CN115100450B

CN115100450B - Intelligent traffic brand automobile big data detection method and system based on artificial intelligence

Info

Publication number: CN115100450B
Application number: CN202211028939.XA
Authority: CN
Inventors: 黄建浩
Original assignee: China Highway Engineering Consultants Corp; CHECC Data Co Ltd
Current assignee: China Highway Engineering Consultants Corp; CHECC Data Co Ltd
Priority date: 2022-08-26
Filing date: 2022-08-26
Publication date: 2023-01-06
Anticipated expiration: 2042-08-26
Also published as: CN115100450A

Abstract

The invention provides a method and a system for detecting big data of an intelligent traffic brand automobile based on artificial intelligence, and relates to the technical field of data detection. Acquiring a plurality of images to be detected; detecting the image to be detected by adopting a multi-target detection mutual inspection method; carrying out target detection by using a target detection method to generate a plurality of trademark areas to be detected; respectively calculating the multi-region entropy similarity of the trademark region to be detected and each trademark template image; distinguishing the brand area to be detected by using the corresponding brand detection model; calculating the similarity between the corresponding trademark template image and the trademark area to be detected based on a matching method of a multi-edge detection operator according to the trademark judgment result; obtaining a detection result of the trademark area to be detected according to the similarity; further counting to obtain the data of each brand of automobile in the image to be detected; and obtaining brand automobile detection data according to the statistics of the data of each brand automobile in each image to be detected. The accuracy of detecting the brand of the automobile trademark is remarkably improved.

Description

Intelligent traffic brand automobile big data detection method and system based on artificial intelligence

Technical Field

The invention relates to the technical field of data detection, in particular to a method and a system for detecting big data of an intelligent traffic brand automobile based on artificial intelligence.

Background

With the gradual improvement of living conditions of people, more and more people buy branded automobiles as transportation tools. Along with the obvious reinforcing of people's purchasing power, more and more brand car drops into market, if can make statistics of the big data of brand car in the road, not only can provide the reference for the consumer, can let the automobile manufacture enterprise master information such as market share, the rate of utilization of oneself car moreover.

The traditional method for detecting the big data of the brand automobile is usually completed by manpower, so that huge manpower resources are wasted, and the accuracy of a result cannot be guaranteed. Although the modern traffic system applies technologies such as target detection and image recognition to big data detection of brand automobiles, higher accuracy cannot be maintained, and with continuous progress of technology, the related technologies in the field of artificial intelligence are more and more mature, and important support can be provided for big data detection of brand automobiles, so that how to fully utilize the technology in the field of artificial intelligence is of great value and significance in providing an intelligent traffic brand automobile big data detection method based on artificial intelligence.

Disclosure of Invention

The invention aims to provide a method and a system for detecting intelligent traffic brand automobile big data based on artificial intelligence, which are used for solving the problem that the prior art applies technologies such as target detection, image recognition and the like to the big data detection of a brand automobile but cannot keep higher accuracy.

In a first aspect, an embodiment of the present application provides a method for detecting intelligent transportation brand automobile big data based on artificial intelligence, including the following steps:

acquiring a plurality of images of each road area as a plurality of images to be detected;

detecting the images to be detected by adopting a multi-target detection mutual-inspection method to generate a plurality of automobile detection images;

respectively carrying out target detection on each automobile detection image by using a target detection method to generate a plurality of trademark areas to be detected;

respectively calculating multi-region entropy similarity of the trademark region to be detected and each trademark template image in a preset vehicle trademark template library to obtain a plurality of multi-region entropy similarity results;

extracting a corresponding brand trademark detection model from a preset brand trademark detection model library according to each multi-region entropy value similarity result;

distinguishing the brand area to be detected by using the corresponding brand detection model to generate a brand distinguishing result;

extracting corresponding trademark template images from a preset vehicle trademark template library according to the result of the entropy similarity of each multi-region;

calculating the similarity between the corresponding trademark template image and the trademark area to be detected based on a matching method of a multi-edge detection operator according to the trademark judgment result;

obtaining a detection result of the trademark area to be detected according to the similarity between the corresponding trademark template image and the trademark area to be detected;

counting according to the detection result of each brand area to be detected to obtain the data of each brand automobile in the image to be detected;

and obtaining brand automobile detection data according to the statistics of the data of each brand automobile in each image to be detected.

In the implementation process, a plurality of images of each road area are obtained to serve as a plurality of images to be detected; detecting the image to be detected by adopting a multi-target detection mutual inspection method; performing target detection by using a target detection method to generate a plurality of trademark areas to be detected; respectively calculating the multi-region entropy similarity of the trademark region to be detected and each trademark template image; extracting a corresponding brand trademark detection model according to each multi-region entropy value similarity result; distinguishing the brand area to be detected by using the corresponding brand detection model; extracting corresponding trademark template images according to the result of the similarity of the multiple regions of entropy values; calculating the similarity between the corresponding trademark template image and the trademark area to be detected based on a matching method of a multi-edge detection operator according to the trademark judgment result; obtaining a detection result of the trademark area to be detected according to the similarity between the corresponding trademark template image and the trademark area to be detected; counting according to the detection result of each brand area to be detected to obtain the data of each brand automobile in the image to be detected; and obtaining brand automobile detection data according to the statistics of the data of each brand automobile in each image to be detected. The method has the advantages that the target detection is carried out on the image to be detected by utilizing the mutual inspection of the multi-target detection method, the automobile detection accuracy is remarkably improved, meanwhile, the part of the trademark to be detected can be directly judged by calculating the multi-region entropy similarity of the automobile trademark template image and the trademark region to be detected, the automobile trademark detection accuracy is remarkably improved, the computing resource consumption of the whole system is also reduced, when the score of the SVM model of the image to be detected of the trademark region is in a critical state, the SVM model is secondarily judged by utilizing a matching method based on a multi-edge detection operator, the problem that the SVM model judges the score edge image inaccurately is solved, and the automobile trademark detection accuracy is remarkably improved.

Based on the first aspect, in some embodiments of the present invention, the step of calculating the multi-region entropy similarity between the trademark region to be detected and each trademark template image in the preset vehicle trademark template library respectively to obtain a plurality of multi-region entropy similarity results includes the following steps:

equally dividing the trademark area to be detected and each trademark template image in a preset vehicle trademark template library respectively to obtain a plurality of equally divided areas of the trademark to be detected and a plurality of equally divided images of the trademark templates;

respectively calculating the entropy value of each trademark equally divided region to be detected and the entropy value of each trademark template equally divided image;

comparing the entropy value of each trademark equal-division area to be detected with the entropy value of the corresponding trademark template equal-division image to obtain a plurality of entropy value comparison results;

and obtaining the multi-region entropy similarity between the trademark region to be detected and each trademark template image according to the comparison result of the plurality of entropy values.

Based on the first aspect, in some embodiments of the present invention, the method further comprises the following steps:

calculating the gray value of each pixel point in the equally divided area of the trademark to be detected;

calculating the probability of each gray level appearing in the equally divided areas of the trademark to be detected according to the gray level value of each pixel point;

and calculating the entropy value of the equal partition area of the trademark to be detected by using a preset entropy value calculation formula according to the probability of the appearance of each gray level in the equal partition area of the trademark to be detected.

acquiring a trademark image of any brand as a positive training sample;

acquiring brand images of other brands except the brand as negative training samples;

respectively screening the positive training sample and the negative training sample to obtain a new positive training sample and a new negative training sample;

and training the new positive training sample and the new negative training sample by using a preset SVM model to obtain a brand trademark detection model corresponding to any brand.

Based on the first aspect, in some embodiments of the present invention, the following steps are further included:

performing significance detection on each trademark image in the positive training sample to obtain a plurality of significance detection results;

screening the positive training samples according to a plurality of significance detection results to obtain a pre-screening positive training book;

carrying out peak signal-to-noise ratio detection on each trademark image in the pre-screened training book to generate a plurality of peak signal-to-noise ratios;

and screening the pre-screened positive training book according to the plurality of peak signal-to-noise ratios to obtain a new positive training book.

Based on the first aspect, in some embodiments of the present invention, the step of calculating, according to the trademark discrimination result, the similarity between the corresponding trademark template image and the trademark region to be detected based on the matching method of the multi-edge detection operator includes the following steps:

judging whether the trademark judgment result is in a critical state, if so, calculating the similarity between the corresponding trademark template image and the trademark area to be detected based on a matching method of a multi-edge detection operator; if not, obtaining the detection result of the trademark area to be detected according to the trademark judgment result.

Based on the first aspect, in some embodiments of the present invention, the step of calculating the similarity between the corresponding trademark template image and the to-be-detected trademark region based on the matching method of the multi-edge detection operator includes the following steps:

respectively filtering the corresponding trademark template image and the trademark area to be detected by using a Canny operator to generate a first trademark template image filtering result and a first trademark area to be detected filtering result;

respectively carrying out Hash coding on the filtering result of the first trademark template image and the filtering result of the first trademark area to be detected to obtain a first trademark template image code and a first trademark area to be detected;

according to the first trademark template image code and the first to-be-detected trademark area code, calculating the similarity between the corresponding trademark template image and the to-be-detected trademark area by using the Euclidean distance to obtain a first similarity;

respectively filtering the corresponding trademark template image and the trademark area to be detected by using a Prewitt operator to generate a second trademark template image filtering result and a second trademark area to be detected filtering result;

respectively carrying out Hash coding on the filtering result of the second trademark template image and the filtering result of the second trademark area to be detected to obtain a second trademark template image code and a second trademark area to be detected code;

calculating the similarity between the corresponding trademark template image and the trademark area to be detected by using the Euclidean distance according to the second trademark template image code and the second trademark area to be detected to obtain a second similarity;

respectively filtering the corresponding trademark template image and the trademark area to be detected by using a Sobel operator to generate a third trademark template image filtering result and a third trademark area to be detected filtering result;

performing hash coding on the filtering result of the third trademark template image and the filtering result of the third trademark area to be detected respectively to obtain a third trademark template image code and a third trademark area to be detected code;

calculating the similarity between the corresponding trademark template image and the trademark area to be detected by using the Euclidean distance according to the third trademark template image code and the third trademark area to be detected to obtain a third similarity;

and obtaining the similarity between the corresponding trademark template image and the to-be-detected trademark area according to the first similarity, the second similarity and the third similarity.

In a second aspect, an embodiment of the present application provides an intelligent transportation brand automobile big data detection system based on artificial intelligence, including:

the to-be-detected image acquisition module is used for acquiring a plurality of images of each road area as a plurality of to-be-detected images;

the first detection module is used for detecting the image to be detected by adopting a multi-target detection mutual inspection method to generate a plurality of automobile detection images;

the second detection module is used for respectively carrying out target detection on each automobile detection image by using a target detection method to generate a plurality of trademark areas to be detected;

the multi-region entropy similarity calculation module is used for calculating multi-region entropy similarity of the trademark region to be detected and each trademark template image in a preset vehicle trademark template library respectively to obtain a plurality of multi-region entropy similarity results;

the brand trademark detection model extraction module is used for extracting corresponding brand trademark detection models from a preset brand trademark detection model library according to the multi-region entropy value similarity result;

the first trademark distinguishing module is used for distinguishing the to-be-detected trademark area by using the corresponding trademark detection model to generate a trademark distinguishing result;

the trademark template image module is used for extracting corresponding trademark template images from a preset vehicle trademark template library according to the result of the entropy value similarity of each multi-region;

the similarity calculation module is used for calculating the similarity between the corresponding trademark template image and the trademark area to be detected based on the matching method of the multi-edge detection operator according to the trademark judgment result;

the detection result module is used for obtaining the detection result of the trademark area to be detected according to the similarity between the corresponding trademark template image and the trademark area to be detected;

the first statistic module is used for counting to obtain the data of each brand of automobile in the image to be detected according to the detection result of each brand area to be detected;

and the second statistical module is used for obtaining brand automobile detection data according to the statistics of the data of each brand automobile in each image to be detected.

In the implementation process, a plurality of images of each road area are acquired as a plurality of images to be detected through an image acquisition module to be detected; the first detection module detects the image to be detected by adopting a multi-target detection mutual inspection method; the second detection module performs target detection by using a target detection method to generate a plurality of trademark areas to be detected; the multi-region entropy similarity calculation module calculates the multi-region entropy similarity between the trademark region to be detected and each trademark template image; the brand trademark detection model extraction module extracts a corresponding brand trademark detection model according to each multi-region entropy value similarity result; the first trademark distinguishing module distinguishes the trademark area to be detected by utilizing a corresponding trademark detection model; the trademark template image module extracts corresponding trademark template images according to the result of the similarity of the multi-region entropy values; the similarity calculation module calculates the similarity between the corresponding trademark template image and the trademark area to be detected based on the matching method of the multi-edge detection operator according to the trademark judgment result; the detection result module obtains a detection result of the trademark area to be detected according to the similarity between the corresponding trademark template image and the trademark area to be detected; the first statistical module is used for obtaining the data of each brand of automobile in the image to be detected according to the detection result of each brand area to be detected; and the second statistical module is used for obtaining brand automobile detection data according to the statistics of the data of each brand automobile in each image to be detected. The method has the advantages that the target detection is carried out on the image to be detected by utilizing the multi-target detection method, the automobile detection accuracy is remarkably improved, meanwhile, the similarity of multi-region entropy values of the automobile trademark template image and the trademark region to be detected is calculated, part of the trademark to be detected can be directly judged, the automobile trademark detection accuracy is remarkably improved, the consumption of computing resources of the whole system is reduced, when the score of an SVM model of the trademark region image to be detected is in a critical state, secondary judgment is carried out on the image by utilizing a matching method based on a multi-edge detection operator, the problem that the score edge image is not accurately judged by the SVM model is solved, and the automobile trademark detection accuracy is remarkably improved.

In a third aspect, an embodiment of the present application provides an electronic device, which includes a memory for storing one or more programs; a processor. The program or programs, when executed by a processor, implement the method of any of the first aspects as described above.

In a fourth aspect, embodiments of the present application provide a computer-readable storage medium, on which a computer program is stored, which, when executed by a processor, implements the method according to any one of the first aspect described above.

The embodiment of the invention at least has the following advantages or beneficial effects:

the embodiment of the invention provides a method and a system for detecting big data of an intelligent traffic brand automobile based on artificial intelligence, wherein a plurality of images of each road area are obtained to be used as a plurality of images to be detected; detecting the image to be detected by adopting a multi-target detection mutual inspection method; performing target detection by using a target detection method to generate a plurality of trademark areas to be detected; respectively calculating the multi-region entropy similarity of the trademark region to be detected and each trademark template image; extracting corresponding brand trademark detection models according to the result of the similarity of the multi-region entropy values; distinguishing the brand area to be detected by using the corresponding brand detection model; extracting corresponding trademark template images according to the multi-region entropy similarity result; according to the trademark discrimination result, calculating the similarity between the corresponding trademark template image and the trademark area to be detected based on a matching method of a multi-edge detection operator; obtaining a detection result of the trademark area to be detected according to the similarity between the corresponding trademark template image and the trademark area to be detected; counting according to the detection result of each brand area to be detected to obtain the data of each brand automobile in the image to be detected; and obtaining brand automobile detection data according to the statistics of the data of each brand automobile in each image to be detected. The method has the advantages that the target detection is carried out on the image to be detected by utilizing the mutual inspection of the multi-target detection method, the automobile detection accuracy is remarkably improved, meanwhile, the part of the trademark to be detected can be directly judged by calculating the multi-region entropy similarity of the automobile trademark template image and the trademark region to be detected, the automobile trademark detection accuracy is remarkably improved, the computing resource consumption of the whole system is also reduced, when the score of the SVM model of the image to be detected of the trademark region is in a critical state, the SVM model is secondarily judged by utilizing a matching method based on a multi-edge detection operator, the problem that the SVM model judges the score edge image inaccurately is solved, and the automobile trademark detection accuracy is remarkably improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.

Fig. 1 is a flowchart of a method for detecting big data of an intelligent transportation brand automobile based on artificial intelligence according to an embodiment of the present invention;

FIG. 2 is a detailed step of step S140 according to an embodiment of the present invention;

FIG. 3 is a diagram illustrating an entropy calculation process provided by an embodiment of the invention;

fig. 4 is a block diagram of a structure of a smart transportation brand automobile big data detection system based on artificial intelligence according to an embodiment of the present invention;

fig. 5 is a block diagram of an electronic device according to an embodiment of the present invention.

Icon: 110-an image acquisition module to be detected; 120-a first detection module; 130-a second detection module; 140-a multi-region entropy similarity calculation module; 150-brand trademark detection model extraction module; 160-first trademark judging module; 170-trademark template image module; 180-similarity calculation module; 190-a detection result module; 200-a first statistical module; 210-a second statistics module; 101-a memory; 102-a processor; 103-communication interface.

Detailed Description

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. The components of the embodiments of the present application, generally described and illustrated in the figures herein, can be arranged and designed in a wide variety of different configurations.

Thus, the following detailed description of the embodiments of the present application, as presented in the figures, is not intended to limit the scope of the claimed application, but is merely representative of selected embodiments of the application. 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 application.

It should be noted that: like reference numbers and letters refer to like items in the following figures, and thus, once an item is defined in one figure, it need not be further defined and explained in subsequent figures. Meanwhile, in the description of the present application, the terms "first", "second", and the like are used only for distinguishing the description, and are not to be construed as indicating or implying relative importance.

It should be noted that, in this document, relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrases "comprising a component of' 8230; \8230;" does not exclude the presence of additional identical elements in the process, method, article, or apparatus that comprises the element.

Examples

Some embodiments of the present application will be described in detail below with reference to the accompanying drawings. The embodiments and features of the embodiments described below can be combined with one another without conflict.

Referring to fig. 1, fig. 1 is a flowchart of a method for detecting big data of an intelligent transportation brand automobile based on artificial intelligence according to an embodiment of the present invention. The intelligent traffic brand automobile big data detection method based on artificial intelligence comprises the following steps:

step S110: acquiring a plurality of images of each road area as a plurality of images to be detected; for a specific traffic road, a road camera is used for photographing to obtain an image to be detected, and photographing is performed at regular intervals in a plurality of road areas in a certain city, for example, all main roads are taken as main roads, and photographing is performed once every 10 seconds, so that a plurality of images of each road area can be obtained.

Step S120: detecting the images to be detected by adopting a multi-target detection mutual inspection method to generate a plurality of automobile detection images; the detection is performed to detect the vehicles in the image to be detected and count the number of the vehicles. The multi-target detection mutual-inspection method refers to detection by using multiple target detection methods, and at least two methods, for example, target detection is performed on the image by using a target detection method A and a target detection method B respectively. If the target detection quantity of the method A and the method B is inconsistent, the higher quantity is taken as the standard. For example: the target detection method A detects 60 automobiles, and the target detection method B detects 62 automobiles, based on 62 automobiles. The target detection can be realized by adopting the existing target detection method, and details are not repeated here.

Step S130: respectively carrying out target detection on each automobile detection image by using a target detection method to generate a plurality of trademark areas to be detected; for any detected vehicle in the image, the trademark area of the automobile is detected by using an object detection method.

Step S140: respectively calculating multi-region entropy similarity of the trademark region to be detected and each trademark template image in a preset vehicle trademark template library to obtain a plurality of multi-region entropy similarity results; the preset vehicle trademark template library comprises commodity template images of various brands, and the commodity template images can be downloaded from official websites of various brands of automobiles.

Referring to fig. 2-3, fig. 2 is a detailed step of step S140 provided in an embodiment of the present invention, and fig. 3 is a diagram of a process of calculating entropy according to an embodiment of the present invention. The above-mentioned calculation of the similarity of the multi-region entropy value is to calculate the similarity of the multi-region entropy value between the trademark region to be detected and each trademark template image, specifically, the above-mentioned calculation process of the similarity of the multi-region entropy value includes the following steps:

firstly, respectively equally dividing a trademark area to be detected and each trademark template image in a preset vehicle trademark template library to obtain a plurality of trademark equally divided areas to be detected and a plurality of trademark template equally divided images; the equal division can be set according to the requirement, for example, the Toyota car trademark template and the trademark to be detected are respectively divided into 4 equal divisions.

Then, respectively calculating the entropy value of each trademark equal division area to be detected and the entropy value of each trademark template equal division image; in the above example, the entropy calculation is performed for 4 regions of two brand images, respectively.

Specifically, the entropy calculation process includes:

firstly, calculating the gray value of each pixel point in the equal division area of the trademark to be detected;

secondly, calculating according to the gray values of all the pixel points to obtain the probability of each gray value appearing in the equally divided areas of the trademark to be detected;

and thirdly, calculating the entropy value of the equal partition area of the trademark to be detected by using a preset entropy value calculation formula according to the probability of the occurrence of each gray level in the equal partition area of the trademark to be detected. The formula for calculating the entropy value is as follows:

wherein, in the step (A),

for each gray level to beThe probability of occurrence in the trademark equally divided region is detected.

Then, comparing the entropy value of each trademark equal-division area to be detected with the entropy value of the corresponding trademark template equal-division image to obtain a plurality of entropy value comparison results; the comparison is to compare the corresponding regions, such as upper left to upper left, lower right to lower right.

And finally, obtaining the multi-region entropy similarity between the trademark region to be detected and each trademark template image according to the plurality of entropy comparison results. Judging the similarity of multiple regions of entropy values according to the result of entropy value comparison, and if the similarity of the multiple regions of entropy values is low, directly determining that the vehicle trademark is not the corresponding brand trademark; if the multi-region entropy similarity is high, the next step can be continued.

Step S150: extracting corresponding brand trademark detection models from a preset brand trademark detection model library according to the result of the entropy value similarity of each multi-region; the preset brand trademark detection model library comprises brand trademark detection models of vehicles of various brands, and for example, if the entropy similarity between the brand trademark detection models and the multi-region entropy value of the Toyota vehicle is high, the Toyota trademark detection models can be extracted.

The process for constructing the brand trademark detection model comprises the following steps: firstly, acquiring a trademark image of any brand as a positive training sample; then, acquiring images of other brands except the brand as negative training samples; for example: a plurality of Toyota car trademark images are selected as positive training samples, network downloading and manual collection can be performed, other brand trademarks are selected as negative training samples, and network downloading and manual collection can be performed. In order to obtain high-quality positive and negative training samples, the positive and negative training samples can be further optimized, namely the positive training sample and the negative training sample are respectively screened to obtain a new positive training sample and a new negative training sample; and finally, training the new positive training sample and the new negative training sample by using a preset SVM model to obtain a brand trademark detection model corresponding to any brand.

The following is a detailed description of a preferred process for a positive training sample, the preferred process for a negative training sample is the same as the preferred process for a positive training sample, and the preferred process for the positive training sample includes the following steps:

firstly, performing significance detection on each trademark image in a positive training sample to obtain a plurality of significance detection results; the significance detection can be realized by adopting the prior art, and is not described herein again.

Secondly, screening the positive training samples according to a plurality of significance detection results to obtain a pre-screening positive training book; if a training sample has few or very small significant regions, it is directly considered as a low quality training sample.

Thirdly, detecting the peak signal-to-noise ratio of each trademark image in the pre-screened training book to generate a plurality of peak signal-to-noise ratios; the peak snr detection can be implemented by using the prior art, and will not be described herein.

And fourthly, screening the pre-screened positive training book according to the plurality of peak value signal-to-noise ratios to obtain a new positive training book. If the peak signal-to-noise ratio of a certain training sample is low, the training sample is directly considered as a low-quality training sample. Those that are not considered as low quality training samples are marked as high quality training samples.

Step S160: distinguishing the brand area to be detected by using the corresponding brand detection model to generate a brand distinguishing result; inputting the trademark area image to be detected into the corresponding trademark detection model, and obtaining SVM model score as a trademark discrimination result. For example, if the score of the SVM model of the to-be-detected trademark area is high, the area image is directly determined to be a Toyota vehicle trademark image; and if the score of the SVM model of the to-be-detected trademark area is low, directly determining that the area image is a non-Toyota vehicle trademark image.

Step S170: extracting corresponding trademark template images from a preset vehicle trademark template library according to the result of the entropy similarity of each multi-region; the vehicle trademark template library comprises vehicle trademark templates of various brands.

Step S180: calculating the similarity between the corresponding trademark template image and the trademark area to be detected based on a matching method of a multi-edge detection operator according to the trademark judgment result; specifically, judging whether the trademark judging result is in a critical state, if so, calculating the similarity between the corresponding trademark template image and the to-be-detected trademark area based on a matching method of a multi-edge detection operator; if not, obtaining the detection result of the trademark area to be detected according to the trademark judgment result. For example: and when the score of the SVM model of the to-be-detected trademark area image is in a critical state, carrying out secondary judgment on the score. Specifically, the similarity between the Toyota brand template image and the brand area image to be detected is calculated by using a matching method based on a multi-edge detection operator, and the brand area image to be detected is judged according to the similarity.

The process of calculating the similarity between the corresponding trademark template image and the trademark area to be detected by the matching method based on the multi-edge detection operator comprises the following steps:

firstly, respectively filtering a corresponding trademark template image and a trademark area to be detected by using a Canny operator to generate a first trademark template image filtering result and a first trademark area to be detected filtering result;

then, respectively carrying out Hash coding on the filtering result of the first trademark template image and the filtering result of the first trademark area to be detected to obtain a first trademark template image code and a first trademark area to be detected code;

then, according to the first trademark template image code and the first trademark area code to be detected, calculating the similarity between the corresponding trademark template image and the trademark area to be detected by using the Euclidean distance to obtain a first similarity;

then, respectively filtering the corresponding trademark template image and the trademark area to be detected by using a Prewitt operator to generate a second trademark template image filtering result and a second trademark area to be detected filtering result;

then, respectively carrying out Hash coding on the filtering result of the second trademark template image and the filtering result of the second trademark area to be detected to obtain a second trademark template image code and a second trademark area to be detected code;

then, according to the second trademark template image code and the second to-be-detected trademark area code, calculating the similarity between the corresponding trademark template image and the to-be-detected trademark area by using the Euclidean distance to obtain a second similarity;

then, respectively filtering the corresponding trademark template image and the trademark area to be detected by using a Sobel operator to generate a third trademark template image filtering result and a third trademark area to be detected filtering result;

then, hash coding is respectively carried out on the filtering result of the third trademark template image and the filtering result of the third trademark area to be detected, and a third trademark template image code and a third trademark area code to be detected are obtained;

then, according to the third trademark template image code and the third to-be-detected trademark area code, calculating the similarity between the corresponding trademark template image and the to-be-detected trademark area by using the Euclidean distance to obtain a third similarity;

and finally, obtaining the similarity between the corresponding trademark template image and the trademark area to be detected according to the first similarity, the second similarity and the third similarity. When at least 2 of the first similarity, the second similarity and the third similarity are higher, the obtained similarity is higher. The similarity may be determined by comparing the first similarity, the second similarity, and the third similarity with a preset similarity threshold, and if the similarity is higher than the threshold, the similarity is considered to be higher, otherwise, the similarity is considered to be lower.

Step S190: obtaining a detection result of the trademark area to be detected according to the similarity between the corresponding trademark template image and the trademark area to be detected; and if the similarity is higher, the trademark area image to be detected is determined as the trademark image of the corresponding brand.

Step S200: counting according to the detection result of each brand area to be detected to obtain the data of each brand automobile in the image to be detected; the number of each brand of automobile in the shot image can be detected by the method, for example, the number of the automobiles is 16 Toyota automobiles, and the number of the automobiles is 3 Toyota automobiles.

Step S210: and counting according to the data of each brand automobile in each image to be detected to obtain brand automobile detection data. The number and occupancy of each brand of automobile can be estimated by taking pictures at regular intervals (e.g., once every 10 seconds) in a plurality of road areas (all major roads are dominant) in a city.

In the implementation process, a plurality of images of each road area are obtained to serve as a plurality of images to be detected; detecting the image to be detected by adopting a multi-target detection mutual inspection method; carrying out target detection by using a target detection method to generate a plurality of trademark areas to be detected; respectively calculating the multi-region entropy similarity of the trademark region to be detected and each trademark template image; extracting a corresponding brand trademark detection model according to each multi-region entropy value similarity result; distinguishing the brand area to be detected by using the corresponding brand detection model; extracting corresponding trademark template images according to the multi-region entropy similarity result; calculating the similarity between the corresponding trademark template image and the trademark area to be detected based on a matching method of a multi-edge detection operator according to the trademark judgment result; obtaining a detection result of the trademark area to be detected according to the similarity between the corresponding trademark template image and the trademark area to be detected; counting according to the detection result of each brand area to be detected to obtain the data of each brand automobile in the image to be detected; and obtaining brand automobile detection data according to the statistics of the data of each brand automobile in each image to be detected. The method has the advantages that the target detection is carried out on the image to be detected by utilizing the mutual inspection of the multi-target detection method, the automobile detection accuracy is remarkably improved, meanwhile, the part of the trademark to be detected can be directly judged by calculating the multi-region entropy similarity of the automobile trademark template image and the trademark region to be detected, the automobile trademark detection accuracy is remarkably improved, the computing resource consumption of the whole system is also reduced, when the score of the SVM model of the image to be detected of the trademark region is in a critical state, the SVM model is secondarily judged by utilizing a matching method based on a multi-edge detection operator, the problem that the SVM model judges the score edge image inaccurately is solved, and the automobile trademark detection accuracy is remarkably improved.

Based on the same inventive concept, the invention further provides an intelligent transportation brand automobile big data detection system based on artificial intelligence, please refer to fig. 4, and fig. 4 is a structural block diagram of the intelligent transportation brand automobile big data detection system based on artificial intelligence provided by the embodiment of the invention. This wisdom traffic brand car big data detection system based on artificial intelligence includes:

an image to be detected acquisition module 110, configured to acquire multiple images of each road area as multiple images to be detected;

the first detection module 120 is configured to detect an image to be detected by using a multi-target detection mutual-inspection method, and generate a plurality of automobile detection images;

the second detection module 130 is configured to perform target detection on each automobile detection image by using a target detection method, so as to generate a plurality of trademark areas to be detected;

the multi-region entropy similarity calculation module 140 is configured to calculate multi-region entropy similarities between the trademark region to be detected and each trademark template image in a preset vehicle trademark template library, and obtain multiple multi-region entropy similarity results;

the brand trademark detection model extraction module 150 is used for extracting corresponding brand trademark detection models from a preset brand trademark detection model library according to the result of the entropy value similarity of each multi-region;

the first trademark distinguishing module 160 is configured to distinguish the to-be-detected trademark area by using the corresponding trademark detection model, and generate a trademark distinguishing result;

a trademark template image module 170, configured to extract a corresponding trademark template image from a preset vehicle trademark template library according to each multi-region entropy similarity result;

the similarity calculation module 180 is used for calculating the similarity between the corresponding trademark template image and the trademark area to be detected based on the matching method of the multi-edge detection operator according to the trademark judgment result;

the detection result module 190 is configured to obtain a detection result of the to-be-detected trademark area according to similarity between the corresponding trademark template image and the to-be-detected trademark area;

the first statistical module 200 is used for obtaining the data of each brand of automobile in the image to be detected according to the detection result of each brand area to be detected;

and the second statistical module 210 is configured to obtain brand automobile detection data according to statistics of each brand automobile data in each image to be detected.

