of high-confidence dense point cloud |
| CN110933031A (en) * | 2019-10-25 | 2020-03-27 | 国网吉林省电力有限公司电力科学研究院 | Intelligent power grid power distribution terminal unit intrusion detection method based on LSTM |
| CN113205104A (en) * | 2021-04-23 | 2021-08-03 | 广西大学 | Point cloud completion method based on deep learning |
| WO2023280558A1 (en) * | 2021-07-06 | 2023-01-12 | Nokia Technologies Oy | Performance improvements of machine vision tasks via learned neural network based filter |
| CN114659463A (en) * | 2022-03-14 | 2022-06-24 | 华南农业大学 | Plant phenotype acquisition device and acquisition method thereof |
Non-Patent Citations (2)
| Title |
|---|
| 徐坤田;李越;: "寓于建筑空间的企业文化――中国工商银行总行营业办公楼建筑文化解析", 现代商业银行, no. 09, pages 66 - 70 * |
| 王志强;王珊;郭大鹏;刘文霞;张馨月;牛妍舒;李月乔;孙晓光;耿建宇;李介夫;: "电力作业远程监控系统中视频的优化传输方法", 电力科学与技术学报, no. 01, pages 171 - 177 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116108601B (en) | 2023-11-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20190080455A1 (en) | Method and device for three-dimensional feature-embedded image object component-level semantic segmentation | |
| CN111797650B (en) | Obstacle identification method, obstacle identification device, computer equipment and storage medium | |
| US11941831B2 (en) | Depth estimation | |
| DE112019002589T5 (en) | DEPTH LEARNING SYSTEM | |
| US20060257049A1 (en) | Representation and coding of panoramic and omnidirectional images | |
| CN115348435B (en) | Geometric fusion of multiple image-based depth images using ray casting | |
| JP2021012683A (en) | Computer vision method and system | |
| CN113807361B (en) | Neural network, target detection method, neural network training method and related products | |
| CN110998671B (en) | Three-dimensional reconstruction method, device, system and storage medium | |
| CN113256718B (en) | Positioning method and device, equipment and storage medium | |
| CN111028279A (en) | Point cloud data processing method and device, electronic equipment and storage medium | |
| JP2024507727A (en) | Rendering a new image of a scene using a geometric shape recognition neural network conditioned on latent variables | |
| CN116228962A (en) | Large scene neuroview synthesis | |
| CN115880435A (en) | Image reconstruction method, model training method, device, electronic device and medium | |
| JP2022521253A (en) | Image processing to determine the thickness of an object | |
| Li et al. | Polarmesh: A star-convex 3d shape approximation for object pose estimation | |
| EP3940636A2 (en) | Method for acquiring three-dimensional perception information based on external parameters of roadside camera, and roadside device | |
| US20240037788A1 (en) | 3d pose estimation in robotics | |
| CN116108601B (en) | Power cable depth geometric information supplementing method, detector, equipment and medium | |
| CN112085842B (en) | Depth value determining method and device, electronic equipment and storage medium | |
| CN116863078A (en) | Three-dimensional human body model reconstruction method, three-dimensional human body model reconstruction device, electronic equipment and readable medium | |
| CN116079727A (en) | Humanoid robot motion simulation method and device based on 3D human body posture estimation | |
| CN116229394A (en) | Automatic driving image recognition method, device and recognition equipment | |
| Ervan et al. | Downsampling of a 3D LiDAR point cloud by a tensor voting based method | |
| Kallasi et al. | Object detection and pose estimation algorithms for underwater manipulation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |