CN115171048A - Asset classification method, system, terminal and storage medium based on image recognition - Google Patents
Asset classification method, system, terminal and storage medium based on image recognition Download PDFInfo
- Publication number
- CN115171048A CN115171048A CN202210860571.7A CN202210860571A CN115171048A CN 115171048 A CN115171048 A CN 115171048A CN 202210860571 A CN202210860571 A CN 202210860571A CN 115171048 A CN115171048 A CN 115171048A
- Authority
- CN
- China
- Prior art keywords
- confidence
- image
- category
- image features
- picture
- 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
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Artificial Intelligence (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Image Analysis (AREA)
Abstract
The application relates to an asset classification method, system, terminal and storage medium based on image recognition, belonging to the field of image recognition; the asset classification method based on image identification comprises the steps of obtaining a web response page picture of video equipment; performing feature extraction on the web response page picture by using a feature extraction technology to obtain image features; calling a preset confidence coefficient relation library, and obtaining a category confidence coefficient corresponding to the image features according to the confidence coefficient relation library and the image features; the confidence coefficient relation library stores image characteristics, category confidence coefficients and corresponding relations between the image characteristics and the category confidence coefficients; based on a preset picture category confidence coefficient calculation rule, obtaining a picture category confidence coefficient of a web response page picture according to the image features and the category confidence coefficients of the image features; and determining the category information of the video equipment according to the image category confidence of the web response page image. The method and the device have the effect of improving the operation convenience of identifying and classifying the video assets.
Description
Technical Field
The present application relates to the field of image recognition, and in particular, to an asset classification method, system, terminal and storage medium based on image recognition.
Background
With the development of the technology of the internet of things, the number of global internet of things equipment is increased explosively, and the problem comes with the global internet of things equipment; on one hand, the network administrator has unknown conditions on the number, type, brand and the like of the network access equipment; on the other hand, some vulnerabilities are for specific brands and types of devices, which are difficult to effectively guard against without knowing the brand type of the device.
The video monitoring network is an important branch of the development of the internet of things technology and becomes an important component in the construction of a smart city; a large number of cameras, hard disk video recorders, video monitoring and operation and maintenance servers and the like exist in a video monitoring network, and the video equipment is video assets.
At present, the video assets are identified and classified by shooting a picture of entity equipment and then identifying and classifying the video assets according to the picture; however, it is difficult to take a picture for each physical device, and the angle and quality of the picture can affect the identification and classification of video assets.
Disclosure of Invention
The application provides an asset classification method, system, terminal and storage medium based on image identification, and has the characteristic of improving the operation convenience of video asset identification and classification.
The asset classification method based on image recognition is provided.
The above object of the present application is achieved by the following technical solutions:
the asset classification method based on image recognition comprises the following steps:
acquiring a web response page picture of video equipment;
performing feature extraction on the web response page picture by using a feature extraction technology to obtain image features;
calling a preset confidence coefficient relation library, and obtaining a category confidence coefficient corresponding to the image features according to the confidence coefficient relation library and the image features;
the confidence coefficient relation library stores image characteristics, category confidence coefficients and corresponding relations between the image characteristics and the category confidence coefficients;
based on a preset picture category confidence coefficient calculation rule, obtaining a picture category confidence coefficient of a web response page picture according to the image characteristics and the category confidence coefficient of the image characteristics;
and determining the category information of the video equipment according to the image category confidence of the web response page image.
By adopting the technical scheme, the web response page picture of the video equipment is obtained firstly, the picture contains image features which possibly represent the type and the brand of the video equipment, then the image features are processed, the category confidence coefficient of the picture is determined according to the confidence coefficient of the image features, and then the category of the video equipment is determined, so that the operation convenience of identifying and classifying the video assets is improved.
The present application may be further configured in a preferred example, the step of obtaining the web response page picture of the video device includes:
simulating and generating an HTTP request by utilizing a network scanning technology;
based on the HTTP request, scanning a web open port of the video equipment to obtain a web response page;
and carrying out screenshot on the web response page to obtain a web response page picture.
In a preferred example, the method for calculating the image category confidence of the web response page image according to the image feature and the category confidence of the image feature based on the preset image category confidence calculation rule may further include:
the picture category confidence calculation rule comprises a first rule;
the first rule includes:
comparing the category confidence of all image features in the web response page picture with a first preset confidence;
if the category confidence of the image features is greater than a first preset confidence;
and regarding the category confidence of the image features as the picture category confidence of the web response page picture.
In a preferred example, the method for obtaining the image category confidence of the web response page image according to the image feature and the category confidence of the image feature based on the preset image category confidence calculation rule may further include:
the picture category confidence degree calculation rule comprises a second rule;
the second rule includes:
when the category confidence degrees of all the image features in the web response page picture are smaller than a first preset confidence degree;
comparing the category confidence of all image features in the web response page picture;
selecting image features with first high class confidence coefficient and second high class confidence coefficient, calculating the confidence coefficient difference of the two image features, comparing the difference with a second preset confidence coefficient, and if the difference is greater than the confidence coefficient threshold and the class information of the two image features is different, marking the image features with first high class confidence coefficient as first features to be calculated;
acquiring category information of a first feature to be calculated;
acquiring the class information of the rest image features, comparing the class information with the class information of the first feature to be calculated, reserving the image features with the same class information in the rest image features, selecting the image feature with the first high confidence coefficient, and marking the image feature as the second feature to be calculated;
and obtaining the image category confidence of the web response page image according to the category confidence of the first feature to be calculated and the category confidence of the second feature to be calculated based on the calculation rule contained in the second rule.
The present application may be further configured in a preferred example, that the calculation rule includes:
M 1 ＝1-[(1-M 2 )×(1-M 3 )]；
wherein M is 1 Indicates picture class confidence, M 2 Representing a first feature to be calculatedClass confidence of, M 3 And representing the class confidence of the second feature to be calculated.
In a preferred example, the method for obtaining the image category confidence of the web response page image according to the image feature and the category confidence of the image feature based on the preset image category confidence calculation rule may further include:
the picture category confidence calculation rule comprises a third rule;
the third rule includes:
when the category confidence degrees of all image features in the web response page picture are smaller than a first preset confidence degree, when the confidence degree difference between the image features with the first high category confidence degree and the image features with the second high category confidence degree is smaller than a confidence degree threshold value, and the category information of the two image features is different;
marking the class information of the image features with the first high class confidence coefficient as first class information;
marking the class information of the image features with the second highest class confidence coefficients as second class information;
calculating the confidence coefficient of the image features corresponding to the first type of information to obtain a first confidence coefficient based on a second rule;
calculating the confidence of the image features corresponding to the second type of information based on a second rule to obtain a second confidence;
and calculating a difference value between the first confidence coefficient and the second confidence coefficient, comparing the difference value with a preset difference value threshold, and if the difference value is greater than the preset difference value threshold, regarding the first confidence coefficient as the image category confidence coefficient of the web response page image.
In a preferred example, the method for obtaining the image category confidence of the web response page image according to the image feature and the category confidence of the image feature based on the preset image category confidence calculation rule may further include:
the picture category confidence calculation rule comprises a fourth rule;
the fourth rule includes:
when the difference value between the first confidence coefficient and the second confidence coefficient is smaller than a preset difference threshold value;
acquiring image features of the first three high levels of category confidence corresponding to the first category information;
based on a calculation rule contained in the fourth rule, obtaining proportion information of the image features according to the confidence coefficient of the image features, and calculating a category confidence coefficient corresponding to the first category information according to the proportion information and the confidence coefficient of the image features;
acquiring image features of the first three high category confidence degrees corresponding to the second category information;
based on the calculation rule contained in the fourth rule, obtaining the proportion information of the image features according to the confidence coefficient of the image features, and calculating the category confidence coefficient corresponding to the second category information according to the proportion information and the confidence coefficient of the image features;
and comparing the two category confidences, and taking the larger category confidence as the picture category confidence of the web response page picture.
The second purpose of the application is to provide an asset classification system based on image recognition.
The second application object of the present application is achieved by the following technical scheme:
an asset classification system based on image recognition, comprising:
the acquisition module is used for acquiring a web response page picture of the video equipment;
the extraction module is used for extracting the features of the web response page picture by using a feature extraction technology to obtain image features;
the calling module is used for calling a preset confidence coefficient relation library and obtaining a category confidence coefficient corresponding to the image features according to the confidence coefficient relation library and the image features; the confidence coefficient relation library is stored with image characteristics, category confidence coefficients and corresponding relations between the image characteristics and the category confidence coefficients;
the computing module is used for obtaining the image category confidence of the web response page image according to the image features and the category confidence of the image features based on a preset image category confidence computing rule;
and the determining module is used for determining the category information of the video equipment according to the image category confidence of the web response page image.
The third purpose of the application is to provide an intelligent terminal.
The third application purpose of the present application is achieved through the following technical scheme:
an intelligent terminal comprises a memory and a processor, wherein the memory stores computer program instructions of the asset classification method based on image recognition, which can be loaded and executed by the processor.
The fourth purpose of the present application is to provide a computer medium capable of storing a corresponding program.
The fourth application purpose of the present application is achieved by the following technical solutions:
a computer readable storage medium storing a computer program that can be loaded by a processor and executed to perform any of the above-described image recognition-based asset classification methods.
In summary, the present application includes at least one of the following beneficial technical effects:
capturing a web response page of a video device, acquiring image characteristics from a picture, determining the confidence coefficient of the picture according to the confidence coefficient of the image characteristics, and further determining the category of the picture; by the mode, the category and the brand of the video equipment can be confirmed without photographing the entity structure of the video equipment, and the operation convenience for identifying and classifying the video assets is improved.
Drawings
Fig. 1 is a schematic flowchart of an asset classification method based on image recognition in an embodiment of the present application.
Fig. 2 is a schematic structural diagram of an asset classification system based on image recognition in an embodiment of the present application.
Description of reference numerals: 1. an acquisition module; 2. an extraction module; 3. a calling module; 4. a calculation module; 5. and determining a module.
Detailed Description
The present embodiment is only for explaining the present application and is not limited to the present application, and those skilled in the art can make modifications without inventive contribution to the present embodiment as needed after reading the present specification, but all of them are protected by patent law within the scope of the claims of the present application.
In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application without making any creative effort belong to the protection scope of the present application.
The embodiments of the present application will be described in further detail below with reference to the drawings.
With the development of science and technology, the internet of things becomes a part of the life of people; from logistics to express delivery, from smart cities to smart homes, the thinking of the interconnection of everything is going deep into the mind; video monitoring is an important branch of the development of the internet of things technology and becomes an important component in the construction of smart cities; in a video surveillance network that has been put into operation, there are a large number of related devices belonging to video assets, and in order to more efficiently inventory and comb these video assets, the present application provides an asset classification method based on image recognition.
The main flow of the method provided by the present application is described below.
As shown in fig. 1:
step S101: and acquiring a web response page picture of the video equipment.
In the embodiment of the application, an HTTP request is generated by utilizing a network scanning technology in a simulation mode, and then a web open port of video equipment is scanned based on the HTTP request to obtain a web response page; then, screenshot is carried out on the web response page to obtain a web response page picture; it is understood that in the above process, the screen shot is performed by accessing the open web service of the video device and then performing the screen shot on the web page that the video device responds to on the basis of the network scanning.
Step S102: and performing feature extraction on the web response page picture by using a feature extraction technology to obtain image features.
In the obtained web response page picture, performing feature extraction on the picture by using a feature extraction technology so as to obtain image features with variable quantity; the image characteristics include information indicating the type or brand of the video equipment corresponding to the picture; the specific extraction process is the same as that in the conventional technology, and is not described herein again.
Step S103: and calling a preset confidence coefficient relational database, and obtaining a category confidence coefficient corresponding to the image features according to the confidence coefficient relational database and the image features.
In the embodiment of the present application, each image feature is assigned a class confidence, where the class confidence may be understood as the probability or confidence of the class a to which the image feature a belongs; that is, if the picture includes the image feature a, the class confidence of the image feature a indicates the confidence that the picture belongs to the class, and the above example is merely an exemplary description.
The image features, the category confidence and the corresponding relationship between the image features and the category confidence are stored in the confidence relationship library, so that the category confidence of each image feature can be deduced according to the image features and the confidence relationship library after the image features and the confidence relationship library are known.
Step S104: and obtaining the image category confidence of the web response page image according to the image features and the category confidence of the image features based on a preset image category confidence calculation rule.
In the embodiment of the present application, the image category confidence calculation rule includes a plurality of sub-rules, and the sub-rules are for dealing with images under different conditions, specifically, the image category confidence calculation rule includes four sub-rules.
The image category confidence calculation rule comprises a first rule which is as follows: comparing the category confidence of all image features in the web response page picture with a first preset confidence; if the category confidence coefficient of one image feature is greater than a first preset confidence coefficient; and regarding the category confidence of the image features as the picture category confidence of the web response page picture.
All image features in the picture and the category confidence of the image features can be obtained according to the steps S101-S103; the first preset confidence is a decisive threshold, that is, if there is an image feature whose confidence exceeds the first preset confidence, the category confidence of the image feature can be directly regarded as the image category confidence of the corresponding web page picture; for example, if the image a includes the image feature a, and the confidence of the image feature a belonging to the i category is 95%, which exceeds the first preset confidence 90%, the confidence of the image feature a of 95% may be directly regarded as the confidence of the image a, that is, the probability that the image a belongs to the i category is indicated to be 95%.
The image category confidence calculation rule includes a second rule, and the second rule is as follows: when the category confidence degrees of all the image features in the web response page picture are smaller than a first preset confidence degree; comparing the category confidence of all image features in the web response page picture; selecting image features with first high class confidence coefficient and second high class confidence coefficient, calculating the confidence coefficient difference between the two image features, comparing the difference with a second preset confidence coefficient, and if the difference is greater than the confidence coefficient threshold and the class information of the two image features is different, marking the image features with first high class confidence coefficient as first features to be calculated; acquiring category information of a first feature to be calculated; acquiring the class information of the rest image features, comparing the class information with the class information of the first feature to be calculated, reserving the image features with the same class information in the rest image features, selecting the image feature with the first high confidence coefficient, and marking the image feature as the second feature to be calculated; and obtaining the image category confidence of the web response page image according to the category confidence of the first feature to be calculated and the category confidence of the second feature to be calculated based on the calculation rule contained in the second rule.
On the premise of using the second rule, the picture does not meet the requirement in the first rule, that is, the confidence degrees of all the image features in the picture are lower than the first preset confidence degree, which indicates that no image feature capable of directly representing the picture exists in the picture at the moment, and then the category confidence degree of the picture needs to be calculated through the second rule at the moment.
Firstly, the category confidence degrees of the image features are compared, the image features with the highest category confidence degree are selected, on the premise that the categories of the image features are different, the confidence degree difference between the two image features is calculated, then the difference value is compared with a second preset confidence degree, if the difference value is larger, the difference between the two image features is larger although the category confidence degrees of the two image features are the highest, and therefore only the first image feature is selected as a judgment reference.
The image features with the first high class confidence coefficient are marked as first to-be-calculated features, so that the features are convenient to distinguish; after the first feature to be calculated is determined, the category to which the first feature to be calculated belongs is obtained, then the image feature which is the same as the category to which the first feature to be calculated belongs is found in the rest image features, the image feature with the highest confidence coefficient is selected, and the image feature is marked as a second feature to be calculated; and then based on a calculation rule contained in the second rule, obtaining the class confidence coefficient of the carcass screen according to the class confidence coefficients of the first feature to be calculated and the second feature to be calculated.
To explain the above process in detail, the following is exemplified; the picture A comprises an image feature a, an image feature b, an image feature c and an image feature d; wherein the confidence of the i category of the image feature a is 85%, the confidence of the ii category of the image feature b is 70%, the confidence of the i category of the image feature c is 68%, and the confidence of the j category of the image feature d is 50%; if the category confidence degrees of the four image features do not exceed the first preset confidence degree by 90%, a second rule is applied; if the confidence difference between the a and the b is calculated to be 15% and is larger than the second preset confidence by 10%, the image feature a is marked as a first feature to be calculated; and b, the category to which the a belongs is i, selecting the image features c of which the same category belongs to i from the rest image features, and then calculating according to a calculation rule.
In the embodiment of the present application, the calculation rule is a specific calculation formula, as follows:
M 1 ＝1-[(1-M 2 )×(1-M 3 )]；
wherein M is 1 Representing picture categoriesConfidence, M 2 Representing the class confidence of the first feature to be computed, M 3 And representing the class confidence of the second feature to be calculated.
From this calculation formula, M can be calculated by substituting the confidences of a and c in the above example into them 1 If the confidence coefficient is 95.2%, the confidence coefficient is the final image category confidence coefficient; under a second rule, selecting the image features with the highest confidence coefficient by comparing the confidence coefficients of the image features, wherein in the second rule, the image features do not meet the requirements of the first rule, which means that even the image features with the highest confidence coefficient cannot be directly used for representing the confidence coefficient of the picture, the image features need to be further processed, and the final picture confidence coefficient is calculated according to a specific calculation formula; by the method, the image category confidence coefficient calculation rule is more rigorous and comprehensive, and can adapt to various conditions, so that the convenience and the accuracy of identifying the image category are improved, and the convenience and the accuracy of identifying the video asset type are further improved.
The image category confidence calculation rules comprise a third rule, and the third rule is as follows: when the category confidence degrees of all image features in the web response page picture are smaller than a first preset confidence degree, when the confidence degree difference between the image features with the first high category confidence degree and the image features with the second high category confidence degree is smaller than a confidence degree threshold value, and the category information of the two image features is different; marking the class information of the image features with the first high class confidence coefficient as first class information; marking the class information of the image features with the second highest class confidence coefficients as second class information; calculating the confidence coefficient of the image features corresponding to the first type of information to obtain a first confidence coefficient based on a second rule; calculating the confidence coefficient of the image features corresponding to the second category information based on a second rule to obtain a second confidence coefficient; and calculating a difference value between the first confidence coefficient and the second confidence coefficient, comparing the difference value with a preset difference value threshold, and if the difference value is greater than the preset difference value threshold, regarding the first confidence coefficient as the image category confidence coefficient of the web response page image.
It is to be understood that the third rule is a supplementary description of the second rule; that is, the second rule is that all image features in the picture do not satisfy the condition that the confidence coefficient is greater than the first preset confidence coefficient, and the confidence coefficient difference between the image features with the first high confidence coefficient and the image features with the second high confidence coefficient is greater than the second preset confidence coefficient, which indicates that the image features in the picture do not have the condition that the confidence coefficient is very high, but one image feature is higher than the rest image features, and the difference is large; the third rule aims at that the confidence coefficient of one image feature is higher than that of the rest image features, but the difference is not very large; this means that, for a picture, image features with the first highest confidence level and the second highest confidence level both affect the class identification of the picture, and then both need to be taken into account when calculating the class confidence level of the picture.
In the calculation process of the third rule, second type information is introduced, and then a first confidence coefficient and a second confidence coefficient corresponding to the first type information and the second type information are calculated respectively based on a calculation formula of the second rule; then calculating the difference value of the two, comparing the difference value with a threshold value, and if the difference value is greater than the threshold value, indicating that the first type of category information is the category information to which the picture belongs; the specific calculation process is similar to the second rule, and is not described herein again.
The image category confidence calculation rule comprises a fourth rule, and the fourth rule is as follows: when the difference value between the first confidence coefficient and the second confidence coefficient is smaller than a preset difference threshold value; acquiring image features of the first three high levels of category confidence corresponding to the first category information; based on a calculation rule contained in the fourth rule, obtaining proportion information of the image features according to the confidence coefficient of the image features, and calculating a category confidence coefficient corresponding to the first category information according to the proportion information and the confidence coefficient of the image features; acquiring image features with the first three high category confidence degrees corresponding to the second category information; based on a calculation rule contained in the fourth rule, obtaining proportion information of the image features according to the confidence coefficient of the image features, and calculating a category confidence coefficient corresponding to the second category information according to the proportion information and the confidence coefficient of the image features; and comparing the two category confidences, and taking the larger category confidence as the picture category confidence of the web response page picture.
The fourth rule is a supplementary explanation to the third rule; in a third rule, the last layer judges that the difference value between the first confidence coefficient and the second confidence coefficient is compared with a threshold value, the third rule aims at the condition that the difference value is larger than the threshold value, and the fourth rule aims at the condition that the difference value is smaller than the threshold value; that is, in the fourth rule, if the result obtained by calculating the two image features as the reference items still cannot show the category tendency of the picture in the process of calculating the confidence degrees of the two categories of information, a third image feature needs to be introduced, and then the calculation mode needs to be changed, so that the calculation accuracy of the confidence degree result of the category of the picture is improved.
Specifically, the fourth rule is exemplified; in one example, picture a includes image feature a, belonging to category i, with a confidence of 85%; image feature b, belonging to ii category with a confidence of 80%; image feature c, belonging to ii category with 70% confidence; image feature d, belonging to ii category with a confidence of 65%; the image features e belong to the category i, and the confidence coefficient is 75 percent; the image features f belong to the i category, and the confidence coefficient is 70%; in this example, picture a meets the requirements of the fourth rule; sorting the image features, wherein the image features belonging to the first type information i are a, e and f, and the image features belonging to the second type information ii are b, c and d; the default three image features of each category are the image features with the first three high confidence degrees of the categories, and the rest image features with lower confidence degrees are not shown; after the image features with the three highest confidence levels before the category confidence level corresponding to the category information are obtained, calculation is carried out by using the calculation rule contained in the fourth rule, and therefore the category confidence level corresponding to the category information is calculated.
In the embodiment of the present application, the fourth rule includes the following calculation rules: determining proportion information of the image features according to the confidence degrees of the image features; calculating the sum of the ratio information of any two image characteristics in the three image characteristics according to the ratio information of the image characteristics; calculating the overall confidence of any two image features in the three image features according to the calculation rule contained in the second rule; selecting three conditions of any two image characteristics in the three image characteristics, multiplying the proportion information corresponding to each condition by the overall confidence coefficient, and adding the three values to obtain the overall confidence coefficient; in this way, the overall confidence degrees of the first category information and the second category information are respectively calculated, the first category information and the second category information are compared, the greater overall confidence degree is selected, and the overall confidence degree is regarded as the category confidence degree of the picture.
The following examples are given; the image A comprises image characteristics a belonging to the category i, and the confidence coefficient is 80%; image feature b, belonging to ii category with 85% confidence; the image characteristics c belong to the category i, and the confidence coefficient is 70%; image feature d, belonging to ii category with a confidence of 75%; the image features e belong to the category i, and the confidence coefficient is 50%; the image feature f, belonging to the ii category, had a confidence of 45%.
Wherein ace belongs to the i category, bdf belongs to the ii category; calculating proportion information of three image characteristics under the category i; 50%/(80% +70% + 50%) =25%, indicating that the e-feature accounts for 25% of the three features; similarly, the proportion of a is calculated to be 40 percent, and the proportion of c is calculated to be 35 percent; the sum of the ac proportion information is 75%, the sum of the ae proportion information is 65%, and the sum of the ce proportion information is 60%; the confidence coefficient of ac is 94%, the confidence coefficient of ae is 90% and the confidence coefficient of ce is 85%, which can be calculated by the calculation rules contained in the second rule; then calculating the product of the confidence of any two image features and the sum of the proportion information to obtain 94% + 75% +90% + 65% +85% + 60% =70.5% +58.5% +51% =180%; here 180% represents the overall confidence of the i category; similarly, the overall confidence of the ii category is calculated to be 184%, the two overall confidences are compared, and finally the overall confidence of the ii category is regarded as the category confidence of the picture.
By the explanation of the rules, the category confidence of the picture can be accurately and conveniently calculated by using the picture category confidence calculation rule in the scheme.
Step S105: and determining the category information of the video equipment according to the image category confidence of the web response page image.
After the class confidence of the picture is determined through the step S104, the class of the picture can be determined according to the confidence; in the embodiment of the present application, the default is the category corresponding to the image category confidence, that is, the category of the image.
The above steps S101 to S105 are an example of the present application, and illustrate how to determine the category to which the picture belongs according to the picture and the image features in the picture, so as to implement identification of the video asset, and improve accuracy and convenience of identification and classification of the video asset.
The application also provides an asset classification system based on image recognition, as shown in fig. 2, the asset classification system based on image recognition comprises an obtaining module 1, a processing module and a display module, wherein the obtaining module is used for obtaining a web response page picture of video equipment; the extraction module 2 is used for extracting the features of the web response page picture by using a feature extraction technology to obtain image features; the calling module 3 is used for calling a preset confidence coefficient relation library and obtaining a category confidence coefficient corresponding to the image features according to the confidence coefficient relation library and the image features; the confidence coefficient relation library stores image characteristics, category confidence coefficients and corresponding relations between the image characteristics and the category confidence coefficients; the calculation module 4 is used for obtaining the image category confidence of the web response page image according to the image features and the category confidence of the image features based on a preset image category confidence calculation rule; and the determining module 5 is used for determining the category information of the video equipment according to the image category confidence of the web response page image.
In order to better execute the program of the method, the application also provides an intelligent terminal which comprises a memory and a processor.
Wherein the memory may be used to store an instruction, a program, code, a set of codes, or a set of instructions. The memory may include a stored program area and a stored data area, wherein the stored program area may store instructions for implementing an operating system, instructions for at least one function, and instructions for implementing the above-described image recognition-based asset classification method, and the like; the storage data area may store data and the like involved in the above-described asset classification method based on image recognition.
A processor may include one or more processing cores. The processor executes or executes the instructions, programs, code sets, or instruction sets stored in the memory, calls data stored in the memory, performs various functions of the present application, and processes the data. The processor may be at least one of an application specific integrated circuit, a digital signal processor, a digital signal processing device, a programmable logic device, a field programmable gate array, a central processing unit, a controller, a microcontroller, and a microprocessor. It is understood that the electronic device for implementing the above processor function may be other electronic devices, and the embodiments of the present application are not limited in particular.
The present application also provides a computer-readable storage medium, for example, comprising: various media capable of storing program codes, such as a usb disk, a removable hard disk, a Read Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk, or an optical disk. The computer readable storage medium stores a computer program that can be loaded by a processor and performs the above-described asset classification method based on image recognition.
The above description is only a preferred embodiment of the application and is illustrative of the principles of the technology employed. It will be appreciated by those skilled in the art that the scope of the disclosure herein is not limited to the particular combination of features described above, but also encompasses other arrangements formed by any combination of the above features or their equivalents without departing from the spirit of the disclosure. For example, the above features may be replaced with (but not limited to) features having similar functions disclosed in the present application.
Claims (10)
1. An asset classification method based on image recognition is characterized by comprising the following steps:
acquiring a web response page picture of video equipment;
performing feature extraction on the web response page picture by using a feature extraction technology to obtain image features;
calling a preset confidence coefficient relational database, and obtaining a category confidence coefficient corresponding to the image features according to the confidence coefficient relational database and the image features; the confidence coefficient relation library is stored with image characteristics, category confidence coefficients and corresponding relations between the image characteristics and the category confidence coefficients; based on a preset picture category confidence coefficient calculation rule, obtaining a picture category confidence coefficient of a web response page picture according to the image characteristics and the category confidence coefficient of the image characteristics;
and determining the category information of the video equipment according to the image category confidence of the web response page image.
2. The image recognition based asset classification method according to claim 1, wherein the step of obtaining a picture of a web response page of a video device comprises:
simulating and generating an HTTP request by utilizing a network scanning technology;
based on the HTTP request, scanning a web open port of the video equipment to obtain a web response page;
and performing screenshot on the web response page to obtain a web response page picture.
3. The asset classification method based on image recognition according to claim 1, wherein the step of obtaining the image class confidence of the web response page image according to the image feature and the class confidence of the image feature based on the preset image class confidence calculation rule comprises:
the picture category confidence calculation rule comprises a first rule;
the first rule includes:
comparing the category confidence of all image features in the web response page picture with a first preset confidence;
if the category confidence of the image features is greater than a first preset confidence;
and regarding the category confidence of the image features as the picture category confidence of the web response page picture.
4. The asset classification method based on image recognition according to claim 3, wherein the step of obtaining the image class confidence of the web response page image according to the image feature and the class confidence of the image feature based on the preset image class confidence calculation rule comprises:
the picture category confidence degree calculation rule comprises a second rule;
the second rule includes:
when the category confidence of all image features in the web response page picture is less than a first preset confidence;
comparing the category confidence of all image features in the web response page picture;
selecting image features with first high class confidence coefficient and second high class confidence coefficient, calculating the confidence coefficient difference of the two image features, comparing the difference with a second preset confidence coefficient, and if the difference is greater than the confidence coefficient threshold and the class information of the two image features is different, marking the image features with first high class confidence coefficient as first features to be calculated;
acquiring category information of a first feature to be calculated;
acquiring the class information of the rest image features, comparing the class information with the class information of the first feature to be calculated, reserving the image features with the same class information in the rest image features, selecting the image feature with the first high confidence coefficient, and marking the image feature as the second feature to be calculated;
and obtaining the image category confidence of the web response page image according to the category confidence of the first feature to be calculated and the category confidence of the second feature to be calculated based on the calculation rule contained in the second rule.
5. The image recognition-based asset classification method according to claim 4, wherein said calculation rule comprises: m 1 ＝1-[(1-M 2 )×(1-M 3 )]；
Wherein M is 1 Indicates picture class confidence, M 2 Representing the class confidence of the first feature to be computed, M 3 And representing the class confidence of the second feature to be calculated.
6. The asset classification method based on image recognition according to claim 4, wherein the step of obtaining the image class confidence of the web response page image according to the image feature and the class confidence of the image feature based on the preset image class confidence calculation rule comprises:
the picture category confidence calculation rules comprise a third rule;
the third rule includes:
when the category confidence of all image features in the web response page picture is smaller than a first preset confidence, when the confidence difference between the image features with the first highest category confidence and the image features with the second highest category confidence are smaller than a confidence threshold, and the category information of the two image features is different;
marking the class information of the image features with the first high class confidence coefficient as first class information;
marking the class information of the image features with the second highest class confidence coefficient as second class information;
calculating the confidence coefficient of the image features corresponding to the first type of information to obtain a first confidence coefficient based on a second rule;
calculating the confidence of the image features corresponding to the second type of information based on a second rule to obtain a second confidence;
and calculating a difference value between the first confidence coefficient and the second confidence coefficient, comparing the difference value with a preset difference value threshold, and if the difference value is greater than the preset difference value threshold, regarding the first confidence coefficient as the image category confidence coefficient of the web response page image.
7. The asset classification method based on image recognition according to claim 6, wherein the step of obtaining the image class confidence of the web response page image according to the image feature and the class confidence of the image feature based on the preset image class confidence calculation rule comprises:
the picture category confidence calculation rule comprises a fourth rule;
the fourth rule includes:
when the difference value between the first confidence coefficient and the second confidence coefficient is smaller than a preset difference threshold value;
acquiring image features of the first three high levels of category confidence corresponding to the first category information;
based on a calculation rule contained in the fourth rule, obtaining proportion information of the image features according to the confidence degrees of the image features, and calculating a category confidence degree corresponding to the first category information according to the proportion information and the confidence degrees of the image features;
acquiring image features of the first three high category confidence degrees corresponding to the second category information;
based on the calculation rule contained in the fourth rule, obtaining the proportion information of the image features according to the confidence coefficient of the image features, and calculating the category confidence coefficient corresponding to the second category information according to the proportion information and the confidence coefficient of the image features;
and comparing the two category confidences, and taking the larger category confidence as the picture category confidence of the web response page picture.
8. An asset classification system based on image recognition, comprising:
the acquisition module (1) is used for acquiring a web response page picture of the video equipment;
the extraction module (2) is used for extracting the features of the web response page pictures by using a feature extraction technology to obtain image features; the calling module (3) is used for calling a preset confidence coefficient relation library and obtaining a category confidence coefficient corresponding to the image features according to the confidence coefficient relation library and the image features; the confidence coefficient relation library stores image characteristics, category confidence coefficients and corresponding relations between the image characteristics and the category confidence coefficients;
the calculation module (4) is used for obtaining the image category confidence coefficient of the web response page image according to the image characteristics and the category confidence coefficient of the image characteristics based on a preset image category confidence coefficient calculation rule;
and the determining module (5) is used for determining the category information of the video equipment according to the image category confidence of the web response page image.
9. An intelligent terminal, comprising a memory and a processor, the memory having stored thereon computer program instructions capable of being loaded by the processor and performing the method of any of claims 1-7.
10. A computer-readable storage medium, in which a computer program is stored which can be loaded by a processor and which executes the method according to any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210860571.7A CN115171048B (en) | 2022-07-21 | 2022-07-21 | Asset classification method, system, terminal and storage medium based on image recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210860571.7A CN115171048B (en) | 2022-07-21 | 2022-07-21 | Asset classification method, system, terminal and storage medium based on image recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115171048A true CN115171048A (en) | 2022-10-11 |
| CN115171048B CN115171048B (en) | 2023-03-17 |
Family
ID=83494312
