[go: up one dir, main page]

CN119478324B - A method and device for removing rain and fog from power UAV inspection images - Google Patents

A method and device for removing rain and fog from power UAV inspection images Download PDF

Info

Publication number
CN119478324B
CN119478324B CN202510038557.2A CN202510038557A CN119478324B CN 119478324 B CN119478324 B CN 119478324B CN 202510038557 A CN202510038557 A CN 202510038557A CN 119478324 B CN119478324 B CN 119478324B
Authority
CN
China
Prior art keywords
fog
tensor
sub
image
inspection
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.)
Active
Application number
CN202510038557.2A
Other languages
Chinese (zh)
Other versions
CN119478324A (en
Inventor
黄志鸿
孙云龙
吴晟
龙彦伯
张可人
乔宏
张辉
钟杭
毛建旭
孔庆宇
张国梁
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Electric Power Research Institute of State Grid Hunan Electric Power Co Ltd
State Grid Hunan Electric Power Co Ltd
Hunan Xiangdian Test Research Institute Co Ltd
State Grid Corp of China SGCC
Original Assignee
Electric Power Research Institute of State Grid Hunan Electric Power Co Ltd
State Grid Hunan Electric Power Co Ltd
Hunan Xiangdian Test Research Institute Co Ltd
State Grid Corp of China SGCC
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Electric Power Research Institute of State Grid Hunan Electric Power Co Ltd, State Grid Hunan Electric Power Co Ltd, Hunan Xiangdian Test Research Institute Co Ltd, State Grid Corp of China SGCC filed Critical Electric Power Research Institute of State Grid Hunan Electric Power Co Ltd
Priority to CN202510038557.2A priority Critical patent/CN119478324B/en
Publication of CN119478324A publication Critical patent/CN119478324A/en
Application granted granted Critical
Publication of CN119478324B publication Critical patent/CN119478324B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; 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 OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/20—Image preprocessing
    • G06V10/30—Noise filtering
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/513—Sparse representations
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/52—Scale-space analysis, e.g. wavelet analysis
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
    • G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
    • G06V10/761—Proximity, similarity or dissimilarity measures
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/17—Terrestrial scenes taken from planes or by drones

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Multimedia (AREA)
  • Health & Medical Sciences (AREA)
  • Business, Economics & Management (AREA)
  • General Health & Medical Sciences (AREA)
  • Economics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Remote Sensing (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • General Business, Economics & Management (AREA)
  • Water Supply & Treatment (AREA)
  • Public Health (AREA)
  • Marketing (AREA)
  • Tourism & Hospitality (AREA)
  • Primary Health Care (AREA)
  • Artificial Intelligence (AREA)
  • Computing Systems (AREA)
  • Databases & Information Systems (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Image Processing (AREA)

Abstract

The invention provides a rain and fog removing method and device for an inspection image of an electric unmanned aerial vehicle, wherein the method comprises the steps of constructing m sub-tensors with different spatial scales for the acquired inspection imageWherein m is an integer and m is greater than or equal to 2,Is an integer andSub-tensors using low rank tensor and spatial critical domain similarity of inspection imagesRemoving rain and fog to obtain an initial rain and fog removal imageFor the initial rain and fog removing image under each spatial scaleDecision average fusion is carried out to generate a final restored image. The method can better utilize the context space structure information to improve the rainwater mist removal performance, can ensure that the electric unmanned aerial vehicle can acquire high-quality inspection images even under unfavorable weather conditions, improves the intelligent operation and inspection safety and reliability of the electric power facilities, can be widely applied to the fields of inspection of the electric unmanned aerial vehicle, and has wide application prospect and market value.

