CN114155394A - Multi-level target detection method and device suitable for multi-target - Google Patents
Multi-level target detection method and device suitable for multi-target Download PDFInfo
- Publication number
- CN114155394A CN114155394A CN202111208213.XA CN202111208213A CN114155394A CN 114155394 A CN114155394 A CN 114155394A CN 202111208213 A CN202111208213 A CN 202111208213A CN 114155394 A CN114155394 A CN 114155394A
- Authority
- CN
- China
- Prior art keywords
- target
- filtering
- classification
- preprocessing
- image
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/73—Deblurring; Sharpening
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a multi-stage target detection method and device suitable for multiple targets. The invention provides a multi-level target detection method suitable for multiple targets, which comprises the steps of acquiring an input image of a multi-level target to be detected, and preprocessing the input image; identifying a target area in the image after linear filtering preprocessing by using an RPN network model, classifying the identified target area to determine classification attribution, classifying and merging the target area after determining classification attribution, distinguishing a background from a foreground in the target area after classification merging, and carrying out image segmentation on the distinguished background and foreground so as to achieve a better foreground detection effect in various complex scenes; the problem that mutual shielding is easy to occur among targets in the detection process and the problems of under-segmentation and over-segmentation of the targets in the detection process are easy to occur is solved, and the robustness under a complex environment is remarkably improved.
Description
Technical Field
The invention relates to the technical field of target detection, in particular to a multi-stage target detection method and device suitable for multiple targets.
Background
The target detection technology is an important subject in the field of computer vision research, and a specific target can be detected in a video by applying the technology, so that the technology is widely applied to the fields of video monitoring, medical diagnosis, intelligent transportation and the like. In the field of video surveillance, pedestrian detection is generally performed by adopting an object detection technology. By detecting the monitoring video frame by frame, the information such as whether the pedestrian exists and the number of the pedestrians can be obtained. Multi-target detection and tracking are always research hotspots in the field of computer vision, and the existing detection method is difficult to obtain a good foreground detection effect in various complex scenes; in addition, in the detection process, the mutual shielding problem easily occurs among targets, the under-segmentation and over-segmentation problems easily occur in the target detection, and the robustness under the complex environment needs to be improved. Therefore, it is necessary to provide a multi-stage object detection method and apparatus suitable for multiple targets to solve the above problems.
Disclosure of Invention
The invention provides a multi-stage target detection method and device suitable for multiple targets, which aim to solve the problem that the existing detection method is difficult to obtain a good foreground detection effect in various complex scenes; in addition, in the detection process, the mutual shielding problem easily occurs among targets, the under-segmentation and over-segmentation problems easily occur in the target detection, and the robustness under the complex environment needs to be improved.
In a first aspect, the present invention provides a multi-stage target detection method suitable for multiple targets, including:
acquiring an input image of a multi-level target to be processed;
preprocessing the input image;
identifying a target region in the image subjected to linear filtering preprocessing by using an RPN network model;
classifying the identified target area to determine a classification attribution;
classifying and merging the target areas after the classification attribution is determined;
distinguishing a background from a foreground in the classified and combined target area;
and carrying out image segmentation on the distinguished background and foreground.
Further, preprocessing the input image includes:
performing filtering preprocessing on the input image based on an image filter;
and carrying out image sharpening pretreatment on the input image after the filtering pretreatment.
Further, in the step of performing filtering preprocessing on the input image based on an image filter, performing filtering preprocessing on the input image by using one or more of the following image filters: block filtering, mean filtering, gaussian filtering, median filtering, bilateral filtering.
Further, in the step of performing image sharpening preprocessing on the input image after filtering preprocessing, sharpening preprocessing is realized based on a Canny operator, a Sobel operator, a Laplacian operator and a Scharr filter.