In the implementation process, the to-be-detected image acquisition module 110 is used for acquiring a plurality of images of each road area as a plurality of to-be-detected images; the first detection module 120 detects the image to be detected by adopting a multi-target detection mutual inspection method; the second detection module 130 performs target detection by using a target detection method to generate a plurality of trademark areas to be detected; the multi-region entropy similarity calculation module 140 calculates the multi-region entropy similarity between the trademark region to be detected and each trademark template image; the brand trademark detection model extraction module 150 extracts a corresponding brand trademark detection model according to each multi-region entropy value similarity result; the first trademark discrimination module 160 discriminates the trademark area to be detected by using the corresponding trademark detection model; the trademark template image module 170 extracts a corresponding trademark template image according to each multi-region entropy similarity result; the similarity calculation module 180 calculates the similarity between the corresponding trademark template image and the trademark area to be detected based on the matching method of the multi-edge detection operator according to the trademark discrimination result; the detection result module 190 obtains a detection result of the trademark area to be detected according to the similarity between the corresponding trademark template image and the trademark area to be detected; the first statistical module 200 obtains the data of each brand of automobile in the image to be detected according to the detection result of each brand area to be detected; the second statistical module 210 obtains brand automobile detection data according to statistics of each brand automobile data in each image to be detected. The method has the advantages that the target detection is carried out on the image to be detected by utilizing the mutual inspection of the multi-target detection method, the automobile detection accuracy is remarkably improved, meanwhile, the part of the trademark to be detected can be directly judged by calculating the multi-region entropy similarity of the automobile trademark template image and the trademark region to be detected, the automobile trademark detection accuracy is remarkably improved, the computing resource consumption of the whole system is also reduced, when the score of the SVM model of the image to be detected of the trademark region is in a critical state, the SVM model is secondarily judged by utilizing a matching method based on a multi-edge detection operator, the problem that the SVM model judges the score edge image inaccurately is solved, and the automobile trademark detection accuracy is remarkably improved.

Referring to fig. 5, fig. 5 is a schematic structural block diagram of an electronic device according to an embodiment of the present disclosure. The electronic device comprises a memory 101, a processor 102 and a communication interface 103, wherein the memory 101, the processor 102 and the communication interface 103 are electrically connected to each other directly or indirectly to realize data transmission or interaction. For example, the components may be electrically connected to each other via one or more communication buses or signal lines. The memory 101 may be used to store software programs and modules, such as program instructions/modules corresponding to an artificial intelligence-based intelligent transportation brand automobile big data detection system provided in an embodiment of the present application, and the processor 102 executes various functional applications and data processing by executing the software programs and modules stored in the memory 101. The communication interface 103 may be used for communicating signaling or data with other node devices.

The Memory 101 may be, but is not limited to, a Random Access Memory (RAM), a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Read-Only Memory (EPROM), an electrically Erasable Read-Only Memory (EEPROM), and the like.

The processor 102 may be an integrated circuit chip having signal processing capabilities. The Processor 102 may be a general-purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but also Digital Signal Processors (DSPs), application Specific Integrated Circuits (ASICs), field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components.

It will be appreciated that the configuration shown in fig. 5 is merely illustrative and that the electronic device may include more or fewer components than shown in fig. 5 or have a different configuration than shown in fig. 5. The components shown in fig. 5 may be implemented in hardware, software, or a combination thereof.

In the embodiments provided in the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The apparatus embodiments described above are merely illustrative, and for example, the flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

In addition, functional modules in the embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.

The functions, if implemented in the form of software functional modules and sold or used as a stand-alone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present application or portions thereof that substantially contribute to the prior art may be embodied in the form of a software product stored in a storage medium and including instructions for causing a computer device (which may be a personal computer, a server, or a network device) to execute all or part of the steps of the method according to the embodiments of the present application. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a magnetic disk or an optical disk, and other various media capable of storing program codes.

The above description is only a preferred embodiment of the present application and is not intended to limit the present application, and various modifications and changes may be made by those skilled in the art. Any modification, equivalent replacement, improvement and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.

It will be evident to those skilled in the art that the application is not limited to the details of the foregoing illustrative embodiments, and that the present application may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the application being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.

Claims

1. A smart traffic brand automobile big data detection method based on artificial intelligence is characterized by comprising the following steps:

detecting the images to be detected by adopting a multi-target detection mutual inspection method to generate a plurality of automobile detection images;

extracting corresponding trademark template images from a preset vehicle trademark template library according to the multi-region entropy similarity result;

according to the trademark discrimination result, calculating the similarity between the corresponding trademark template image and the to-be-detected trademark region based on the matching method of the multi-edge detection operator, wherein the similarity comprises the following steps: judging whether the trademark judging result is in a critical state, if so, calculating the similarity between the corresponding trademark template image and the to-be-detected trademark area based on the matching method of the multi-edge detection operator; if not, obtaining a detection result of the trademark area to be detected according to the trademark judgment result;

the step of calculating the similarity between the corresponding trademark template image and the trademark area to be detected based on the matching method of the multi-edge detection operator comprises the following steps: respectively filtering the corresponding trademark template image and the trademark area to be detected by using a Canny operator to generate a first trademark template image filtering result and a first trademark area to be detected filtering result; respectively carrying out Hash coding on the filtering result of the first trademark template image and the filtering result of the first trademark area to be detected to obtain a first trademark template image code and a first trademark area to be detected; according to the first trademark template image code and the first trademark area code to be detected, calculating the similarity between the corresponding trademark template image and the trademark area to be detected by utilizing the Euclidean distance to obtain a first similarity; respectively filtering the corresponding trademark template image and the trademark area to be detected by using a Prewitt operator to generate a second trademark template image filtering result and a second trademark area to be detected filtering result; respectively carrying out Hash coding on the filtering result of the second trademark template image and the filtering result of the second trademark area to be detected to obtain a second trademark template image code and a second trademark area to be detected code; calculating the similarity between the corresponding trademark template image and the trademark area to be detected by using the Euclidean distance according to the second trademark template image code and the second trademark area to be detected to obtain a second similarity; respectively filtering the corresponding trademark template image and the trademark area to be detected by using a Sobel operator to generate a third trademark template image filtering result and a third trademark area to be detected filtering result; respectively carrying out Hash coding on the filtering result of the third trademark template image and the filtering result of the third trademark area to be detected to obtain a third trademark template image code and a third trademark area to be detected; calculating the similarity between the corresponding trademark template image and the trademark area to be detected by using the Euclidean distance according to the third trademark template image code and the third trademark area to be detected to obtain a third similarity; obtaining the similarity between the corresponding trademark template image and the trademark area to be detected according to the first similarity, the second similarity and the third similarity;

2. The intelligent traffic brand automobile big data detection method based on artificial intelligence as claimed in claim 1, wherein the step of calculating the multi-region entropy similarity of each brand template image in the brand area to be detected and the preset vehicle brand template library respectively to obtain a plurality of multi-region entropy similarity results comprises the steps of:

respectively calculating the entropy value of each trademark equant region to be detected and the entropy value of each trademark template equant image;

and obtaining the multi-region entropy similarity between the trademark region to be detected and each trademark template image according to the plurality of entropy comparison results.

3. The intelligent transportation brand automobile big data detection method based on artificial intelligence as claimed in claim 2, further comprising the steps of:

4. The intelligent transportation brand automobile big data detection method based on artificial intelligence as claimed in claim 1, further comprising the steps of:

acquiring a trademark image of any brand as a positive training sample;

5. The intelligent transportation brand car big data detection method based on artificial intelligence of claim 4, further comprising the following steps:

screening the positive training sample according to a plurality of significance detection results to obtain a pre-screened positive training sample;

carrying out peak signal-to-noise ratio detection on each trademark image in the pre-screened training sample to generate a plurality of peak signal-to-noise ratios;

and screening the pre-screened positive training sample according to the plurality of peak signal-to-noise ratios to obtain a new positive training sample.

6. The utility model provides a wisdom traffic brand car big data detecting system based on artificial intelligence which characterized in that includes:

the multi-region entropy similarity calculation module is used for calculating multi-region entropy similarities of the trademark region to be detected and each trademark template image in a preset vehicle trademark template library respectively to obtain a plurality of multi-region entropy similarity results;

the trademark template image module is used for extracting corresponding trademark template images from a preset vehicle trademark template library according to the multi-region entropy similarity result;

the similarity calculation module is used for calculating the similarity between the corresponding trademark template image and the trademark area to be detected based on the matching method of the multi-edge detection operator according to the trademark judgment result, and comprises the following steps: judging whether the trademark judging result is in a critical state, if so, calculating the similarity between the corresponding trademark template image and the to-be-detected trademark area based on the matching method of the multi-edge detection operator; if not, obtaining a detection result of the trademark area to be detected according to the trademark judgment result;

the step of calculating the similarity between the corresponding trademark template image and the trademark area to be detected based on the matching method of the multi-edge detection operator comprises the following steps:

according to the third trademark template image code and the third trademark area code to be detected, calculating the similarity between the corresponding trademark template image and the trademark area to be detected by using the Euclidean distance to obtain a third similarity;

obtaining the similarity between the corresponding trademark template image and the trademark area to be detected according to the first similarity, the second similarity and the third similarity;

the first statistical module is used for obtaining the data of each brand of automobile in the image to be detected according to the detection result of each brand area to be detected;

7. An electronic device, comprising:

a memory for storing one or more programs;

a processor;

the one or more programs, when executed by the processor, implement the method of any of claims 1-5.

8. A computer-readable storage medium, on which a computer program is stored which, when being executed by a processor, carries out the method according to any one of claims 1-5.