CN114359792A - Deep learning-based method and device for identifying insulating gloves on electric power operation site - Google Patents
Deep learning-based method and device for identifying insulating gloves on electric power operation site Download PDFInfo
- Publication number
- CN114359792A CN114359792A CN202111544435.9A CN202111544435A CN114359792A CN 114359792 A CN114359792 A CN 114359792A CN 202111544435 A CN202111544435 A CN 202111544435A CN 114359792 A CN114359792 A CN 114359792A
- Authority
- CN
- China
- Prior art keywords
- data set
- training data
- color distribution
- information
- insulating
- 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 53
- 238000013135 deep learning Methods 0.000 title claims abstract description 35
- 238000012549 training Methods 0.000 claims abstract description 75
- 238000012706 support-vector machine Methods 0.000 claims abstract description 50
- 238000003062 neural network model Methods 0.000 claims abstract description 22
- 238000013528 artificial neural network Methods 0.000 claims abstract description 15
- 238000002372 labelling Methods 0.000 claims abstract description 6
- 239000013598 vector Substances 0.000 claims description 62
- 238000004590 computer program Methods 0.000 claims description 13
- 239000003086 colorant Substances 0.000 claims description 6
- 239000011159 matrix material Substances 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 4
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 238000012545 processing Methods 0.000 abstract description 2
- 238000004422 calculation algorithm Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000001514 detection method Methods 0.000 description 3
- 238000000513 principal component analysis Methods 0.000 description 3
- 238000013136 deep learning model Methods 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000000354 decomposition reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
Images
Landscapes
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The embodiment of the invention relates to the technical field of image processing, and discloses a method and a device for identifying insulating gloves in an electric power operation site based on deep learning. The method comprises the following steps: collecting pictures containing insulating gloves worn by power operators and normal hand information, manually labeling the pictures, and constructing a first training data set; training a deep neural network initial model for detecting the insulating gloves and the human hands by using the first training data set to obtain a deep neural network model; cutting all the insulating gloves and hand information in the first training data set according to the manually marked content to construct a second training data set; extracting the features of all the images in the second training data set to train the support vector machine; and determining the probability of the image to be recognized as the insulating glove and the human hand by using the deep neural network model and the support vector machine. By implementing the embodiment of the invention, the accuracy of the identification of the insulating gloves can be improved.
Description
Technical Field
The invention relates to the technical field of image recognition, in particular to a method and a device for recognizing insulating gloves on an electric power operation site based on deep learning.
Background
The electric power operation personnel need to wear the insulating gloves when carrying out the operation, and the method of whether the insulating gloves are worn correctly to the traditional inspection operation personnel is manual supervision, is limited by the operating time and the working scene, and the supervision personnel can not guarantee to supervise at any time. In the prior art, a computer vision method is considered, and a target detection algorithm is adopted to identify gloves worn on the hands of operators, so that although a certain identification effect can be achieved, misjudgment can often occur.
Disclosure of Invention
Aiming at the defects, the embodiment of the invention discloses a method and a device for identifying insulating gloves in an electric power operation field based on deep learning, which can effectively improve the accuracy of identifying the insulating gloves.
The embodiment of the invention discloses a method for identifying insulating gloves in an electric power operation field based on deep learning in a first aspect, which comprises the following steps:
collecting pictures containing insulating gloves worn by power operators and normal hand information, manually labeling the pictures, and constructing a first training data set;
training a deep neural network initial model for detecting the insulating gloves and the human hands by using the first training data set to obtain a deep neural network model;
cutting all the insulating gloves and hand information in the first training data set according to the manually marked content to construct a second training data set;
extracting the features of all the images in the second training data set, and using the extracted features as input information of a support vector machine, wherein the type of the images belongs to is output information of the support vector machine, so as to train the support vector machine, and the type of the images belongs to is an insulating glove or a hand;
determining position information of a target of an image to be recognized and a probability p of the target being an insulating glove and a human hand using the deep neural network modelg1、ph1Cutting the image to be recognized according to the position information, inputting the cut image to be recognized into a support vector machine, and obtaining the probability p of the target serving as the insulating glove and the human handg2、ph2Obtaining the probability p that the image to be identified is an insulating glove or a human handg、ph：
pg＝θnn·pg1+θsvm·pg2
ph＝θnn·ph1+θsvm·ph2
Wherein, thetannAnd thetasvmPreset weights, p, for the output probabilities of the deep neural network model and the support vector machine, respectivelyg1+ph1＝pg2+ph2＝θnn+θsvm＝1；
If soP isg＞phAnd p isg＞psThen the target is considered to be an insulating glove, wherein psIs a preset threshold.
As a preferred embodiment, in the first aspect of the embodiment of the present invention, the extracting features of all images in the second training data set, and using the extracted features as input information of a support vector machine includes:
extracting RGB information of all images in the second training data set, and calculating a color distribution histogram of each image based on the RGB information;
and constructing a color distribution vector, and using the color distribution vector as input information of a support vector machine.
As a preferred embodiment, in the first aspect of the embodiment of the present invention, before extracting RGB information of all images in the second training data set, the method further includes:
and uniformly scaling the cut images to the same size.
As a preferred embodiment, in the first aspect of the embodiments of the present invention, constructing a color distribution vector includes:
r, G, B color distribution vectors of three channels are respectively constructed, the dimensionality of the color distribution vector constructed by each channel is a 256 multiplied by 1 vector, the stored information of each dimension is the number of pixel points of corresponding colors in a color distribution histogram, and the dimensionalities of the color distribution vectors of the three channels are spliced to obtain the color distribution vector with the dimensionality of 768 multiplied by 1.
As a preferred embodiment, in the first aspect of the embodiment of the present invention, after the constructing the color distribution vector, the method further includes:
and performing dimensionality reduction operation on the color distribution vector with the dimensionality of 768 multiplied by 1.
As a preferred embodiment, in the first aspect of the embodiment of the present invention, the performing the dimensionality reduction operation on the color distribution vector with the dimensionality of 768 × 1 includes:
color feature vector [ v ] for any one of the images in the second training data set1 v2 … v768]And (3) performing decentralized operation to obtain decentralized color feature vectors:
[x1 x2 … x768]＝[v1-μ v2-μ … v768-μ]
wherein mu is the mean value of the two,
obtaining a covariance matrix d (x):
where ω is a feature vector, ωTω＝1，
λiThe eigenvalues of the ith de-centered color eigenvector;
decomposing eigenvalues of the covariance matrix D (x), and sequencing the obtained eigenvalues from small to large;
determining the eigenvector omega corresponding to the first d ordered eigenvalues1,ω2,…,ωd；
Obtaining the color feature vector x 'after dimensionality reduction'i：
Wherein d is more than or equal to 1 and less than or equal to 768.
The second aspect of the embodiment of the invention discloses an electric power operation field insulating glove recognition device based on deep learning, which comprises:
the acquisition unit is used for acquiring pictures containing the information that the electric power working personnel wear the insulating gloves and the normal hands, manually marking the pictures and constructing a first training data set;
the first training unit is used for training a deep neural network initial model for detecting the insulating gloves and the human hands by using the first training data set to obtain a deep neural network model;
the cutting unit is used for cutting all the insulating gloves and hand information in the first training data set according to the manually marked content to construct a second training data set;
the second training unit is used for extracting the features of all the images in the second training data set, and taking the extracted features as input information of a support vector machine, wherein the type of the images belongs to is output information of the support vector machine, so as to train the support vector machine, and the type of the images belongs to is an insulating glove or a human hand;
a recognition unit for determining position information of a target of an image to be recognized and a probability p of the target being an insulating glove and a human hand using the deep neural network modelg1、ph1Cutting the image to be recognized according to the position information, inputting the cut image to be recognized into a support vector machine, and obtaining the probability p of the target serving as the insulating glove and the human handg2、ph2Obtaining the probability p that the image to be identified is an insulating glove or a human handg、ph：
pg＝θnn·pg1+θsvm·pg2
ph＝θnn·ph1+θsvm·ph2
Wherein, thetannAnd thetasvmPreset weights, p, for the output probabilities of the deep neural network model and the support vector machine, respectivelyg1+ph1＝pg2+ph2＝θnn+θsvm＝1；
If said p isg＞phAnd p isg＞psThen the target is considered to be an insulating glove, wherein psIs a preset threshold.
As a preferred embodiment, in the second aspect of the embodiment of the present invention, the second training unit includes:
the zooming subunit is used for zooming the clipped images to the same size in a unified manner;
an extraction subunit, configured to extract RGB information of all images in the second training data set, and calculate a color distribution histogram of each image based on the RGB information;
the building subunit is used for respectively building R, G, B color distribution vectors of three channels, the dimensionality of the color distribution vector built by each channel is a 256 × 1 vector, the stored information of each dimension is the number of pixel points of corresponding colors in a color distribution histogram, and the dimensionalities of the color distribution vectors of the three channels are spliced to obtain the color distribution vector with the dimensionality of 768 × 1;
and the dimensionality reduction subunit is used for performing dimensionality reduction operation on the color distribution vector with the dimensionality of 768 multiplied by 1.
A third aspect of an embodiment of the present invention discloses an electronic device, including: a memory storing executable program code; a processor coupled with the memory; the processor calls the executable program code stored in the memory for executing the deep learning-based electric power operation field insulating glove identification method disclosed by the first aspect of the embodiment of the invention.
A fourth aspect of the embodiments of the present invention discloses a computer-readable storage medium storing a computer program, wherein the computer program causes a computer to execute the method for identifying an insulating glove in an electric power working site based on deep learning disclosed in the first aspect of the embodiments of the present invention.
The fifth aspect of the embodiment of the present invention discloses a computer program product, which, when running on a computer, causes the computer to execute the method for identifying the insulated gloves in the electric power working field based on deep learning disclosed in the first aspect of the embodiment of the present invention.
The sixth aspect of the embodiment of the present invention discloses an application publishing platform, where the application publishing platform is configured to publish a computer program product, where when the computer program product runs on a computer, the computer is enabled to execute the method for identifying insulating gloves in an electric power operation field based on deep learning disclosed in the first aspect of the embodiment of the present invention.
Compared with the prior art, the embodiment of the invention has the following beneficial effects:
the embodiment of the invention realizes the identification of the insulating gloves by combining the deep neural network and the support vector machine, and effectively improves the efficiency of the identification of the insulating gloves.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, and it is apparent that the drawings in the following description are only some embodiments of the present invention, and for a person of ordinary skill in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a schematic flow chart of a deep learning-based method for identifying insulating gloves in an electric power operation field, disclosed by an embodiment of the invention;
FIG. 2 is a schematic structural diagram of an electric power operation field insulating glove recognition device based on deep learning according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.
It should be noted that the terms "first", "second", "third", "fourth", and the like in the description and the claims of the present invention are used for distinguishing different objects, and are not used for describing a specific order. The terms "comprises," "comprising," and any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed, but may include other steps or elements not expressly listed or inherent to such process, method, article, or apparatus.
The embodiment of the invention discloses a method and a device for identifying insulating gloves in an electric power operation site based on deep learning, which can improve the efficiency of image identification and are described in detail in the following with reference to the attached drawings.
Example one
The method is characterized in that an electric power operator is required to wear insulating gloves during operation, in order to monitor whether the electric power operator wears the insulating gloves normally, a target detection technology in computer vision is applied, a deep neural network of the insulating gloves and human hands is detected through training, meanwhile, an operation field image of the electric power operator is collected, the image is placed in the neural network for recognition, and a corresponding reasoning result is given. And if the neural network detects that the hands of the workers appear, judging that the workers do not correctly wear the insulating gloves. But is limited by the recognition accuracy of the neural network, and misjudgment often occurs. The invention provides a method and a device for identifying insulating gloves on an electric power operation site based on deep learning.
Referring to fig. 1, fig. 1 is a schematic flow chart of a method for identifying an insulating glove in an electric power operation field based on deep learning according to an embodiment of the present invention. As shown in fig. 1, the method for identifying insulating gloves in an electric power working site based on deep learning comprises the following steps:
s110, collecting pictures containing the information that the electric power working personnel wear the insulating gloves and have normal hands, manually labeling the pictures, and constructing a first training data set.
And the data set of the training sample, namely the first training data set, is used for training the deep learning model and the support vector machine so as to obtain a corresponding recognition model. The collected pictures mainly comprise two types, one type is pictures of the insulating gloves worn by the power operators, the other type is pictures containing normal hand information, and for the first type, the pictures of the insulating gloves which are suitable for any color and shape are preferably selected, so that the pictures of the insulating gloves with each color and shape are obtained as much as possible in the collection process. The second type of picture may be normal hand information of any person, and it is preferable to collect hand information of the power worker.
The collected images are manually marked, the manual marking information mainly comprises type information and outline information, wherein the type information (the type information to which the images belong) is insulating gloves or human hands and can be only one of the insulating gloves or the human hands, and the outline information refers to the outline corresponding to the type information, namely the outline of the insulating gloves or the human hands, drawn on the basis of the corresponding type information.
And S120, training the initial deep neural network model for detecting the insulating gloves and the human hands by using the first training data set to obtain the deep neural network model.
The deep learning model can be any existing deep learning network, and illustratively, a deep neural network, such as a YOLOv5 algorithm model, can be adopted, and the algorithm has high recognition speed and recognition accuracy and is suitable for the requirement of high real-time performance of the electric power operation site. And training the YOLOv5 neural network initial model by using the existing first training data set to obtain a final deep neural network model.