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210860571.7A Active CN115171048B (en) | 2022-07-21 | 2022-07-21 | Asset classification method, system, terminal and storage medium based on image recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115171048B (en) |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107690660A (en) * | 2016-12-21 | 2018-02-13 | 深圳前海达闼云端智能科技有限公司 | Image-recognizing method and device |
| US20180108358A1 (en) * | 2016-10-19 | 2018-04-19 | Mastercard International Incorporated | Voice Categorisation |
| CN108974043A (en) * | 2018-08-09 | 2018-12-11 | 铁路愿景欧洲有限公司 | Railroad track survey system |
| CN111177371A (en) * | 2019-12-05 | 2020-05-19 | 腾讯科技（深圳）有限公司 | Classification method and related device |
| CN111432206A (en) * | 2020-04-24 | 2020-07-17 | 腾讯科技（北京）有限公司 | Video definition processing method and device based on artificial intelligence and electronic equipment |
| CN111738042A (en) * | 2019-10-25 | 2020-10-02 | 北京沃东天骏信息技术有限公司 | Identification method, device and storage medium |
| CN112288014A (en) * | 2020-10-30 | 2021-01-29 | 云南电网有限责任公司电力科学研究院 | Data mining-based equipment full life cycle management method |
| CN112487270A (en) * | 2019-09-12 | 2021-03-12 | 北京白帽汇科技有限公司 | Method and device for asset classification and accuracy verification based on picture identification |
| CN112819815A (en) * | 2021-02-25 | 2021-05-18 | 汇鸿智能科技(辽宁)有限公司 | Metallographic structure identification method and device, computer equipment and storage medium |
| CN113780315A (en) * | 2020-06-09 | 2021-12-10 | 顺丰科技有限公司 | Image recognition method and device, network equipment and computer readable storage medium |
| CN114677573A (en) * | 2022-05-30 | 2022-06-28 | 上海捷勃特机器人有限公司 | Visual classification method, system, device and computer readable medium |
-
2022
- 2022-07-21 CN CN202210860571.7A patent/CN115171048B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20180108358A1 (en) * | 2016-10-19 | 2018-04-19 | Mastercard International Incorporated | Voice Categorisation |
| CN107690660A (en) * | 2016-12-21 | 2018-02-13 | 深圳前海达闼云端智能科技有限公司 | Image-recognizing method and device |
| CN108974043A (en) * | 2018-08-09 | 2018-12-11 | 铁路愿景欧洲有限公司 | Railroad track survey system |
| CN112487270A (en) * | 2019-09-12 | 2021-03-12 | 北京白帽汇科技有限公司 | Method and device for asset classification and accuracy verification based on picture identification |
| CN111738042A (en) * | 2019-10-25 | 2020-10-02 | 北京沃东天骏信息技术有限公司 | Identification method, device and storage medium |
| CN111177371A (en) * | 2019-12-05 | 2020-05-19 | 腾讯科技（深圳）有限公司 | Classification method and related device |
| CN111432206A (en) * | 2020-04-24 | 2020-07-17 | 腾讯科技（北京）有限公司 | Video definition processing method and device based on artificial intelligence and electronic equipment |
| CN113780315A (en) * | 2020-06-09 | 2021-12-10 | 顺丰科技有限公司 | Image recognition method and device, network equipment and computer readable storage medium |
| CN112288014A (en) * | 2020-10-30 | 2021-01-29 | 云南电网有限责任公司电力科学研究院 | Data mining-based equipment full life cycle management method |
| CN112819815A (en) * | 2021-02-25 | 2021-05-18 | 汇鸿智能科技(辽宁)有限公司 | Metallographic structure identification method and device, computer equipment and storage medium |
| CN114677573A (en) * | 2022-05-30 | 2022-06-28 | 上海捷勃特机器人有限公司 | Visual classification method, system, device and computer readable medium |
Non-Patent Citations (1)
| Title |
|---|
| 曹治国 等: "基于特征的景象匹配置信度融合计算" * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115171048B (en) | 2023-03-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110851835A (en) | Image model detection method and device, electronic equipment and storage medium | |
| CN107209942B (en) | Object detection method and image retrieval system | |
| JP2001325593A (en) | Duplicate picture detecting method in automatic albuming system | |
| CN111738120A (en) | Person identification method, person identification device, electronic equipment and storage medium | |
| CN111191591B (en) | Watermark detection and video processing method and related equipment | |
| CN114650447B (en) | Method and device for determining video content abnormality degree and computing equipment | |
| CN110796053B (en) | Video detection method and device, electronic equipment and computer readable storage medium | |
| CN110287361B (en) | Figure picture screening method and device | |
| CN112990792B (en) | Method and device for automatically detecting infringement risk and electronic equipment | |
| CN113076961B (en) | Image feature library updating method, image detection method and device | |
| CN112906671B (en) | Method and device for identifying false face-examination picture, electronic equipment and storage medium | |
| Hu et al. | Fast source camera identification using matching signs between query and reference fingerprints | |
| CN112884866B (en) | Coloring method, device, equipment and storage medium for black-and-white video | |
| CN112132892A (en) | Target position marking method, device and equipment | |
| CN114155363A (en) | Converter station vehicle identification method and device, computer equipment and storage medium | |
| CN111803956B (en) | Method and device for determining game plug-in behavior, electronic equipment and storage medium | |
| CN113495886A (en) | Method and device for detecting pollution sample data for model training | |
| CN115171048B (en) | Asset classification method, system, terminal and storage medium based on image recognition | |
| CN111340139A (en) | Method and device for judging complexity of image content | |
| CN111611388A (en) | Account classification method, device and equipment | |
| CN111368128A (en) | Target picture identification method and device and computer readable storage medium | |
| CN110969602B (en) | Image definition detection method and device | |
| CN111062388B (en) | Advertisement character recognition method, system, medium and equipment based on deep learning | |
| CN110674830B (en) | Image privacy identification method and device, computer equipment and storage medium | |
| CN114550062A (en) | Method and device for determining moving object in image, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |