CN113052260A - Transformer substation foreign matter identification method and system based on image registration and target detection - Google Patents
Transformer substation foreign matter identification method and system based on image registration and target detection Download PDFInfo
- Publication number
- CN113052260A CN113052260A CN202110430795.XA CN202110430795A CN113052260A CN 113052260 A CN113052260 A CN 113052260A CN 202110430795 A CN202110430795 A CN 202110430795A CN 113052260 A CN113052260 A CN 113052260A
- Authority
- CN
- China
- Prior art keywords
- detected
- image
- target detection
- picture
- foreign matter
- 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.)
- Pending
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 74
- 238000000034 method Methods 0.000 title claims abstract description 41
- 239000013598 vector Substances 0.000 claims description 19
- 230000009466 transformation Effects 0.000 claims description 14
- 238000012545 processing Methods 0.000 claims description 13
- 239000011159 matrix material Substances 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 6
- 230000002159 abnormal effect Effects 0.000 claims description 5
- 230000000694 effects Effects 0.000 abstract description 5
- 239000013589 supplement Substances 0.000 abstract description 3
- 238000012549 training Methods 0.000 description 3
- 241001465754 Metazoa Species 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 235000005770 birds nest Nutrition 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000002265 prevention Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
- 235000005765 wild carrot Nutrition 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/33—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods
- G06T7/337—Determination of transform parameters for the alignment of images, i.e. image registration using feature-based methods involving reference images or patches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/30—Determination of transform parameters for the alignment of images, i.e. image registration
- G06T7/35—Determination of transform parameters for the alignment of images, i.e. image registration using statistical methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Biology (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Biophysics (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Probability & Statistics with Applications (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a transformer substation foreign matter identification method and a transformer substation foreign matter identification system based on image registration and target detection, which belong to the technical field of target detection and comprise the following steps: s1: inputting an image; s2: image registration; s3: searching for inconsistent areas; s4: non-foreign object detection; s5: a foreign matter region is obtained. The method finds out the area which appears at first, judges whether the foreign matters exist in the area or not, determines whether the types of the foreign matters are few or not, determines whether the samples are many, and has good effect on the trained target detection model which is not the foreign matters; in addition, for the detected non-foreign matters, through artificial re-judgment, the filtered non-foreign matter categories can rapidly supplement samples to train a non-foreign matter target detection model, and the whole scheme is easy to optimize and is worthy of popularization and use.
Description
Technical Field
The invention relates to the technical field of target detection, in particular to a transformer substation foreign matter identification method and system based on image registration and target detection.
Background
When the transformer substation normally operates, some objects which do not belong to the transformer substation inherently appear, and the existence of the objects threatens the normal operation of the transformer substation, such as plastic bags hung on a transmission line, various small animals biting the electric wire, short-circuit faults caused by nesting in the transformer substation and the like.
At present, foreign matter detection is carried out on foreign matters possibly existing in a transformer substation by directly using a target detection algorithm. Since foreign objects in a substation are a broad concept, such as plastic bags, birds, bird nests, animals, etc., the categories cannot be exhaustive. The shape of the same foreign matter such as kite, plastic bag, etc. is varied. The target detection algorithm needs to know the class of the target in advance and collect enough samples to use. This clearly deviates from the actual requirement, resulting in many foreign bodies being undetectable. Therefore, a transformer substation foreign matter identification method and system based on image registration and target detection are provided.
Disclosure of Invention
The technical problem to be solved by the invention is as follows: the method marks all objects which do not belong to the transformer substation foreign matter, then adopts the target detection algorithm to filter the target which is not the foreign matter, and the rest is treated as the foreign matter, so that the foreign matter in the transformer substation is detected and positioned, and the method has great significance for early warning and fault tracing of the transformer substation.
The invention solves the technical problems through the following technical scheme, and the invention comprises the following steps:
s1: inputting an image
Inputting an image to be detected, and simultaneously inputting a matching template picture corresponding to the image to be detected;
s2: image registration
Performing image registration processing on the image to be detected and the template picture in the step S1;
s3: finding inconsistent areas
Calculating the similarity between the to-be-detected picture and the transformed template picture by adopting an SSIM image comparison method, and finding out an inconsistent area of the two pictures;
s4: non-foreign object detection
Detecting inconsistent areas in the picture to be detected by adopting a target detection network;
s5: obtaining a foreign matter region
Based on the detection result of step S4, the recognizable region is removed, and the remaining inconsistent region is determined to be a foreign region.
Preferably, in step S1, the matching template picture is a picture of the substation during normal operation, and the picture to be detected is a picture of which the shooting area is the same as the matching template picture.
Preferably, the specific process in step S2 is as follows:
s21: extracting key points of the two graphs, and then extracting key point descriptors;
s22: after two groups of key points are extracted from the two images, corresponding key points in the two images are associated or matched;
s23: calculating an optimal transformation matrix between the two point sets according to the group of matching point sets obtained in the step S22;
s24: and performing similarity transformation on the matched template picture by using the optimal transformation matrix, wherein the coordinates of the key points of the transformed matched template picture are the same as the coordinates of the corresponding key points in the picture to be detected.
Preferably, in step S21, the keypoint descriptor is a feature vector of a keypoint, which represents information of an image around the keypoint, and the keypoint is a set of a class of points, which represents a point, a line, and edge feature information in the image.
Preferably, in step S22, the specific calculation process is as follows:
s221: calculating the similarity between the feature vector of a certain point in the picture to be detected and the feature vectors of all points in the template picture, selecting the point with the closest distance, and preliminarily obtaining a matching point set corresponding to the two pictures;
s222: and then detecting abnormal values in the matching point set, and removing to obtain a final matching point set.
Preferably, in the step S4, the target detection network is a YOLO-V3 network or a Cascade-Rcnn network, and the trained target detection network can identify a plurality of non-foreign targets.
Preferably, after the detection in step S4, the recognizable target can be found in the inconsistent area by the target detection network and determined as a non-foreign object, and when no target is detected in the inconsistent area, the entire inconsistent area is determined as a foreign object area.
The invention also provides a transformer substation foreign matter identification system based on image registration and target detection, which adopts the identification method to detect and position the foreign matters in the transformer substation and comprises the following steps:
the input module is used for inputting an image to be detected and simultaneously inputting a matching template picture corresponding to the image to be detected;
a registration module, configured to perform image registration processing on the image to be detected and the template picture in step S1;
the searching module is used for calculating the similarity between the to-be-detected picture and the transformed template picture by adopting an SSIM image comparison method, and finding out the inconsistent area of the two pictures;
the non-foreign object detection module is used for detecting inconsistent areas in the picture to be detected by adopting a target detection network;
a foreign matter region determination module for removing the recognizable region according to the detection result of step S4, and determining the remaining inconsistent region as a foreign matter region;
the central processing module is used for sending instructions to other modules to complete related actions;
the input module, the registration module, the searching module, the non-foreign matter detection module and the foreign matter region distinguishing module are all electrically connected with the central processing module.
Compared with the prior art, the invention has the following advantages: according to the transformer substation foreign matter identification method based on image registration and target detection, the appearing region is found out firstly, and the foreign matters are judged in the region, the rest are the foreign matters, the types of the foreign matters are not many, the samples are also many, and the trained target detection model of the foreign matters is good in effect; in addition, for the detected non-foreign matters, through artificial re-judgment, the filtered non-foreign matter categories can rapidly supplement samples to train a non-foreign matter target detection model, and the whole scheme is easy to optimize and is worthy of popularization and use.
Drawings
FIG. 1 is a schematic overall flow chart of a second embodiment of the present invention;
FIG. 2 is a schematic diagram illustrating the registration effect of two images with translation according to a second embodiment of the present invention;
FIG. 3a is a normal picture of the same shooting area and matching template in the second embodiment of the present invention;
FIG. 3b is a diagram of a picture to be detected according to the second embodiment of the present invention;
FIG. 3c is the inconsistency between the areas in FIG. 3a and FIG. 3b found in the second embodiment of the present invention.
Detailed Description
The following examples are given for the detailed implementation and specific operation of the present invention, but the scope of the present invention is not limited to the following examples.
Example one
The embodiment provides a technical scheme: the transformer substation foreign matter identification method based on image registration and target detection comprises the following steps:
s1: inputting an image
Inputting an image to be detected, and simultaneously inputting a matching template picture corresponding to the image to be detected;
s2: image registration
Performing image registration processing on the image to be detected and the template picture in the step S1;
s3: finding inconsistent areas
Calculating the similarity between the to-be-detected picture and the transformed template picture by adopting an SSIM image comparison method, and finding out an inconsistent area of the two pictures;
s4: non-foreign object detection
Detecting inconsistent areas in the picture to be detected by adopting a target detection network;
s5: obtaining a foreign matter region
Based on the detection result of step S4, the recognizable region is removed, and the remaining inconsistent region is determined to be a foreign region.
In this embodiment, in the step S1, the matching template picture is a picture of the substation during normal operation, and the picture to be detected is a picture of which the shooting area is the same as the matching template picture.
In this embodiment, the specific process in step S2 is as follows:
s21: extracting key points of the two graphs, and then extracting key point descriptors;
s22: after two groups of key points are extracted from the two images, corresponding key points in the two images are associated or matched;
s23: calculating an optimal transformation matrix between the two point sets according to the group of matching point sets obtained in the step S22;
s24: and performing similarity transformation on the matched template picture by using the optimal transformation matrix, wherein the coordinates of the key points of the transformed matched template picture are the same as the coordinates of the corresponding key points in the picture to be detected.
In this embodiment, in step S21, the keypoint descriptor is a feature vector of a keypoint, which represents information of an image around the keypoint, and the keypoint is a set of a type of keypoint, which represents information of a point, a line, and an edge feature in the image.
In this embodiment, in step S22, the specific calculation process is as follows:
s221: calculating the similarity between the feature vector of a certain point in the picture to be detected and the feature vectors of all points in the template picture, selecting the point with the closest distance, and preliminarily obtaining a matching point set corresponding to the two pictures.
S222: and then detecting abnormal values in the matching point set, and removing to obtain a final matching point set.
In this embodiment, in the step S4, the target detection network adopts a YOLO-V3 network or a Cascade-Rcnn network, and a plurality of non-foreign objects can be identified by the trained target detection network.
In the present embodiment, after the detection in step S4, the recognizable target can be found in the inconsistent area by the target detection network and is determined as a non-foreign object, and when no target is detected in the inconsistent area, the entire inconsistent area is determined as a foreign object area.
The embodiment also provides a transformer substation foreign matter identification system based on image registration and target detection, and the method for detecting and positioning the foreign matters in the transformer substation comprises the following steps:
the input module is used for inputting an image to be detected and simultaneously inputting a matching template picture corresponding to the image to be detected;
a registration module, configured to perform image registration processing on the image to be detected and the template picture in step S1;
the searching module is used for calculating the similarity between the to-be-detected picture and the transformed template picture by adopting an SSIM image comparison method, and finding out the inconsistent area of the two pictures;
the non-foreign object detection module is used for detecting inconsistent areas in the picture to be detected by adopting a target detection network;
a foreign matter region determination module for removing the recognizable region according to the detection result of step S4, and determining the remaining inconsistent region as a foreign matter region;
the central processing module is used for sending instructions to other modules to complete related actions;
the input module, the registration module, the searching module, the non-foreign matter detection module and the foreign matter region distinguishing module are all electrically connected with the central processing module.
Example two
As shown in fig. 1, the present embodiment provides a transformer substation foreign object identification method based on image registration and target detection, including the following steps:
step 1: inputting an image
Inputting an image to be detected, and simultaneously inputting a matching template corresponding to the image to be detected, wherein the matching template is a picture of the transformer substation in normal operation, and the picture to be detected is a picture with the same shooting area as the matching template;
step 2: image registration
Processing the two pictures in the step 1 by using an image registration algorithm;
the specific process in step S2 is as follows:
s21: extracting key points: it is a collection of points that generally represent characteristic information such as points, lines, edges, etc. in an image. These points can be found in the same area of different images as indicated by the dots in fig. 2. Then extracting a key point descriptor, wherein the descriptor is a feature vector of the key point (when extraction of key points such as SIFT and the like is finished, coordinates of the key point and the corresponding feature vector can be automatically output), the information of images around the key point is represented, and the feature vector can still keep consistency when illumination and visual angle change slightly; common characteristic points are SIFT, SURF, ORB and BRIEF;
s22: and (3) feature matching: after two groups of key points are extracted from the two images, the corresponding key points in the two images need to be associated or matched. As shown in fig. 2, two points connected by a straight line are corresponding matching points. The specific calculation process includes firstly calculating the similarity between the feature vector of a certain point in the picture to be detected and the feature vectors of all points in the template picture, selecting the point with the closest distance (the most similar point), and preliminarily obtaining a matching point set corresponding to the two pictures. Then detecting abnormal values in the matching point set by using a RANSAC algorithm, and removing the abnormal values to obtain a final matching point set;
and (4) similarity calculation, namely calculating the cosine distance of the two characteristic vectors by adopting the pair of characteristic vectors X and Y extracted in the step S21, wherein the distance is larger if the two characteristic vectors are more similar. The cosine distance calculation formula is:
wherein A is a feature vector of one of the two images, B is a feature vector of the other of the two images, and A and B are matched.
S23: calculating a similarity transformation matrix: calculating an optimal transformation matrix M between two point sets by using a minimum mean square error or RANSAC method according to a group of matching point sets obtained by feature matching;
the RANSAC method process is as follows:
1): first, a subset of the data is randomly selected, the subset is assumed to be local points, and a model is fitted with the local points.
2): testing other data in the data by using the model calculated in the step 1), wherein points meeting the model are called local interior points, and the points which are not met are called local exterior points by expanding a local interior point set.
3): if enough points are classified as local points, the estimated model is reasonable and the iteration stops.
4): then, the estimation model is updated by the expanded intra-office point set, and the step 2) is switched.
S24: image transformation: and performing similarity transformation on the template picture by using the optimal transformation matrix, wherein the coordinates of key points of the transformed template picture are the same as the coordinates of corresponding key points in the picture to be detected.
And step 3: and calculating the similarity between the picture to be detected and the transformed template picture by adopting an SSIM (structural similarity) image comparison algorithm, and finding out an inconsistent area of the two pictures.
SSIM algorithm: all of them are called structural similarity index, i.e. the structural similarity measurement technique, and the formula is:
where μ is the mean of the pixels within the image block and σ is the standard deviation of the pixels within the image block. C1And C2Is constant, the prevention denominator is 0.
And 4, step 4: and detecting inconsistent areas in the picture to be detected by adopting a target detection network (target detection model).
The target detection network generally adopts a network such as YOLO-V3, Cascade-Rcnn and the like. Other types of deep learning networks cannot be adopted, and only selection can be carried out in the target detection network. Various targets such as people, vehicles, safety helmets and the like can be identified through the trained target detection network. Only the disclosed target detection network training is needed, and no special improvement is needed.
And 5: and (4) removing the recognizable area according to the recognition result of the step (4), and judging the residual inconsistent area as a foreign matter area.
In this embodiment, the recognizable area is an area that can only be detected by the target detection network, and this part of the area can be directly removed from the area to be detected by setting the corresponding pixel of the image to 0.
Through the step 4, the recognizable target can be found in the inconsistent area through the target detection network, the recognizable target is judged to be a non-foreign object, such as a vehicle and a person, if the recognizable target is a foreign object, such as a kite, and any target cannot be detected in the inconsistent area, the whole inconsistent area is judged to be a foreign object area.
As shown in fig. 3, fig. 3a is a normal picture with the same shooting area as the matching template, fig. 3b is a picture to be detected, and fig. 3c is an inconsistent area in fig. 3a and fig. 3b found by using the structural similarity image comparison algorithm.
To sum up, compared with the prior art, in the transformer substation foreign object identification method based on image registration and target detection in the embodiment, which is the foreign object is known in the existing scheme, and then a large amount of foreign object data is collected to perform model training for use, but the types of the foreign objects cannot be exhausted, and the foreign objects are not frequently appeared (samples are few), which is very bad for training a target detection model and affects the final effect; the method firstly finds out the area which appears, judges whether the area is foreign matter or not, determines whether the type of the foreign matter is small or not, determines whether the number of samples is large, and has good effect on the trained target detection model which is not foreign matter; in addition, for the detected non-foreign matters, through artificial re-judgment, the filtered non-foreign matter categories can rapidly supplement samples to train a non-foreign matter target detection model, and the whole scheme is easy to optimize and is worthy of popularization and use.
Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention, and that variations, modifications, substitutions and alterations can be made to the above embodiments by those of ordinary skill in the art within the scope of the present invention.
Claims (8)
1. The transformer substation foreign matter identification method based on image registration and target detection is characterized by comprising the following steps of:
s1: inputting an image
Inputting an image to be detected, and simultaneously inputting a matching template picture corresponding to the image to be detected;
s2: image registration
Performing image registration processing on the image to be detected and the matched template picture in the step S1;
s3: finding inconsistent areas
Calculating the similarity between the to-be-detected picture and the transformed template picture by adopting an image comparison method, and finding out an inconsistent area of the two pictures;
s4: non-foreign object detection
Detecting inconsistent areas in the picture to be detected by adopting a target detection network;
s5: obtaining a foreign matter region
Based on the detection result of step S4, the recognizable region is removed, and the remaining inconsistent region is determined to be a foreign region.
2. The transformer substation foreign object identification method based on image registration and target detection as claimed in claim 1, wherein: in the step S1, the matching template picture is a picture of the transformer substation during normal operation, and the picture to be detected is a picture of which the shooting area is the same as the matching template picture.
3. The transformer substation foreign object identification method based on image registration and target detection as claimed in claim 1, wherein: the specific process in step S2 is as follows:
s21: extracting key points of the two graphs, and then extracting key point descriptors;
s22: after two groups of key points are extracted from the two images, corresponding key points in the two images are associated or matched;
s23: calculating an optimal transformation matrix between the two point sets according to the group of matching point sets obtained in the step S22;
s24: and performing similarity transformation on the matched template picture by using the optimal transformation matrix, wherein the coordinates of the key points of the transformed matched template picture are the same as the coordinates of the corresponding key points in the picture to be detected.
4. The transformer substation foreign object identification method based on image registration and target detection as claimed in claim 3, wherein: in step S21, the keypoint descriptor is a feature vector of a keypoint, which represents information of an image around the keypoint, and the keypoint is a set of a type of point, which represents a point, a line, and edge feature information in the image.
5. The transformer substation foreign object identification method based on image registration and target detection as claimed in claim 3, wherein: in step S22, the specific calculation process is as follows:
s221: calculating the similarity between the feature vector of a certain point in the picture to be detected and the feature vectors of all points in the template picture, selecting the point with the closest distance, and preliminarily obtaining a matching point set corresponding to the two pictures;
s222: and then detecting abnormal values in the matching point set, and removing to obtain a final matching point set.
6. The transformer substation foreign object identification method based on image registration and target detection as claimed in claim 1, wherein: in the step S4, the target detection network uses a YOLO-V3 network or a Cascade-Rcnn network, and identifies a plurality of non-foreign objects through the trained target detection network.
7. The transformer substation foreign object identification method based on image registration and target detection as claimed in claim 6, wherein: after the detection in step S4, the recognizable target can be found in the inconsistent area by the target detection network and is determined as a non-foreign object, and when no target is detected in the inconsistent area, the entire inconsistent area is determined as a foreign object area.
8. The transformer substation foreign matter identification system based on image registration and target detection is characterized in that the foreign matter in the transformer substation is detected and positioned by adopting the identification method of any one of claims 1-7, and the method comprises the following steps:
the input module is used for inputting an image to be detected and simultaneously inputting a matching template picture corresponding to the image to be detected;
a registration module, configured to perform image registration processing on the image to be detected and the template picture in step S1;
the searching module is used for calculating the similarity between the to-be-detected picture and the transformed template picture by adopting an image comparison method and finding out the inconsistent area of the two pictures;
the non-foreign object detection module is used for detecting inconsistent areas in the picture to be detected by adopting a target detection network;
a foreign matter region determination module for removing the recognizable region according to the detection result of step S4, and determining the remaining inconsistent region as a foreign matter region;
the central processing module is used for sending instructions to other modules to complete related actions;
the input module, the registration module, the searching module, the non-foreign matter detection module and the foreign matter region distinguishing module are all electrically connected with the central processing module.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110430795.XA CN113052260A (en) | 2021-04-21 | 2021-04-21 | Transformer substation foreign matter identification method and system based on image registration and target detection |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110430795.XA CN113052260A (en) | 2021-04-21 | 2021-04-21 | Transformer substation foreign matter identification method and system based on image registration and target detection |
Publications (1)
Publication Number | Publication Date |
---|---|
CN113052260A true CN113052260A (en) | 2021-06-29 |
Family
ID=76519971
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110430795.XA Pending CN113052260A (en) | 2021-04-21 | 2021-04-21 | Transformer substation foreign matter identification method and system based on image registration and target detection |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113052260A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114092478A (en) * | 2022-01-21 | 2022-02-25 | 合肥中科类脑智能技术有限公司 | Anomaly detection method |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080144972A1 (en) * | 2006-11-09 | 2008-06-19 | University Of Delaware | Geometric registration of images by similarity transformation using two reference points |
WO2010053109A1 (en) * | 2008-11-10 | 2010-05-14 | 日本電気株式会社 | Image matching device, image matching method, and image matching program |
CN102722731A (en) * | 2012-05-28 | 2012-10-10 | 南京航空航天大学 | Efficient image matching method based on improved scale invariant feature transform (SIFT) algorithm |
CN108154496A (en) * | 2017-11-21 | 2018-06-12 | 武汉中元华电科技股份有限公司 | A kind of power equipment appearance suitable for electric operating robot changes recognition methods |
CN109636848A (en) * | 2018-12-17 | 2019-04-16 | 武汉天乾科技有限责任公司 | A kind of oil-gas pipeline method for inspecting based on unmanned plane |
CN111680609A (en) * | 2020-06-03 | 2020-09-18 | 合肥中科类脑智能技术有限公司 | Foreign matter identification system and method based on image registration and target detection |
CN112308887A (en) * | 2020-09-30 | 2021-02-02 | 西北工业大学 | Real-time registration method for multi-source image sequence |
-
2021
- 2021-04-21 CN CN202110430795.XA patent/CN113052260A/en active Pending
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20080144972A1 (en) * | 2006-11-09 | 2008-06-19 | University Of Delaware | Geometric registration of images by similarity transformation using two reference points |
WO2010053109A1 (en) * | 2008-11-10 | 2010-05-14 | 日本電気株式会社 | Image matching device, image matching method, and image matching program |
CN102722731A (en) * | 2012-05-28 | 2012-10-10 | 南京航空航天大学 | Efficient image matching method based on improved scale invariant feature transform (SIFT) algorithm |
CN108154496A (en) * | 2017-11-21 | 2018-06-12 | 武汉中元华电科技股份有限公司 | A kind of power equipment appearance suitable for electric operating robot changes recognition methods |
CN109636848A (en) * | 2018-12-17 | 2019-04-16 | 武汉天乾科技有限责任公司 | A kind of oil-gas pipeline method for inspecting based on unmanned plane |
CN111680609A (en) * | 2020-06-03 | 2020-09-18 | 合肥中科类脑智能技术有限公司 | Foreign matter identification system and method based on image registration and target detection |
CN112308887A (en) * | 2020-09-30 | 2021-02-02 | 西北工业大学 | Real-time registration method for multi-source image sequence |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114092478A (en) * | 2022-01-21 | 2022-02-25 | 合肥中科类脑智能技术有限公司 | Anomaly detection method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109118479B (en) | Capsule network-based insulator defect identification and positioning device and method | |
CN109101924B (en) | Machine learning-based road traffic sign identification method | |
US10445602B2 (en) | Apparatus and method for recognizing traffic signs | |
JP2022537857A (en) | Automatic Determination System and Method for Degree of Damage by Automobile Parts Based on Deep Learning | |
CN111784633B (en) | Insulator defect automatic detection algorithm for electric power inspection video | |
CN110674874B (en) | Fine-grained image identification method based on target fine component detection | |
CN110033431B (en) | Non-contact detection device and detection method for detecting corrosion area on surface of steel bridge | |
CN103473545B (en) | A kind of text image method for measuring similarity based on multiple features | |
CN112419299A (en) | Bolt loss detection method, device, equipment and storage medium | |
CN111680609B (en) | Foreign matter identification system and method based on image registration and target detection | |
CN109858438B (en) | Lane line detection method based on model fitting | |
CN112733914B (en) | Underwater target visual identification classification method based on support vector machine | |
CN111259707B (en) | Training method of special linear lane line detection model | |
CN114529839A (en) | Unmanned aerial vehicle routing inspection-oriented power transmission line hardware anomaly detection method and system | |
CN111582126A (en) | Pedestrian re-identification method based on multi-scale pedestrian contour segmentation fusion | |
CN108133211B (en) | Power distribution cabinet detection method based on mobile terminal visual image | |
CN109993107B (en) | Mobile robot obstacle visual detection method based on non-iterative K-means algorithm | |
CN113052260A (en) | Transformer substation foreign matter identification method and system based on image registration and target detection | |
CN114429649A (en) | Target image identification method and device | |
CN117036342B (en) | Chip defect identification method and system | |
CN116703895B (en) | Small sample 3D visual detection method and system based on generation countermeasure network | |
CN113592839A (en) | Distribution network line typical defect diagnosis method and system based on improved fast RCNN | |
CN111291712B (en) | Forest fire recognition method and device based on interpolation CN and capsule network | |
CN111950556A (en) | License plate printing quality detection method based on deep learning | |
Amin et al. | An automatic number plate recognition of Bangladeshi vehicles |
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 | ||
RJ01 | Rejection of invention patent application after publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20210629 |