The method of inspecting insulating gloves and human hands, i.e. hands, is an object detection (ObjectDetection) technique in computer vision technology, the task of which is to find all objects of interest (objects) in an image, while determining their category and position.
S130, according to the manually marked content, cutting all the insulating gloves and hand information in the first training data set to construct a second training data set.
According to the contour information in the manual labeling information, cutting the contour of the insulating glove or the human hand contained in each picture in each first training data set to obtain new pictures, wherein the number of the new pictures is the same as that of the pictures in the first training data set; a second training data set is constructed based on these new pictures. The manner of clipping may be by means of OpenCV techniques.
And S140, extracting the features of all the images in the second training data set, and using the extracted features as input information of a support vector machine, wherein the type of the images belongs to is the output information of the support vector machine, so as to train the support vector machine, and the type of the images belongs to is insulating gloves or human hands.
In order to ensure the uniformity of the cut images, all the images (called RGB images) in the second training data set after cutting are uniformly scaled to w × h size, wherein w is the width of the images, and h is the height of the images. It should be noted here that the scaling includes two cases of scaling up or scaling down, so that the sizes of all RGB images are the same, the number of pixel points is the same, in the scaling up process, the manner of adding pixel points may be interpolation, etc., and in the scaling down process, the equidistant sampling method or the local averaging method, etc. may be adopted.
The zoomed image has three channels of RGB, and color distribution histograms are calculated for the three channels respectively, namely the horizontal axis represents the depth of color on the channel, and the vertical axis represents the number of pixel points with corresponding color depth in the image. And calculating the color distribution condition of the image through the color division histogram.
For an RGB image, the color depth of each channel is represented by 256 values which are 0-255, a vector with the dimension of 256 multiplied by 1 is constructed for a single channel, the information stored in each dimension is the number of pixel points of corresponding colors in a color distribution histogram, and the vectors obtained by the three channels are spliced to obtain the vector with the dimension of 768 multiplied by 1.
In order to improve the accuracy and the calculation efficiency of the support vector machine, the vectors with the dimensionality of 768 × 1 need to be reduced. There are various ways of reducing the dimensions, and for example, PCA (principal component analysis), LLE (local linear embedding), random forest/combination, LDA (linear discriminant analysis), and the like can be used.
In the preferred embodiment of the present invention, the dimensionality reduction operation is implemented by using a linear PCA algorithm, which specifically includes the following steps:
A. the dimension of m images is 768 multiplied by 1 color feature vectors, and any one color feature vector is [ v1 v2 … v768]The color feature vector is de-centered, i.e. [ x ]1 x2 … x768]＝[v1-μ v2-μ … v768-μ]Wherein
B. Finding out a direction omega dispersed as much as possible after the original characteristic points are mapped, wherein omega is a characteristic vector omega corresponding to characteristic valuesTω is 1, i.e. the variance of the projection of the original feature point on ω is made as large as possible, i.e.:
for the last term of the above equation, the covariance matrix of the samples is solved.
C. And carrying out eigenvalue decomposition on the covariance matrix of the sample, and sequencing the obtained eigenvalues from large to small.
D. The eigenvector omega corresponding to the front d of the obtained eigenvalue1,ω2,…,ωdBy passing
Obtaining the color feature vector x 'after dimensionality reduction'i：
Wherein d is more than or equal to 1 and less than or equal to 768, mapping the 768-dimensional samples into d-dimensional samples, and mapping the new x'iThe color feature information after dimension reduction can be obtained by selecting the maximum feature vector as the color feature vector after the final dimension reduction.
Inputting the color feature vector subjected to dimensionality reduction into a support vector machine (the support vector machine at the moment is an initial support vector machine), taking the type information of each RGB image, namely the information of the insulating glove or the human hand, as the output information of the initial support vector machine, training the initial support vector machine, and obtaining a support vector machine model capable of finally identifying the insulating glove and the human hand.
S150, respectively determining the probability of the image to be recognized as the insulating glove and the human hand by using the deep neural network model and the support vector machine model.
The image to be recognized is a related image of the power operator captured in the power operation site, and may be an entire image of the power operator or a partial image including a hand region of the power operator. Preferably adapted to the picture of the first training data set.
Through the trained deep neural network model, the position information of a target (an insulating glove or a human hand) in the image to be recognized and the probability p of the target serving as the insulating glove or the human hand can be obtainedg1、ph1。
Then, the target is cut according to the position information of the target to obtain a cut image, the cut image is correspondingly zoomed (the size of the cut image is w multiplied by h), and then a support vector machine model is input to obtain the probability p that the target is used as an insulating glove and a human handg2、ph2Then, the probability p of the image to be recognized being an insulating glove and a human hand is calculatedg、ph：
pg＝θnn·pg1+θsvm·pg2
ph＝θnn·ph1+θsvm·ph2
Wherein, thetannAnd thetasvmPreset weights, p, for the output probabilities of the deep neural network model and the support vector machine, respectivelyg1+ph1＝pg2+ph2＝θnn+θsvm＝1；
If said p isg＞phIf the object is an insulating glove, it is confirmed that the power operator wears the insulating glove during the field operation, and the operation is in accordance withOtherwise, the target is regarded as a hand, and the power operator does not meet the operation requirement.
In order to make the identification more accurate, in a preferred embodiment of the present invention, a preset threshold p may be further setsWhen p isg＞phAnd p isg＞psThen, the target is considered as an insulating glove, and the operation requirement is met.
In conclusion, by implementing the embodiment of the invention, the defect that the deep learning neural network frequently makes misjudgment on the insulating gloves and the hands can be overcome, and the accuracy of identification can be effectively improved by performing secondary feature extraction on the identified content.
Example two
Referring to fig. 2, fig. 2 is a schematic structural diagram of an electric power operation field insulating glove recognition apparatus based on deep learning according to an embodiment of the present invention. As shown in fig. 2, the deep learning-based electric power working site insulating glove recognition apparatus may include:
the acquisition unit 210 is used for acquiring pictures containing insulating gloves worn by power operators and normal hand information, manually labeling the pictures and constructing a first training data set;
a first training unit 220, configured to train a deep neural network initial model for detecting the insulated glove and the human hand using the first training data set to obtain a deep neural network model;
the cutting unit 230 is used for cutting all the insulating gloves and hand information in the first training data set according to the manually marked content to construct a second training data set;
a second training unit 240, configured to extract features of all images in the second training data set, and train the support vector machine by using the extracted features as input information of the support vector machine, where a type of the image belongs to is output information of the support vector machine, and the type of the image belongs to is an insulating glove or a human hand;
a recognition unit 250 for determining position information of a target of an image to be recognized using the deep neural network modelProbability p of the target being an insulating glove and a human handg1、ph1Cutting the image to be recognized according to the position information, inputting the cut image to be recognized into a support vector machine, and obtaining the probability p of the target serving as the insulating glove and the human handg2、ph2Obtaining the probability p that the image to be identified is an insulating glove or a human handg、ph：
pg＝θnn·pg1+θsvm·pg2
ph＝θnn·ph1+θsvm·ph2
Wherein, thetannAnd thetasvmPreset weights, p, for the output probabilities of the deep neural network model and the support vector machine, respectivelyg1+ph1＝pg2+ph2＝θnn+θsvm＝1；
If said p isg＞phAnd p isg＞psThen the target is considered to be an insulating glove, wherein psIs a preset threshold.
Preferably, the second training unit comprises:
the zooming subunit is used for zooming the clipped images to the same size in a unified manner;
an extraction subunit, configured to extract RGB information of all images in the second training data set, and calculate a color distribution histogram of each image based on the RGB information;
the building subunit is used for respectively building R, G, B color distribution vectors of three channels, the dimensionality of the color distribution vector built by each channel is a 256 × 1 vector, the stored information of each dimension is the number of pixel points of corresponding colors in a color distribution histogram, and the dimensionalities of the color distribution vectors of the three channels are spliced to obtain the color distribution vector with the dimensionality of 768 × 1;
the dimensionality reduction subunit is used for performing dimensionality reduction operation on the color distribution vector with the dimensionality of 768 multiplied by 1;
and the training subunit is used for taking the color distribution vector subjected to the dimensionality reduction as input information of the support vector machine, taking type information corresponding to the color distribution vector as output information of the support vector machine, and training the support vector machine to obtain a final support vector machine model capable of identifying the insulating gloves and human hands.
EXAMPLE III
Referring to fig. 3, fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the disclosure. As shown in fig. 3, the electronic device may include:
a memory 310 storing executable program code;
a processor 320 coupled to the memory 310;
the processor 320 calls the executable program code stored in the memory 310 to execute part or all of the steps in the deep learning-based power job site insulating glove identification method in the first embodiment.
The embodiment of the invention discloses a computer-readable storage medium which stores a computer program, wherein the computer program enables a computer to execute part or all of the steps in the deep learning-based power operation field insulating glove identification method in the first embodiment.
The embodiment of the invention also discloses a computer program product, wherein when the computer program product runs on a computer, the computer is enabled to execute part or all of the steps in the deep learning-based power operation field insulating glove identification method in the first embodiment.
The embodiment of the invention also discloses an application publishing platform, wherein the application publishing platform is used for publishing the computer program product, and when the computer program product runs on a computer, the computer is enabled to execute part or all of the steps in the deep learning-based electric power operation field insulating glove identification method in the first embodiment.
In various embodiments of the present invention, it should be understood that the sequence numbers of the processes do not mean the execution sequence necessarily in order, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.
The units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiment.
In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
The integrated units, if implemented as software functional units and sold or used as a stand-alone product, may be stored in a computer accessible memory. Based on such understanding, the technical solution of the present invention, which is a part of or contributes to the prior art in essence, or all or part of the technical solution, can be embodied in the form of a software product, which is stored in a memory and includes several requests for causing a computer device (which may be a personal computer, a server, a network device, or the like, and may specifically be a processor in the computer device) to execute part or all of the steps of the method according to the embodiments of the present invention.
In the embodiments provided herein, it should be understood that "B corresponding to a" means that B is associated with a from which B can be determined. It should also be understood, however, that determining B from a does not mean determining B from a alone, but may also be determined from a and/or other information.
It will be understood by those of ordinary skill in the art that some or all of the steps of the methods of the embodiments may be implemented by instructions associated with a program, which may be stored in a computer-readable storage medium, such as a Read-Only Memory (ROM), a Random Access Memory (RAM), a Programmable Read-Only Memory (PROM), an Erasable Programmable Read-Only Memory (EPROM), a One-time Programmable Read-Only Memory (OTPROM), an Electrically Erasable Programmable Read-Only Memory (EEPROM), a Compact Disc-Read-Only Memory (CD-ROM), or other Memory, a CD-ROM, or other disk, or a combination thereof, A tape memory, or any other medium readable by a computer that can be used to carry or store data.
The method and the device for identifying the insulating gloves on the electric power operation site based on deep learning disclosed by the embodiment of the invention are described in detail, a specific example is applied in the method for explaining the principle and the implementation mode of the invention, and the description of the embodiment is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (10)
1. A deep learning-based method for identifying insulating gloves in an electric power operation site is characterized by comprising the following steps:
collecting pictures containing insulating gloves worn by power operators and normal hand information, manually labeling the pictures, and constructing a first training data set;
training a deep neural network initial model for detecting the insulating gloves and the human hands by using the first training data set to obtain a deep neural network model;
cutting all the insulating gloves and hand information in the first training data set according to the manually marked content to construct a second training data set;
extracting the features of all the images in the second training data set, and using the extracted features as input information of a support vector machine, wherein the type of the images belongs to is output information of the support vector machine, so as to train the support vector machine, and the type of the images belongs to is an insulating glove or a hand;
determining position information of a target of an image to be recognized and a probability p of the target being an insulating glove and a human hand using the deep neural network modelg1、ph1Cutting the image to be recognized according to the position information, inputting the cut image to be recognized into a support vector machine, and obtaining the probability p of the target serving as the insulating glove and the human handg2、ph2Obtaining the probability p that the image to be identified is an insulating glove or a human handg、ph：
pg＝θnn·pg1+θsvm·pg2
ph＝θnn·ph1+θsvm·ph2
Wherein, thetannAnd thetasvmPreset weights, p, for the output probabilities of the deep neural network model and the support vector machine, respectivelyg1+ph1＝pg2+ph2＝θnn+θsvm＝1；
If said p isg＞phAnd p isg＞psThen the target is considered to be an insulating glove, wherein psIs a preset threshold.
2. The deep learning-based power job site insulating glove recognition method according to claim 1, wherein the extracting features of all images in the second training data set and using the extracted features as input information of a support vector machine comprises:
extracting RGB information of all images in the second training data set, and calculating a color distribution histogram of each image based on the RGB information;
and constructing a color distribution vector, and using the color distribution vector as input information of a support vector machine.
3. The deep learning-based power job site insulating glove recognition method according to claim 2, wherein before extracting the RGB information of all the images in the second training data set, further comprising:
and uniformly scaling the cut images to the same size.
4. The deep learning-based electric power job site insulating glove identification method according to claim 2, wherein constructing a color distribution vector comprises:
r, G, B color distribution vectors of three channels are respectively constructed, the dimensionality of the color distribution vector constructed by each channel is a 256 multiplied by 1 vector, the stored information of each dimension is the number of pixel points of corresponding colors in a color distribution histogram, and the dimensionalities of the color distribution vectors of the three channels are spliced to obtain the color distribution vector with the dimensionality of 768 multiplied by 1.
5. The deep learning-based electric power job site insulating glove identification method according to claim 4, wherein the constructing a color distribution vector further comprises:
and performing dimensionality reduction operation on the color distribution vector with the dimensionality of 768 multiplied by 1.
6. The deep learning-based electric power working field insulating glove identification method according to claim 5, wherein the dimensionality reduction operation on the 768 x 1 color distribution vector comprises:
color feature vector [ v ] for any one of the images in the second training data set1 v2…v768]And (3) performing decentralized operation to obtain decentralized color feature vectors:
[x1 x2…x768]＝[v1-μ v2-μ…v768-μ]
wherein mu is the mean value of the two,
obtaining a covariance matrix d (x):
where ω is a feature vector, ωTω＝1，
λiThe eigenvalues of the ith de-centered color eigenvector;
decomposing eigenvalues of the covariance matrix D (x), and sequencing the obtained eigenvalues from small to large;
determining the eigenvector omega corresponding to the first d ordered eigenvalues1,ω2,…,ωd；
Obtaining the color feature vector x 'after dimensionality reduction'i：
Wherein d is more than or equal to 1 and less than or equal to 768.
7. The utility model provides an on-spot insulating gloves recognition device of electric power operation based on degree of depth study which characterized in that, it includes:
the acquisition unit is used for acquiring pictures containing the information that the electric power working personnel wear the insulating gloves and the normal hands, manually marking the pictures and constructing a first training data set;
the first training unit is used for training a deep neural network initial model for detecting the insulating gloves and the human hands by using the first training data set to obtain a deep neural network model;
the cutting unit is used for cutting all the insulating gloves and hand information in the first training data set according to the manually marked content to construct a second training data set;
the second training unit is used for extracting the features of all the images in the second training data set, and taking the extracted features as input information of a support vector machine, wherein the type of the images belongs to is output information of the support vector machine, so as to train the support vector machine, and the type of the images belongs to is an insulating glove or a human hand;
a recognition unit for determining position information of a target of an image to be recognized and a probability p of the target being an insulating glove and a human hand using the deep neural network modelg1、ph1Cutting the image to be recognized according to the position information, inputting the cut image to be recognized into a support vector machine, and obtaining the probability p of the target serving as the insulating glove and the human handg2、ph2Obtaining the probability p that the image to be identified is an insulating glove or a human handg、ph：
pg＝θnn·pg1+θsvm·pg2
ph＝θnn·ph1+θsvm·ph2
Wherein, thetannAnd thetasvmPreset weights, p, for the output probabilities of the deep neural network model and the support vector machine, respectivelyg1+ph1＝pg2+ph2＝θnn+θsvm＝1；
If said p isg＞phAnd p isg＞psThen the target is considered to be an insulating glove, wherein psIs a preset threshold.
8. The deep learning-based electric power job site insulating glove recognition apparatus according to claim 7, wherein the second training unit comprises:
the zooming subunit is used for zooming the clipped images to the same size in a unified manner;
an extraction subunit, configured to extract RGB information of all images in the second training data set, and calculate a color distribution histogram of each image based on the RGB information;
the building subunit is used for respectively building R, G, B color distribution vectors of three channels, the dimensionality of the color distribution vector built by each channel is a 256 × 1 vector, the stored information of each dimension is the number of pixel points of corresponding colors in a color distribution histogram, and the dimensionalities of the color distribution vectors of the three channels are spliced to obtain the color distribution vector with the dimensionality of 768 × 1;
and the dimensionality reduction subunit is used for performing dimensionality reduction operation on the color distribution vector with the dimensionality of 768 multiplied by 1.
9. An electronic device, comprising: a memory storing executable program code; a processor coupled with the memory; the processor calls the executable program code stored in the memory for executing a deep learning based power job site insulating glove identification method according to any one of claims 1 to 6.
10. A computer-readable storage medium storing a computer program, wherein the computer program causes a computer to execute a deep learning-based power job site insulating glove recognition method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111544435.9A CN114359792B (en) | 2021-12-16 | Electric power operation site insulation glove identification method and device based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111544435.9A CN114359792B (en) | 2021-12-16 | Electric power operation site insulation glove identification method and device based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114359792A true CN114359792A (en) | 2022-04-15 |
| CN114359792B CN114359792B (en) | 2024-08-02 |
Family
ID=
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115471874A (en) * | 2022-10-28 | 2022-12-13 | 山东新众通信息科技有限公司 | Construction site dangerous behavior identification method based on monitoring video |
| CN115830635A (en) * | 2022-12-09 | 2023-03-21 | 南通大学 | PVC glove identification method based on key point detection and target identification |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109344905A (en) * | 2018-10-22 | 2019-02-15 | 王子蕴 | A kind of transmission facility automatic fault recognition methods based on integrated study |
| US20190258878A1 (en) * | 2018-02-18 | 2019-08-22 | Nvidia Corporation | Object detection and detection confidence suitable for autonomous driving |
| CN113033328A (en) * | 2021-03-05 | 2021-06-25 | 杭州追猎科技有限公司 | Personnel mask wearing state detection and identification method based on deep learning |
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190258878A1 (en) * | 2018-02-18 | 2019-08-22 | Nvidia Corporation | Object detection and detection confidence suitable for autonomous driving |
| CN109344905A (en) * | 2018-10-22 | 2019-02-15 | 王子蕴 | A kind of transmission facility automatic fault recognition methods based on integrated study |
| CN113033328A (en) * | 2021-03-05 | 2021-06-25 | 杭州追猎科技有限公司 | Personnel mask wearing state detection and identification method based on deep learning |
Non-Patent Citations (2)
| Title |
|---|
| BAINING ZHAO 等: "Detection and Location of Safety Protective Wear in Power Substation Operation Using Wear-enhanced YOLOv3 Algorithm", 《IEEE ACCESS》, vol. 4, 13 August 2021 (2021-08-13), pages 1 - 10 * |
| 杨莉琼 等: "基于机器学习方法的安全帽佩戴行为检测", 《中国安全生产科学技术》, vol. 15, no. 10, 30 October 2019 (2019-10-30), pages 152 - 157 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115471874A (en) * | 2022-10-28 | 2022-12-13 | 山东新众通信息科技有限公司 | Construction site dangerous behavior identification method based on monitoring video |
| CN115471874B (en) * | 2022-10-28 | 2023-02-07 | 山东新众通信息科技有限公司 | Construction site dangerous behavior identification method based on monitoring video |
| CN115830635A (en) * | 2022-12-09 | 2023-03-21 | 南通大学 | PVC glove identification method based on key point detection and target identification |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105740780B (en) | Method and device for detecting living human face | |
| CN109214280B (en) | Shop identification method and device based on street view, electronic equipment and storage medium | |
| CN108073855B (en) | Facial expression recognition method and system | |
| US10395091B2 (en) | Image processing apparatus, image processing method, and storage medium identifying cell candidate area | |
| CN109977253B (en) | Semantic and content-based rapid image retrieval method and device | |
| CN111178252A (en) | Multi-feature fusion identity recognition method | |
| CN111275040A (en) | Positioning method and device, electronic equipment and computer readable storage medium | |
| CN112926601A (en) | Image recognition method, device and equipment based on deep learning and storage medium | |
| CN105246149B (en) | Geographical position identification method and device | |
| CN109919128B (en) | Control instruction acquisition method and device and electronic equipment | |
| CN114882010A (en) | Surface defect detection method based on picture recognition | |
| CN113033488B (en) | Medical feature recognition method and device, electronic equipment and storage medium | |
| CN108537990B (en) | All-in-one machine fraud judgment method, device, equipment and computer readable storage medium | |
| CN114219936A (en) | Object detection method, electronic device, storage medium, and computer program product | |
| CN113888496A (en) | Sperm morphology recognition method based on deep convolutional neural network and storage medium | |
| CN107368847B (en) | Crop leaf disease identification method and system | |
| CN104504692B (en) | The extracting method of notable object in image based on region contrast | |
| CN111582278B (en) | Portrait segmentation method and device and electronic equipment | |
| CN114359792B (en) | Electric power operation site insulation glove identification method and device based on deep learning | |
| CN114359792A (en) | Deep learning-based method and device for identifying insulating gloves on electric power operation site | |
| CN114299426A (en) | Electric power operation field work clothes identification method and device based on deep learning | |
| CN115661903A (en) | Map recognizing method and device based on spatial mapping collaborative target filtering | |
| CN113239861B (en) | Method for determining head motion of driver, storage medium, and electronic device | |
| CN115019152A (en) | Image shooting integrity judgment method and device | |
| CN107180244B (en) | Image detection method and device based on cascade classifier |
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 |