Further, classifying the identified target area to determine a classification attribution, comprising:
extracting at least one candidate region feature of the target region;
obtaining at least one first probability vector corresponding to at least two major classes based on the candidate region characteristics, classifying each major class, and respectively obtaining at least one second probability vector corresponding to at least two minor classes in the major classes;
determining a classification probability that the target belongs to the subclass based on the first probability vector and the second probability vector;
judging whether the classification probability of the target belonging to the subclass is greater than a preset threshold value or not;
and if the classification probability of the target belonging to the subclass is greater than a preset threshold value, judging the target to belong to the subclass.
In a second aspect, the present invention provides a multi-stage object detection apparatus suitable for multiple objects, comprising:
an acquisition unit configured to acquire an input image of a multi-level target to be performed;
the preprocessing unit is used for preprocessing the input image;
the identification unit is used for identifying the target area in the image subjected to the linear filtering pretreatment by using the RPN network model;
the classification unit is used for classifying the identified target area to determine classification attribution;
the merging unit is used for classifying and merging the target areas after the classification attribution is determined;
the distinguishing unit is used for distinguishing the background from the foreground of the classified and combined target area;
and the segmentation unit is used for carrying out image segmentation on the distinguished background and foreground.
Further, the pre-processing comprises:
the filtering subunit is used for carrying out filtering preprocessing on the input image based on an image filter;
and the sharpening subunit is used for carrying out image sharpening preprocessing on the input image after the filtering preprocessing.
Further, the filtering subunit is configured to perform filtering preprocessing on the input image by using one or more of the following image filters: block filtering, mean filtering, gaussian filtering, median filtering, bilateral filtering.
Further, the sharpening subunit is configured to implement sharpening preprocessing based on a Canny operator, a Sobel operator, a Laplacian operator, and a Scharr filter.
Further, the classification unit includes:
the extraction subunit is used for extracting at least one candidate region characteristic of the target region;
the classification subunit is used for obtaining at least one first probability vector corresponding to at least two major classes based on the candidate region characteristics, classifying each major class and respectively obtaining at least one second probability vector corresponding to at least two minor classes in the major classes;
a determining subunit, configured to determine, based on the first probability vector and the second probability vector, a classification probability that the target belongs to the subclass;
a judging subunit, configured to judge whether the classification probability that the target belongs to the subclass is greater than a preset threshold;
and the judging subunit is used for judging the target to belong to the classification of the subclass under the condition that the classification probability of the target belonging to the subclass is greater than a preset threshold value. .
The invention has the following beneficial effects: the invention provides a multi-level target detection method and device suitable for multiple targets, which are characterized in that an input image of a multi-level target to be detected is acquired, and the input image is preprocessed; identifying a target area in the image after linear filtering preprocessing by using an RPN network model, classifying the identified target area to determine classification attribution, classifying and merging the target area after determining classification attribution, distinguishing a background from a foreground in the target area after classification merging, and carrying out image segmentation on the distinguished background and foreground so as to achieve a better foreground detection effect in various complex scenes; the problem that mutual shielding is easy to occur among targets in the detection process and the problems of under-segmentation and over-segmentation of the targets in the detection process are easy to occur is solved, and the robustness under a complex environment is remarkably improved.
Drawings
In order to more clearly illustrate the technical solution of the present invention, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious to those skilled in the art that other drawings can be obtained according to the drawings without any inventive exercise.
Fig. 1 is a flowchart of a multi-stage target detection method suitable for multiple targets according to an embodiment of the present invention.
Fig. 2 is a flowchart of an embodiment of a multi-stage target detection method suitable for multiple targets according to the present invention.
Fig. 3 is a schematic diagram of a multi-stage object detection apparatus suitable for multiple objects according to an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the technical solutions of the present invention will be clearly and completely described below with reference to the specific embodiments of the present invention and the accompanying drawings. It is to be understood that the described embodiments are merely exemplary of the invention, and not restrictive of the full scope of the invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention. The technical solutions provided by the embodiments of the present invention are described in detail below with reference to the accompanying drawings.
Referring to fig. 1, the present invention provides a multi-stage target detection method for multiple targets, including:
step S101, acquiring an input image of a multi-level target to be processed.
Step S102, preprocessing the input image.
In this embodiment, the preprocessing the input image may specifically include: performing filtering preprocessing on the input image based on an image filter; and carrying out image sharpening pretreatment on the input image after the filtering pretreatment.
Specifically, in the step of performing filtering preprocessing on the input image based on an image filter, one or more of the following image filters are used to perform filtering preprocessing on the input image: block filtering, mean filtering, gaussian filtering, median filtering, bilateral filtering. The smoothing preprocessing is also called fuzzy preprocessing and is mainly used for eliminating noise parts in images, the smoothing preprocessing is commonly used for reducing noise points or distortion on the images, the smoothing mainly uses image filtering, the method commonly used for image smoothing is an image filter, and the commonly used image filters include the following steps: block filtering-BoxBlur function, mean filtering (neighborhood average filtering) -Blur function, Gaussian filtering-Gaussian filtering function, median filtering-mediaBlur function, bilateral filtering-bilatelfilter function.
Specifically, in the step of performing image sharpening preprocessing on the input image after filtering preprocessing, sharpening preprocessing is realized based on a Canny operator, a Sobel operator, a Laplacian operator and a Scharr filter. The image sharpening operation is to highlight the boundary and other details of the image, and the image sharpening is realized through various operators and filters.
And step S103, identifying the target area in the image after the linear filtering pretreatment by using an RPN network model.
And step S104, classifying the identified target area to determine classification attribution.
As shown in fig. 2, in this embodiment, classifying the identified target area to determine a classification attribution includes:
in step S201, at least one candidate region feature of the target region is extracted.
Step S202, based on the candidate region features, at least one first probability vector corresponding to at least two major classes is obtained, each major class is classified, and at least one second probability vector corresponding to at least two minor classes in the major classes is obtained respectively.
Specifically, the classification is performed by a first classifier based on at least one candidate region feature to obtain at least one first probability vector corresponding to at least two major classes, and the classification is performed on each major class by at least two second classifiers based on at least one candidate region feature to obtain at least one second probability vector corresponding to at least two minor classes in the major classes, respectively.
Step S203, determining the classification probability of the target belonging to the subclass based on the first probability vector and the second probability vector.
Step S204, judging whether the classification probability of the target belonging to the subclass is larger than a preset threshold value.
Step S205, if the classification probability of the target belonging to the subclass is larger than a preset threshold, the target is judged to belong to the subclass.
And step S105, classifying and merging the target areas after the classification attribution is determined.
And step S106, distinguishing the background and the foreground of the target area after classification and combination.
And step S107, carrying out image segmentation on the distinguished background and foreground.
According to the embodiment, the multi-level target detection method suitable for multiple targets, provided by the invention, comprises the steps of acquiring an input image of a multi-level target to be detected, and preprocessing the input image; identifying a target area in the image after linear filtering preprocessing by using an RPN network model, classifying the identified target area to determine classification attribution, classifying and merging the target area after determining classification attribution, distinguishing a background from a foreground in the target area after classification merging, and carrying out image segmentation on the distinguished background and foreground so as to achieve a better foreground detection effect in various complex scenes; the problem that mutual shielding is easy to occur among targets in the detection process and the problems of under-segmentation and over-segmentation of the targets in the detection process are easy to occur is solved, and the robustness under a complex environment is remarkably improved.
Referring to fig. 3, the present invention provides a multi-stage target detection apparatus for multiple targets, comprising:
an acquisition unit 31 for acquiring an input image of a multi-level target to be performed;
a preprocessing unit 32 for preprocessing the input image;
the identification unit 33 is configured to identify a target region in the image after the linear filtering preprocessing by using an RPN network model;
a classification unit 34 configured to classify the identified target region to determine a classification home;
a merging unit 35, configured to perform classification merging on the target areas to which the classification attribution is determined;
a distinguishing unit 36, configured to distinguish a background from a foreground in the classified and combined target area;
and a segmentation unit 37, configured to perform image segmentation on the distinguished background and foreground.
Further, the pre-processing comprises:
the filtering subunit is used for carrying out filtering preprocessing on the input image based on an image filter;
and the sharpening subunit is used for carrying out image sharpening preprocessing on the input image after the filtering preprocessing.
Further, the filtering subunit is configured to perform filtering preprocessing on the input image by using one or more of the following image filters: block filtering, mean filtering, gaussian filtering, median filtering, bilateral filtering.
Further, the sharpening subunit is configured to implement sharpening preprocessing based on a Canny operator, a Sobel operator, a Laplacian operator, and a Scharr filter.
Further, the classification unit includes:
the extraction subunit is used for extracting at least one candidate region characteristic of the target region;
the classification subunit is used for obtaining at least one first probability vector corresponding to at least two major classes based on the candidate region characteristics, classifying each major class and respectively obtaining at least one second probability vector corresponding to at least two minor classes in the major classes;
a determining subunit, configured to determine, based on the first probability vector and the second probability vector, a classification probability that the target belongs to the subclass;
a judging subunit, configured to judge whether the classification probability that the target belongs to the subclass is greater than a preset threshold;
and the judging subunit is used for judging the target to belong to the classification of the subclass under the condition that the classification probability of the target belonging to the subclass is greater than a preset threshold value.
An embodiment of the present invention further provides a storage medium, and a storage medium, where a computer program is stored in the storage medium, and when the computer program is executed by a preprocessor, part or all of the steps in each embodiment of the multi-target multi-level target detection method provided by the present invention are implemented. The storage medium may be a magnetic disk, an optical disk, a Read-only memory (ROM) or a Random Access Memory (RAM).
Those skilled in the art will readily appreciate that the techniques of the embodiments of the present invention may be implemented as software plus a required general purpose hardware platform. Based on such understanding, the technical solutions in the embodiments of the present invention may be essentially or partially implemented in the form of a software product, which may be stored in a storage medium, such as ROM/RAM, magnetic disk, optical disk, etc., and includes several instructions for enabling a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the method according to the embodiments or some parts of the embodiments.
The same and similar parts in the various embodiments in this specification may be referred to each other. In particular, as for the embodiment of the multi-stage object detection device suitable for multiple objects, since it is basically similar to the method embodiment, the description is simple, and the relevant points can be referred to the description in the method embodiment.
The above-described embodiments of the present invention should not be construed as limiting the scope of the present invention.
Claims (10)
1. A multi-stage target detection method suitable for multiple targets is characterized by comprising the following steps:
acquiring an input image of a multi-level target to be processed;
preprocessing the input image;
identifying a target region in the image subjected to linear filtering preprocessing by using an RPN network model;
classifying the identified target area to determine a classification attribution;
classifying and merging the target areas after the classification attribution is determined;
distinguishing a background from a foreground in the classified and combined target area;
and carrying out image segmentation on the distinguished background and foreground.
2. The method of claim 1, wherein pre-processing the input image comprises:
performing filtering preprocessing on the input image based on an image filter;
and carrying out image sharpening pretreatment on the input image after the filtering pretreatment.
3. The method of claim 2, wherein the step of performing filter pre-processing on the input image based on an image filter performs filter pre-processing on the input image using one or more of the following image filters: block filtering, mean filtering, gaussian filtering, median filtering, bilateral filtering.
4. The method of claim 2, wherein in the step of performing the image sharpening preprocessing on the input image after the filtering preprocessing, the sharpening preprocessing is performed based on a Canny operator, a Sobel operator, a Laplacian operator, and a Scharr filter.
5. The method of claim 1, wherein classifying the identified target region to determine a classification home comprises:
extracting at least one candidate region feature of the target region;
obtaining at least one first probability vector corresponding to at least two major classes based on the candidate region characteristics, classifying each major class, and respectively obtaining at least one second probability vector corresponding to at least two minor classes in the major classes;
determining a classification probability that the target belongs to the subclass based on the first probability vector and the second probability vector;
judging whether the classification probability of the target belonging to the subclass is greater than a preset threshold value or not;
and if the classification probability of the target belonging to the subclass is greater than a preset threshold value, judging the target to belong to the subclass.
6. A multi-stage object detection apparatus adapted for multiple targets, comprising:
an acquisition unit configured to acquire an input image of a multi-level target to be performed;
the preprocessing unit is used for preprocessing the input image;
the identification unit is used for identifying the target area in the image subjected to the linear filtering pretreatment by using the RPN network model;
the classification unit is used for classifying the identified target area to determine classification attribution;
the merging unit is used for classifying and merging the target areas after the classification attribution is determined;
the distinguishing unit is used for distinguishing the background from the foreground of the classified and combined target area;
and the segmentation unit is used for carrying out image segmentation on the distinguished background and foreground.
7. The apparatus of claim 6, wherein the pre-processing comprises:
the filtering subunit is used for carrying out filtering preprocessing on the input image based on an image filter;
and the sharpening subunit is used for carrying out image sharpening preprocessing on the input image after the filtering preprocessing.
8. The apparatus of claim 7, wherein the filtering subunit is configured to perform filtering pre-processing on the input image using one or more of the following image filters: block filtering, mean filtering, gaussian filtering, median filtering, bilateral filtering.
9. The apparatus of claim 7, wherein the sharpening subunit is to implement a sharpening pre-process based on a Canny operator, a Sobel operator, a Laplacian operator, and a Scharr filter.
10. The apparatus of claim 6, wherein the classification unit comprises:
the extraction subunit is used for extracting at least one candidate region characteristic of the target region;
the classification subunit is used for obtaining at least one first probability vector corresponding to at least two major classes based on the candidate region characteristics, classifying each major class and respectively obtaining at least one second probability vector corresponding to at least two minor classes in the major classes;
a determining subunit, configured to determine, based on the first probability vector and the second probability vector, a classification probability that the target belongs to the subclass;
a judging subunit, configured to judge whether the classification probability that the target belongs to the subclass is greater than a preset threshold;
and the judging subunit is used for judging the target to belong to the classification of the subclass under the condition that the classification probability of the target belonging to the subclass is greater than a preset threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111208213.XA CN114155394A (en) | 2021-10-18 | 2021-10-18 | Multi-level target detection method and device suitable for multi-target |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111208213.XA CN114155394A (en) | 2021-10-18 | 2021-10-18 | Multi-level target detection method and device suitable for multi-target |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114155394A true CN114155394A (en) | 2022-03-08 |
Family
ID=80462739
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111208213.XA Pending CN114155394A (en) | 2021-10-18 | 2021-10-18 | Multi-level target detection method and device suitable for multi-target |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114155394A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117036798A (en) * | 2023-08-04 | 2023-11-10 | 宁波市电力设计院有限公司 | A method and system for image recognition of power transmission and distribution lines based on deep learning |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070183663A1 (en) * | 2006-02-07 | 2007-08-09 | Haohong Wang | Intra-mode region-of-interest video object segmentation |
| CN109285178A (en) * | 2018-10-25 | 2019-01-29 | 北京达佳互联信息技术有限公司 | Image partition method, device and storage medium |
| CN109712118A (en) * | 2018-12-11 | 2019-05-03 | 武汉三江中电科技有限责任公司 | A kind of substation isolating-switch detection recognition method based on Mask RCNN |
| CN110751079A (en) * | 2019-10-16 | 2020-02-04 | 北京海益同展信息科技有限公司 | Article detection method, apparatus, system and computer readable storage medium |
| CN110879950A (en) * | 2018-09-06 | 2020-03-13 | 北京市商汤科技开发有限公司 | Multi-stage target classification and traffic sign detection method and device, equipment and medium |
| CN112364931A (en) * | 2020-11-20 | 2021-02-12 | 长沙军民先进技术研究有限公司 | Low-sample target detection method based on meta-feature and weight adjustment and network model |
| CN113012149A (en) * | 2021-04-14 | 2021-06-22 | 北京铁道工程机电技术研究所股份有限公司 | Intelligent cleaning robot path planning method and system |
-
2021
- 2021-10-18 CN CN202111208213.XA patent/CN114155394A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20070183663A1 (en) * | 2006-02-07 | 2007-08-09 | Haohong Wang | Intra-mode region-of-interest video object segmentation |
| CN110879950A (en) * | 2018-09-06 | 2020-03-13 | 北京市商汤科技开发有限公司 | Multi-stage target classification and traffic sign detection method and device, equipment and medium |
| CN109285178A (en) * | 2018-10-25 | 2019-01-29 | 北京达佳互联信息技术有限公司 | Image partition method, device and storage medium |
| CN109712118A (en) * | 2018-12-11 | 2019-05-03 | 武汉三江中电科技有限责任公司 | A kind of substation isolating-switch detection recognition method based on Mask RCNN |
| CN110751079A (en) * | 2019-10-16 | 2020-02-04 | 北京海益同展信息科技有限公司 | Article detection method, apparatus, system and computer readable storage medium |
| CN112364931A (en) * | 2020-11-20 | 2021-02-12 | 长沙军民先进技术研究有限公司 | Low-sample target detection method based on meta-feature and weight adjustment and network model |
| CN113012149A (en) * | 2021-04-14 | 2021-06-22 | 北京铁道工程机电技术研究所股份有限公司 | Intelligent cleaning robot path planning method and system |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117036798A (en) * | 2023-08-04 | 2023-11-10 | 宁波市电力设计院有限公司 | A method and system for image recognition of power transmission and distribution lines based on deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Stalder et al. | Cascaded confidence filtering for improved tracking-by-detection | |
| CN101344922B (en) | Human face detection method and device | |
| WO2018019194A1 (en) | Image recognition method, terminal, and nonvolatile storage medium | |
| Zhang et al. | Knowledge-based eye detection for human face recognition | |
| EP3510526B1 (en) | Particle boundary identification | |
| Radha et al. | A novel approach to extract text from license plate of vehicles | |
| Lee et al. | Blurred image region detection and segmentation | |
| CN106446921A (en) | High-voltage power transmission line barrier identification method and apparatus | |
| Hidayatullah et al. | Automatic sperms counting using adaptive local threshold and ellipse detection | |
| Paunwala et al. | A novel multiple license plate extraction technique for complex background in Indian traffic conditions | |
| Fomani et al. | License plate detection using adaptive morphological closing and local adaptive thresholding | |
| CN114581658A (en) | Target detection method and device based on computer vision | |
| Zhan et al. | A robust split-and-merge text segmentation approach for images | |
| CN114913112A (en) | Method, device and equipment for detecting double edges of wafer | |
| CN114155394A (en) | Multi-level target detection method and device suitable for multi-target | |
| CN112329572B (en) | Rapid static living body detection method and device based on frame and flash point | |
| Zaharin et al. | Comparison of human detection using background subtraction and frame difference | |
| Mol et al. | Text recognition using poisson filtering and edge enhanced maximally stable extremal regions | |
| Mubin et al. | Identification of parking lot status using circle blob detection | |
| Delforouzi et al. | A vision-based method for automatic crack detection in railway sleepers | |
| Yang et al. | A modified method of vehicle extraction based on background subtraction | |
| CN114373067A (en) | Bar code reading method and device based on image feature region triggering reading | |
| Suresh et al. | A survey on occlusion detection | |
| Satish et al. | Edge assisted fast binarization scheme for improved vehicle license plate recognition | |
| Gim et al. | A novel framework for white blood cell segmentation based on stepwise rules and morphological features |
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 |