Description

Rain and fog removing method and device for inspection image of electric unmanned aerial vehicle
Technical Field
The invention relates to the technical field of image processing, in particular to a method and a device for removing rain and fog in an inspection image of an electric unmanned aerial vehicle.
Background
The electric unmanned aerial vehicle is widely applied to the inspection of the electric overhead line due to the simplicity and convenience in operation and high-efficiency operation capability. Particularly in areas with complex geographic conditions, such as mountain areas, the unmanned aerial vehicle can overcome the terrain obstacle, and the effective monitoring of the power transmission line is realized. However, in severe natural environments, such as rainy, snowy and cloudy weather, the quality of the inspection image acquired by the unmanned aerial vehicle can be affected, so that part of spatial textures and detailed information in the image are missing or blurred.
In this case, the image restoration technique becomes particularly important. The definition of the interfered image can be restored to a certain extent, and the detection precision is improved. The application of the technology is helpful for ensuring that the electric unmanned aerial vehicle can provide high-quality inspection data even under unfavorable weather conditions (such as rain and fog interference), thereby ensuring the safe operation and maintenance of electric facilities.
In view of the above, there is a strong need for a method and a device for removing rain and fog in inspection images of an electric unmanned aerial vehicle to solve the problems in the prior art.
Disclosure of Invention
The invention aims to provide a rain and fog removing method for an inspection image of an electric unmanned aerial vehicle, which aims to remove the interference of rain and fog on the inspection image and meet the high-quality imaging requirement on the inspection image, and the specific technical scheme is as follows:
the method for removing rain and fog of the inspection image of the electric unmanned aerial vehicle comprises the following steps:
s1, constructing m sub tensors with different spatial scales for the acquired inspection image Wherein m is an integer and m is greater than or equal to 2,Is an integer and;
S2, utilizing low-rank tensor and spatial critical domain similarity of inspection images to sub-tensorRemoving rain and fog to obtain an initial rain and fog removal image;
Step S3, performing initial rain and fog removal on the images in each spatial scaleDecision average fusion is carried out to generate a final restored image。
In the above technical solution, preferably, in step S1, the sub-tensor for the inspection image YExpressed as:
(1),
wherein, Representing neutron tensors in inspection imagesIs set in the number of (3),Represents the nth spatial scaleSub tensor。
In the above technical scheme, a sliding window with a step length of 1 is set, and a sub-tensor is takenIs of the spatial scale size ofThenWherein M, N are two space dimensions of the inspection image respectively, Q is the channel number of the inspection image,Representing sub-tensorsIs used for the spatial dimension of the (c) in the (c),Satisfying both less than M and N.
In the above technical solution, preferably, in step S2, the low-rank tensor and the spatial critical domain similarity of the inspection image are used for the sub-tensorThe solution formula for removing rain and fog is as follows:
(8)
wherein, Indicating that the center is at the nth spatial scaleIs used for the low-rank sub-tensor of (c),Indicating that the center is at the nth spatial scaleIs used to determine the sparse sub-tensor of (c),Indicating that the center is at the nth spatial scaleIs used to determine the sub-tensor of (c),For the spatial critical domain similarity constraint, λ is a parameter used to balance the weight between the low rank and sparse sub-tensors,As the kernel norm of the sub-tensor,Is the 1-norm of the sub-tensor,Is the fourier norm of the matrix,Is a preset weight value.
In the above technical solution, preferably, the spatial critical domain similarity constraint termThe method comprises the following steps:
(9)
wherein, Indicating that the center is at the nth spatial scaleIs used to determine the sub-tensor of (c),Indicating that the center is at the nth spatial scaleIs used to determine the sub-tensor of (c),Indicating that the center is at the nth spatial scaleIs used to determine the sub-tensor of (c),Indicating that the center is at the nth spatial scaleIs used to determine the sub-tensor of (c),Indicating that the center is at the nth spatial scaleIs a sub-tensor of (c).
In the above technical solution, preferably, the formula (8) is converted into the formula (10) and the formula (11) for iterative solution:
(10)
(11)
obtaining the solved low-rank tensor under the nth spatial scale when the iteration ending condition is met Then the rain and fog removing image is initially carried outExpressed as:
(12)
wherein, Representing that the (k+1) th iteration center is at the (n) th spatial scaleIs used for the low-rank sub-tensor of (c),Representing that the kth iteration center is at the nth spatial scaleIs used to determine the sparse sub-tensor of (c),Representing neutron tensors in inspection imagesIs set in the number of (3),Represents the nth spatial scaleAnd a low rank sub-tensor.
In the above technical scheme, preferably when meetingObtaining solved low rank tensor;
In the formula (8)、Wherein Q is the number of channels of the inspection image,Representing sub-tensorsIs used for the spatial dimension of the (c) in the (c),A value is preset for the gaussian noise intensity.
In the above technical solution, it is preferable that in step S3, the initial rain and fog removing image under each spatial scale is obtained according to formula (13)Decision average fusion is carried out to generate a final restored image:
(13)。
The invention also provides a device for removing the rain and fog of the inspection image of the electric unmanned aerial vehicle, which comprises a memory and a processor, wherein the processor executes the method for removing the rain and fog of the inspection image of the electric unmanned aerial vehicle when running the computer instructions stored in the memory.
The invention further provides another device for removing rain and fog of the inspection image of the electric unmanned aerial vehicle, which comprises an image acquisition unit and a data calculation unit, wherein the image acquisition unit is used for shooting electric equipment to obtain the inspection image, and the data calculation unit is used for processing the inspection image according to the method for removing rain and fog of the inspection image of the electric unmanned aerial vehicle.
The technical scheme of the invention has the following beneficial effects:
The rain and fog removing method can better utilize the context space structure information to improve the performance of removing the rain and fog, can ensure that even under unfavorable weather conditions (such as rain and fog interference), an electric unmanned aerial vehicle can acquire high-quality inspection images, improves the safety and reliability of intelligent operation and inspection of electric facilities, can be widely applied to the fields of inspection of electric unmanned aerial vehicles and the like, and has wide application prospect and market value.
In addition to the objects, features and advantages described above, the present invention has other objects, features and advantages. The present invention will be described in further detail with reference to the drawings.
Drawings
The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and together with the description serve to explain the application. In the drawings:
fig. 1 is a flow chart of a method for removing rain and fog from an inspection image of an electric unmanned aerial vehicle.
Detailed Description
The present invention will be described more fully hereinafter in order to facilitate an understanding of the present invention, and preferred embodiments of the present invention are set forth. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.
Examples:
referring to fig. 1, the embodiment provides a method for removing rain and fog from an inspection image of an electric unmanned aerial vehicle, which comprises the following steps:
s1, constructing m sub tensors with different spatial scales for the acquired inspection image Wherein m is an integer and m is greater than or equal to 2,Is an integer and;
Preferably, the power equipment is photographed by the power unmanned aerial vehicle, and an original visible light image Y (namely, a patrol image) in a jpg format is obtained. The inspection image can be regarded as a three-dimensional tensor, and the data scale size is expressed asWherein M, N are two spatial dimensions of the inspection image respectively, Q is the number of channels of the inspection image, specifically rgb three channels, i.e. q=3.
Because of the background complexity of the power inspection image, the spatial structure information in the image is difficult to fully excavate by a single spatial scale, so that the spatial structure information in the inspection image is fully excavated by using m sub-tensors with different spatial scales in the embodiment, and specifically, the firstSub-tensors of individual spatial scalesThe construction method is as follows:
Designing a sub-tensor Its spatial dimension is,Representing sub-tensorsIs used for the spatial dimension of the (c) in the (c),A sliding window with the step length of 1 is designed, and each inspection image can be divided intoSub tensorThen the sub-tensor for the inspection image YCan be expressed as:
(1)
wherein, ,Represents the nth spatial scaleSub tensor。
Further, the inspection image Y is represented as a sub-tensor of m different spatial scales:
(2)
S2, utilizing low-rank tensor and spatial critical domain similarity of inspection images to sub-tensor Removing rain and fog to obtain an initial rain and fog removal image;
The prior art "C. Lu, J. Feng, Y. Chen, W. Liu, Z. Lin, and S. Yan, ''Tensor robust principal component analysis: Exact recovery of corrupted low-rank tensors via convex optimization,'' in Proc. IEEE Conf. Comput. Vis. Pattern Recognit., Jun. 2016, pp. 5249–5257." proposes a model for image restoration using a low-rank tensor restoration algorithm, which sets an input image Y that can be decomposed into a low-rank tensor L, a sparse tensor S, and a noise tensor N, expressed as:
(3)
Meanwhile, prior art "C. Lu, J. Feng, Y. Chen, W. Liu, Z. Lin, and S. Yan, ''Tensor robust principal component analysis: Exact recovery of corrupted low-rank tensors via convex optimization,'' in Proc. IEEE Conf. Comput. Vis. Pattern Recognit., Jun. 2016, pp. 5249–5257." also discloses a solved objective function expressed as:
(4)
Where lambda is a parameter for balancing the weight magnitudes between the low rank tensor L and the sparse tensor S, As the kernel norm of the tensor,As a 1-norm of the tensor,Is the fourier norm of the matrix,Preset value for Gaussian noise intensity, and is generally taken。
Because of the similarity of spatially neighboring pixels of the inspection image, a linear combination of several basis vectors can be used to represent each pixel in the background, and thus the entire area of the inspection image typically has a low rank characteristic. Compared with the whole image area, the rainwater and fog target is usually a small area target, and the rainwater and fog target has sparse characteristics, so that the rainwater and fog can be removed from the inspection image by using the formula (4).
The low-rank tensor recovery algorithm is characterized in that the rain and fog of the image are removed by utilizing the low-rank characteristic of the image. However, due to the complexity of the inspection image, the rain and fog targets in the image may not be removed well using only the low rank features of the image. Therefore, the embodiment further innovates the pair tensor by utilizing the low-rank tensor and the spatial threshold similarity of the patrol image based on the prior artTo remove rain and fog, in order to improve the effect of removing rain and fog of images, the embodiment constructs a spatial critical domain similarity constraint termAdding to equation (4) yields equation (5):
(5)
Specifically, in the present embodiment Expressed as:
(6)
wherein, The representation being centered atIs used to determine the sub-tensor of (c),The representation being centered atIs used to determine the sub-tensor of (c),The representation being centered atIs used to determine the sub-tensor of (c),The representation being centered atIs a sub-tensor of (c).
The objective function for converting equation (5) into a penalty-containing term is:
(7)
wherein, Is a preset weight value.
Equation (7) is a solving equation for performing low-rank tensor decomposition on the whole inspection image, and the sub-tensor is calculated according to equation (1)Can be converted into the following formula to be solved:
(8)
wherein, Indicating that the center is at the nth spatial scaleIs used for the low-rank sub-tensor of (c),Indicating that the center is at the nth spatial scaleIs used to determine the sparse sub-tensor of (c),Indicating that the center is at the nth spatial scaleIs used to determine the sub-tensor of (c),Is a spatial critical domain similarity constraint term, and in the formula (8), lambda is a parameter for balancing weight magnitude between a low rank tensor and a sparse tensor,As the kernel norm of the sub-tensor,Is the 1-norm of the sub-tensor,Is the fourier norm of the matrix,Is a preset weight value, and further, is set in a formula (8)、。
Further, the spatial critical domain similarity constraint termBy mining the similarity information in the space domain, the smoothness of image restoration can be improved, the influence of noise is further eliminated or reduced as much as possible, and the image quality is improved, wherein the similarity constraint term of the space critical domain in the formula (8)The method comprises the following steps:
(9)
wherein, Indicating that the center is at the nth spatial scaleIs used to determine the sub-tensor of (c),Indicating that the center is at the nth spatial scaleIs used to determine the sub-tensor of (c),Indicating that the center is at the nth spatial scaleIs used to determine the sub-tensor of (c),Indicating that the center is at the nth spatial scaleIs used to determine the sub-tensor of (c),Indicating that the center is at the nth spatial scaleIs a sub-tensor of (c).
Further, the formula (8) can be converted into the formula (10) and the formula (11) to perform iterative solution:
(10)
(11)
wherein, Representing that the (k+1) th iteration center is at the (n) th spatial scaleIs used for the low-rank sub-tensor of (c),Representing that the kth iteration center is at the nth spatial scaleIs a sparse sub-tensor of (c).
When meeting the requirementsObtaining solved low rank tensorObtaining an overall low rank tensor after recovery at the nth spatial scale according to equation (1)(I.e., the initial rain and fog removal image) is expressed as:
(12)
Preferably, in this embodiment, the iterative solution is performed on the formula (10) and the formula (11) by using a succession iterative algorithm, which is referred to in the prior art "X. Yuan and J. Yang, ''Sparse and low-rank matrix decomposition via alternating direction methods,' Pacific J. Optim., vol. 9, no. 1, pp. 167–180, 2013.", and is not described in detail in this embodiment.
Step S3, performing initial rain and fog removal on the images in each spatial scaleAnd carrying out decision average fusion to generate a final recovery image.
The initial rain and fog removal image under different spatial scales is carried out according to the formula (13)Fusing to obtain a fused recovery image:
(13)
wherein, And finally removing the recovered image of the rain mist after fusion.
The embodiment also provides a device for removing the rain and fog of the inspection image of the electric unmanned aerial vehicle, which comprises a memory and a processor, wherein the processor executes the method for removing the rain and fog of the inspection image of the electric unmanned aerial vehicle when running the computer instructions stored in the memory.
The embodiment provides another device for removing rain and fog of the inspection image of the electric unmanned aerial vehicle, which comprises an image acquisition unit and a data calculation unit, wherein the image acquisition unit is used for shooting electric equipment to obtain the inspection image, and the data calculation unit is used for processing the inspection image according to the method for removing the rain and fog of the inspection image of the electric unmanned aerial vehicle.
The rain and fog removing method can better utilize the context space structure information to improve the performance of removing the rain and fog, can ensure that the electric unmanned aerial vehicle can acquire high-quality inspection images even under unfavorable weather conditions (such as rain and fog interference), improves the intelligent operation and inspection safety and reliability of electric facilities, can be widely applied to the fields of inspection of the electric unmanned aerial vehicle, and has wide application prospects and market values.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the present invention, but various modifications and variations can be made to the present invention by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (8)

1.一种电力无人机巡检图像去雨雾方法，其特征在于，包括以下步骤：1. A method for removing rain and fog from power UAV inspection images, characterized in that it includes the following steps: 步骤S1、对采集的巡检图像构建m种不同空间尺度的子张量；其中，m为整数且m大于等于2，为整数且；Step S1: construct m sub-tensors of different spatial scales for the collected inspection images ; Wherein, m is an integer greater than or equal to 2, is an integer and ; 步骤S2、利用巡检图像的低秩张量和空间临域相似性对子张量进行去雨雾，得到初始去雨雾图像；Step S2: Use the low-rank tensor of the inspection image and the spatial neighborhood similarity to obtain the sub-tensor Remove rain and fog to obtain the initial rain and fog-free image ; 步骤S3、对各空间尺度下的初始去雨雾图像进行决策平均融合，生成最终的恢复图像；Step S3: Initial rain and fog removal images at each spatial scale Perform decision averaging fusion to generate the final restored image ; 步骤S2中利用巡检图像的低秩张量和空间临域相似性对子张量进行去雨雾的求解公式为：In step S2, the low-rank tensor of the inspection image and the spatial neighborhood similarity are used to obtain the sub-tensor The solution formula for removing rain and fog is: （8） (8) 其中，表示第n种空间尺度下中心在的低秩子张量，表示第n种空间尺度下中心在的稀疏子张量，表示第n种空间尺度下中心在的子张量，为空间临域相似性约束项，λ是一个用于平衡低秩子张量和稀疏子张量之间权重大小的参数，为子张量的核范数，为子张量的1范数，为矩阵的傅里叶范数，为预设权重值；in, Indicates that the center is at the nth spatial scale The low-rank subtensor of Indicates that the center is at the nth spatial scale The sparse subtensor of Indicates that the center is at the nth spatial scale The sub-tensor of is the spatial neighborhood similarity constraint, λ is a parameter used to balance the weights between the low-rank sub-tensor and the sparse sub-tensor. is the nuclear norm of the subtensor, is the 1-norm of the subtensor, is the Fourier norm of the matrix, is the preset weight value; 所述空间临域相似性约束项具体为：The spatial neighborhood similarity constraint Specifically: （9） (9) 其中，表示第n种空间尺度下中心在的子张量，表示第n种空间尺度下中心在的子张量，表示第n种空间尺度下中心在的子张量，表示第n种空间尺度下中心在的子张量。in, Indicates that the center is at the nth spatial scale The sub-tensor of Indicates that the center is at the nth spatial scale The sub-tensor of Indicates that the center is at the nth spatial scale The sub-tensor of Indicates that the center is at the nth spatial scale A sub-tensor of . 2.根据权利要求1所述的电力无人机巡检图像去雨雾方法，其特征在于，步骤S1中巡检图像Y用子张量表示为：2. The method for removing rain and fog from inspection images of electric power drones according to claim 1 is characterized in that in step S1, the inspection image Y is obtained by using a sub-tensor It is expressed as: （1） (1) 其中，表示巡检图像中子张量的个数，表示第n种空间尺度下的第个子张量。in, Represents the subtensor in the inspection image The number of represents the nth spatial scale Subtensor . 3.根据权利要求2所述的电力无人机巡检图像去雨雾方法，其特征在于：3. The method for removing rain and fog from power drone inspection images according to claim 2 is characterized in that: 设置一个步长为1的滑动窗口，取子张量的空间尺度大小为，则；其中，M、N分别为巡检图像的两个空间维度，Q为巡检图像的通道数，表示子张量的空间维度，满足同时小于M和N。Set a sliding window with a step size of 1 and take the sub-tensor The spatial scale is ,but ; Wherein, M and N are the two spatial dimensions of the inspection image, Q is the number of channels of the inspection image, Represents a sub-tensor The spatial dimension of Satisfies that it is less than both M and N. 4.根据权利要求1所述的电力无人机巡检图像去雨雾方法，其特征在于，将公式（8）转化为公式（10）和公式（11）进行迭代求解：4. The method for removing rain and fog from power drone inspection images according to claim 1 is characterized in that formula (8) is converted into formula (10) and formula (11) for iterative solution: （10） (10) （11） (11) 当满足迭代结束条件时得到求解后的在第n种空间尺度下的低秩子张量，则初始去雨雾图像表示为：When the iteration end condition is met, the low-rank subtensor at the nth spatial scale is obtained after the solution , then the initial rain and fog removal image It is expressed as: （12） (12) 其中，表示第n种空间尺度下第k+1次迭代中心在的低秩子张量，表示第n种空间尺度下第k次迭代中心在的稀疏子张量，表示巡检图像中子张量的个数，表示第n种空间尺度下的第个低秩子张量。in, Indicates that the k +1th iteration center at the nth spatial scale is The low-rank subtensor of Indicates that the kth iteration center at the nth spatial scale is The sparse subtensor of Represents the subtensor in the inspection image The number of represents the nth spatial scale A low-rank subtensor. 5.根据权利要求4所述的电力无人机巡检图像去雨雾方法，其特征在于：5. The method for removing rain and fog from power drone inspection images according to claim 4 is characterized in that: 当满足时得到求解后的低秩子张量；When satisfied The low-rank subtensor obtained after the solution is ; 在公式（8）中、；其中，Q为巡检图像的通道数，表示子张量的空间维度，为高斯噪声强度预设值。In formula (8) , ; Where, Q is the number of channels of the inspection image, Represents a sub-tensor The spatial dimension of Preset value for Gaussian noise intensity. 6.根据权利要求1所述的电力无人机巡检图像去雨雾方法，其特征在于，步骤S3中根据公式（13）对各空间尺度下的初始去雨雾图像进行决策平均融合，生成最终的恢复图像：6. The method for removing rain and fog from power UAV inspection images according to claim 1 is characterized in that in step S3, the initial rain and fog removal images at each spatial scale are calculated according to formula (13). Perform decision averaging fusion to generate the final restored image : （13）。 (13). 7.一种电力无人机巡检图像去雨雾装置，其特征在于，包括存储器和处理器，所述处理器运行存储器中存储的计算机指令时，执行如权利要求1-6任意一项所述的电力无人机巡检图像去雨雾方法。7. A device for removing rain and fog from inspection images of electric power drones, characterized in that it comprises a memory and a processor, and when the processor runs the computer instructions stored in the memory, it executes the method for removing rain and fog from inspection images of electric power drones as described in any one of claims 1-6. 8.一种电力无人机巡检图像去雨雾装置，其特征在于，包括图像获取单元和数据计算单元；所述图像获取单元用于拍摄电力设备以获得巡检图像；所述数据计算单元根据如权利要求1-6任意一项所述的电力无人机巡检图像去雨雾方法对巡检图像进行处理。8. A device for removing rain and fog from inspection images of electric power drones, characterized in that it comprises an image acquisition unit and a data calculation unit; the image acquisition unit is used to photograph electric power equipment to obtain inspection images; the data calculation unit processes the inspection images according to the method for removing rain and fog from inspection images of electric power drones as described in any one of claims 1-6.
CN202510038557.2A 2025-01-10 2025-01-10 A method and device for removing rain and fog from power UAV inspection images Active CN119478324B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202510038557.2A CN119478324B (en) 2025-01-10 2025-01-10 A method and device for removing rain and fog from power UAV inspection images

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202510038557.2A CN119478324B (en) 2025-01-10 2025-01-10 A method and device for removing rain and fog from power UAV inspection images

Publications (2)

Publication Number Publication Date
CN119478324A CN119478324A (en) 2025-02-18
CN119478324B true CN119478324B (en) 2025-04-15

Family

ID=94596780

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202510038557.2A Active CN119478324B (en) 2025-01-10 2025-01-10 A method and device for removing rain and fog from power UAV inspection images

Country Status (1)

Country Link
CN (1) CN119478324B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109859119A (en) * 2019-01-07 2019-06-07 南京邮电大学 A kind of video image rain removing method restored based on adaptive low-rank tensor
CN115439344A (en) * 2022-08-02 2022-12-06 武汉大学 Mixed noise hyperspectral image restoration method combining double low rank and spatial spectrum total variation

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8989465B2 (en) * 2012-01-17 2015-03-24 Mayo Foundation For Medical Education And Research System and method for medical image reconstruction and image series denoising using local low rank promotion
US10217018B2 (en) * 2015-09-15 2019-02-26 Mitsubishi Electric Research Laboratories, Inc. System and method for processing images using online tensor robust principal component analysis
CN109977989B (en) * 2019-01-17 2021-04-20 北京工业大学 Image tensor data processing method

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109859119A (en) * 2019-01-07 2019-06-07 南京邮电大学 A kind of video image rain removing method restored based on adaptive low-rank tensor
CN115439344A (en) * 2022-08-02 2022-12-06 武汉大学 Mixed noise hyperspectral image restoration method combining double low rank and spatial spectrum total variation

Also Published As

Publication number Publication date
CN119478324A (en) 2025-02-18

Similar Documents

Publication Publication Date Title
Wang et al. Haze concentration adaptive network for image dehazing
CN110827198B (en) Multi-camera panoramic image construction method based on compressed sensing and super-resolution reconstruction
CN111986108A (en) Complex sea-air scene image defogging method based on generation countermeasure network
CN105976383A (en) Power transmission equipment fault diagnosis method based on limit learning machine image recognition
CN112487874B (en) Method and system for eliminating background noise based on event camera
CN113362281B (en) Infrared and visible light image fusion method based on WSN-LatLRR
Hu et al. Seasondepth: Cross-season monocular depth prediction dataset and benchmark under multiple environments
CN118887453A (en) A visible light and infrared image fusion target detection method based on deep learning
CN118470659A (en) Waterlogging detection method and device based on denoising diffusion model from the perspective of urban monitoring
CN111275751A (en) Unsupervised absolute scale calculation method and system
CN119478324B (en) A method and device for removing rain and fog from power UAV inspection images
Ruan et al. Distill drops into data: Event-based rain-background decomposition network
CN119850467A (en) Polarized image defogging method based on double-iteration regular separation
CN114821434A (en) Space-time enhanced video anomaly detection method based on optical flow constraint
Chen et al. Multi‐scale single image dehazing based on the fusion of global and local features
CN117808721A (en) Low-illumination image enhancement method, device, equipment and medium based on deep learning
CN108596831B (en) Super-resolution reconstruction method based on AdaBoost example regression
CN111104875A (en) Moving target detection method under rain and snow weather conditions
CN116645280A (en) Video compression sensing restoration method and system based on generative network and low-rank regularization
Ye et al. Unsupervised Monocular depth estimation based on dual attention mechanism and depth-aware loss
Zaher et al. Clouds motion estimation from ground-based sky camera and satellite images
CN115937021A (en) Polarization Dehazing Method Based on Frequency Domain Feature Separation and Iterative Optimization of Atmospheric Light
Yan et al. Image defogging method for transmission channel inspection by UAV based on deep multi-patch layered network
CN114757222A (en) Multidimensional time series autoencoder denoising method for towed sensor array
CN114140739A (en) Video deblurring method based on patch matching and synthesis

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