CN114187568A - Road sign breakage detection method, apparatus and storage medium - Google Patents
Road sign breakage detection method, apparatus and storage medium Download PDFInfo
- Publication number
- CN114187568A CN114187568A CN202111530371.7A CN202111530371A CN114187568A CN 114187568 A CN114187568 A CN 114187568A CN 202111530371 A CN202111530371 A CN 202111530371A CN 114187568 A CN114187568 A CN 114187568A
- Authority
- CN
- China
- Prior art keywords
- image
- signage
- label
- similarity
- sign
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The embodiment of the application discloses a road sign breakage detection method and device and a storage medium, and belongs to the technical field of intelligent traffic. In the embodiment of the application, a plurality of label image sets are stored in advance, and the label image set matched with the type and the position information of the label to be detected is acquired from the plurality of label image sets according to the type and the position information of the label to be detected, so that the probability that the acquired label image set is the image set containing the label to be detected can be improved. Moreover, each of the plurality of signboard image sets comprises a plurality of signboard images which are acquired in advance and are actually installed on the road signboard, so that whether the to-be-detected signboard is damaged or not is judged by taking the acquired signboard image set as a reference according to the similarity between the image of the to-be-detected signboard and the acquired signboard image in the signboard image set, whether the to-be-detected signboard is damaged or not can be judged more accurately, and data support can be provided for timely replacement of the road signboard.
Description
Technical Field
The present disclosure relates to the field of traffic management, and more particularly, to a method and an apparatus for detecting breakage of a road sign, and a storage medium.
Background
In road traffic, road signs are one of the most basic and important safety facilities, and provide a road driving environment as safe and efficient as possible by transferring road-related conditions to motor vehicles, non-motor vehicles, and pedestrians using figures, symbols, lines, characters, etc. of a specific color. However, when the road sign is intact, the road sign cannot be guided well when cracks, rust, damages, etc. occur, and even traffic accidents may occur. Therefore, it is very important to detect and maintain the damage of the road sign at regular time.
Disclosure of Invention
The embodiment of the application provides a road sign breakage detection method, a road sign breakage detection device and a storage medium, which can more accurately detect the breakage condition of a road sign, so that data support is provided for timely replacement of the road sign. The technical scheme is as follows:
in one aspect, a method for detecting breakage of a road sign is provided, the method comprising:
acquiring the label type and position information of a label to be detected contained in the first label image;
acquiring a first sign image set from a plurality of stored sign image sets according to the sign type and the position information of the sign to be detected, wherein each sign image set comprises a plurality of sign images of a road sign acquired in advance;
and carrying out damage detection on the to-be-detected label according to the similarity between the first label image and the label image in the first label image set.
Optionally, each signage image set of the plurality of signage image sets corresponds to one signage type and location information;
the acquiring a first signage image set from a plurality of stored signage image sets according to the signage type and the position information of the signage to be detected comprises:
acquiring candidate sign image sets from the plurality of sign image sets, wherein the distance between the position information corresponding to each candidate sign image set and the position information of the sign to be detected is within a reference threshold range;
acquiring a second label image set from the acquired candidate label image set, wherein the label type corresponding to each second label image set is the same as the label type of the label to be detected;
a first set of placard images is selected from the second set of placard images.
Optionally, said selecting a first set of signage images from said second set of signage images comprises:
determining a similarity of the first signage image and each signage image in each second signage image set;
determining the similarity average value of the first label image and the label images in the corresponding second label image set according to the similarity of the first label image and each label image in each second label image set;
and selecting the second label image set with the maximum corresponding similarity average value from the plurality of second label image sets as the first label image set.
Optionally, the detecting damage to the to-be-detected signage according to the similarity between the first signage image and the signage image in the first signage image set includes:
acquiring a similarity reference value of the first label image set;
and carrying out damage detection on the to-be-detected label according to the similarity of the first label image and each label image in the first label image set and the similarity reference value.
Optionally, the detecting the breakage of the to-be-detected signage according to the similarity of the first signage image and the similarity reference value of each signage image in the first signage image set, includes:
determining an average value of the similarity of the first placard image and the first placard image set according to the similarity of the first placard image and the placard image in the first placard image set;
and if the average value of the similarity of the first label image and the first label image set is smaller than the reference value of the similarity, determining that the label to be detected is damaged.
Optionally, the detecting the breakage of the to-be-detected signage according to the similarity of the first signage image and the signage image in the first signage image set and the similarity reference value, further includes:
determining an absolute value of a difference between the average of the similarity of the first placard image and the first placard image set and the similarity reference value;
and determining the damage degree of the to-be-detected label according to the reference difference interval in which the absolute value of the difference is positioned.
Optionally, the method further comprises:
acquiring a plurality of sign images collected in advance and sign type and position information of a road sign in each sign image;
classifying the plurality of sign images according to the sign types and the position information of the road signs in each sign image to obtain a plurality of sign image sets, wherein each sign image set corresponds to one sign type and position information;
a similarity reference value is determined for each signage image set.
Optionally, the determining the similarity reference value of each signage image set includes:
acquiring an image quality score of each sign image in a third sign image set, wherein the third sign image set is a sign image set of any road sign;
and determining a similarity reference value of the third label image set according to the image quality score of each label image in the third label image set and the similarity between each label image and each label image in other label images except the third label image set.
Optionally, the determining a similarity reference value of the third signage image set according to the image quality score of each signage image in the third signage image set and the similarity between each signage image and each signage image in other signage images except the third signage image set comprises:
determining a similarity weighted average corresponding to each signage image according to the image quality score of each signage image in the third signage image set and the similarity between the corresponding signage image and each signage image in other signage images except the third signage image set;
determining a similarity weighted average value corresponding to the third placard image set according to the similarity weighted average value and the image quality score corresponding to each placard image in the third placard image set;
and deleting the signage images of which the similarity weighted average value corresponding to the third signage image set is smaller than the similarity weighted average value corresponding to the third signage image set, and determining the similarity reference value of the third signage image set according to the similarity weighted average value corresponding to each remaining signage image and the image quality score.
In another aspect, a road sign breakage detection device is provided, the device comprising:
the first acquisition module is used for acquiring the label type and the position information of the label to be detected contained in the first label image;
the second acquisition module is used for acquiring a first sign image set from a plurality of stored sign image sets according to the sign type and the position information of the sign to be detected, wherein each sign image set comprises a plurality of pre-acquired sign images of one road sign;
and the detection module is used for carrying out damage detection on the to-be-detected label according to the similarity between the first label image and the label image in the first label image set.
Optionally, each signage image set of the plurality of signage image sets corresponds to one signage type and location information;
the second obtaining module includes:
the first acquisition submodule is used for acquiring candidate sign image sets from the plurality of sign image sets, and the distance between the position information corresponding to each candidate sign image set and the position information of the sign to be detected is within a reference threshold range;
the second acquisition sub-module is used for acquiring a second label image set from the acquired candidate label image set, and the label type corresponding to each second label image set is the same as the label type of the label to be detected;
a selection sub-module for selecting a first set of signage images from the second set of signage images.
Optionally, the selection submodule is configured to:
determining a similarity of the first signage image and each signage image in each second signage image set;
determining the similarity average value of the first label image and the label images in the corresponding second label image set according to the similarity of the first label image and each label image in each second label image set;
and selecting the second label image set with the maximum corresponding similarity average value from the plurality of second label image sets as the first label image set.
Optionally, the detection module includes:
a third obtaining submodule, configured to obtain a similarity reference value of the first signage image set;
and the detection submodule is used for carrying out damage detection on the label to be detected according to the similarity of the first label image and each label image in the first label image set and the similarity reference value.
Optionally, the detection submodule is configured to:
determining an average value of the similarity of the first placard image and the first placard image set according to the similarity of the first placard image and the placard image in the first placard image set;
and if the average value of the similarity of the first label image and the first label image set is smaller than the reference value of the similarity, determining that the label to be detected is damaged.
Optionally, the detection submodule is further configured to:
determining an absolute value of a difference between the average of the similarity of the first placard image and the first placard image set and the similarity reference value;
and determining the damage degree of the to-be-detected label according to the reference difference interval in which the absolute value of the difference is positioned.
Optionally, the apparatus further comprises:
the third acquisition module is used for acquiring a plurality of sign images acquired in advance and sign types and position information of the road signs in each sign image;
the classification module is used for classifying the plurality of sign images according to the sign types and the position information of the road signs in each sign image to obtain a plurality of sign image sets, and each sign image set corresponds to one sign type and position information;
and the determining module is used for determining the similarity reference value of each label image set.
Optionally, the determining module includes:
a fourth obtaining submodule, configured to obtain an image quality score of each signage image in a third signage image set, where the third signage image set is a signage image set of any road signage;
and the determining submodule is used for determining the similarity reference value of the third label image set according to the image quality score of each label image in the third label image set and the similarity between each label image and each label image in other label images except the third label image set.
Optionally, the determining sub-module is configured to:
determining a similarity weighted average corresponding to each signage image according to the image quality score of each signage image in the third signage image set and the similarity between the corresponding signage image and each signage image in other signage images except the third signage image set;
determining a similarity weighted average value corresponding to the third placard image set according to the similarity weighted average value and the image quality score corresponding to each placard image in the third placard image set;
and deleting the signage images of which the similarity weighted average value corresponding to the third signage image set is smaller than the similarity weighted average value corresponding to the third signage image set, and determining the similarity reference value of the third signage image set according to the similarity weighted average value corresponding to each remaining signage image and the image quality score.
In another aspect, a computer device is provided, the computer device comprising a processor and a memory, the memory being configured to store a computer program, the processor being configured to execute the program stored on the memory to implement the steps of the road sign breakage detection method described above.
In another aspect, a computer-readable storage medium is provided, in which a computer program is stored, which, when executed by a computer, implements the steps of the road sign breakage detection method described above.
In another aspect, a computer program product comprising instructions is provided, which when run on a computer, causes the computer to perform the steps of the above-described road sign breakage detection method.
The beneficial effects brought by the technical scheme provided by the embodiment of the application at least comprise:
in the embodiment of the application, a plurality of label image sets are stored in advance, and the first label image set is obtained from the plurality of label image sets stored in advance according to the type and the position information of the label to be detected, so that the probability that the obtained label image set is the image set containing the label to be detected can be improved. Moreover, each of the plurality of signboard image sets comprises a plurality of signboard images which are acquired in advance and are actually installed on the road signboard, so that whether the to-be-detected signboard is damaged or not is judged by taking the acquired signboard image set as a reference according to the similarity between the image of the to-be-detected signboard and the acquired signboard image in the signboard image set, whether the to-be-detected signboard is damaged or not can be judged more accurately, and data support can be provided for timely replacement of the road signboard.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a system architecture diagram according to a road sign breakage detection method provided in an embodiment of the present application;
fig. 2 is a flowchart of a method for creating a road signboard base library according to an embodiment of the present application;
fig. 3 is a flowchart of a road sign breakage detection method according to an embodiment of the present application;
fig. 4 is a schematic view of a road sign breakage detection device provided in the embodiment of the present application;
fig. 5 is a schematic structural diagram of a server for detecting breakage of a road sign according to an embodiment of the present application.
Detailed Description
To make the objects, technical solutions and advantages of the present application more clear, embodiments of the present application will be described in further detail below with reference to the accompanying drawings.
Before explaining the embodiments of the present application in detail, a system architecture related to the embodiments of the present application will be described.
Fig. 1 is a system architecture diagram according to a road sign breakage detection method provided in an embodiment of the present application. As shown in fig. 1, the system includes an image capturing apparatus 101 and a server 102. Wherein the image capturing device 101 may communicate with the server 102 via a wireless network.
The image collecting device 101 is configured to collect a label image of a label to be detected and obtain position information when the label image is collected, and further send the label image and the position information to the server 102. For example, the image acquisition device 101 first acquires an original image including the to-be-detected label and including other backgrounds, and then performs background removal processing on the original image to obtain the label image of the to-be-detected label. Optionally, the image capturing device 101 may further identify the type of the sign to be detected according to the sign image, and further send the identified type of the sign to the server 102. After receiving the sign image, the sign type, and the position information of the sign to be detected, the server 102 may perform damage detection on the sign to be detected by the method provided in the embodiment of the present application based on a plurality of sign image sets stored in the road sign base in advance.
It should be noted that the image capturing device 101 may be a vehicle-mounted camera, and a GPS (Global Positioning System) device is mounted on the vehicle-mounted camera, so that when the vehicle-mounted camera 101 captures a label image of a label to be detected, the GPS device may be used to acquire position information corresponding to the label to be detected. Optionally, an AI (Artificial Intelligence) algorithm chip is further mounted on the vehicle-mounted camera, and based on this, after the vehicle-mounted camera 101 collects a label image of the label to be detected, the label image may be further processed to identify the label type of the label to be detected.
In addition, in the embodiment of the present application, before the breakage detection is performed on the sign to be detected, the image capturing device 101 may further capture each road sign mounted on the road multiple times, and acquire the position information at each capture, thereby obtaining multiple sign images of each road sign in each road sign and the position information of the road sign included in each sign image. Alternatively, the sign type of the road sign in each sign image may also be identified. The acquired plurality of placard images and the position information and the placard type corresponding to each placard image are sent to the server 102. The server 102 may establish a road sign base library by the method provided in the embodiment of the present application based on the received multiple sign images and the position information and the sign type corresponding to each sign image, so as to perform breakage detection on the sign to be detected based on the road sign base library in the following.
The server 102 may be an individual server, a server cluster, or a cloud platform, which is not limited in this embodiment of the present application.
In the embodiment of the application, the server may perform damage detection on the signage to be detected based on the signage image in the plurality of signage image sets stored in the road sign base library, and based on this, a process of establishing the road sign base library is introduced first. Fig. 2 is a flowchart of a method for creating a road sign base library according to an embodiment of the present disclosure. As shown in fig. 2, the method comprises the steps of:
step 201: sign type and position information of a road sign in a plurality of sign images and each sign image acquired in advance are acquired.
In the embodiment of the application, the image acquisition equipment acquires images for multiple times aiming at each road sign installed on a road, so that multiple sign images are obtained. The image capturing apparatus may acquire, at the time of capturing each sign image, position information at the time of capturing as position information of a road sign included in the sign image, and recognize, after capturing the sign image, a sign type of the road sign in the sign image. Thereafter, the image-capturing device may transmit the captured placard image and the position information and the placard type of the road placard included in the placard image to the server. Accordingly, the server receives the plurality of sign images transmitted from the image pickup device and the sign type and position information of the road sign in each sign image.
It should be noted that the image capturing device may be a vehicle-mounted camera. In this case, when a vehicle mounted with a vehicle-mounted camera patrols a plurality of road signs mounted on a road, the image of the plurality of road signs mounted on the road may be captured, and the image of the plurality of road signs may be captured once for one patrol, and the image of the plurality of road signs may be captured a plurality of times for a plurality of rounds, so that the collected sign images may be different depending on factors such as a shooting angle of view, a field of view, and an illumination intensity even for the same road sign. In addition, the vehicle-mounted camera may be equipped with a GPS device, so that the vehicle-mounted camera can perform positioning by the GPS device every time the vehicle-mounted camera performs image capturing, and obtain corresponding position information. Then, the vehicle-mounted camera can identify the sign type of the road sign included in the collected sign image through the AI model, so as to obtain the sign type of the road sign in the sign image.
Alternatively, the image capturing device may be other forms of intelligent devices with image capturing functions, such as an intelligent robot for road patrol, a drone, and the like. In this case, the image pickup apparatus can also obtain the plurality of sign images and the position information and the sign type of the road sign included in each sign image in the manner described above.
Alternatively, in a possible implementation manner, after the image capturing device captures the sign image of the road sign and obtains the position information of the road sign, the image capturing device may directly transmit the sign image and the corresponding position information to the server without recognizing the sign type of the road sign in the sign image. After receiving the sign image and the corresponding location information, the server may identify the sign type of the road sign in the sign image through the AI model.
In addition, in the embodiment of the present application, when the image capturing device captures an image of a road sign, in order to facilitate subsequent image processing, an image area including the road sign may be extracted from the captured image, and the extracted image area may be used as a sign image of the road sign.
Step 202: and classifying the plurality of sign images according to the sign types and the position information of the road signs in each sign image to obtain sign image sets respectively corresponding to the plurality of road signs.
After the plurality of signage images and the signage type and the position information of the road signage in each signage image are acquired, the server may classify the plurality of signage images according to the signage type and the position information of the road signage in each signage image, so as to obtain signage image sets corresponding to the plurality of road signage respectively, wherein each signage image set corresponds to one signage type and one position information.
In this embodiment, the server may classify the plurality of placard image sets according to the position information of the road placard in each of the plurality of placard images to obtain a plurality of classification sets. Since the signage types of the road signage images included in each classification set may be different, the server may classify the road signage images included in each classification set again according to the signage types of the road signage, so as to obtain signage image sets corresponding to the road signage.
For example, since each of the plurality of placard images corresponds to one location information, the server may mark the corresponding location information of each of the plurality of placard images in the map. The server may then determine a plurality of central location points with a greater concentration of locations based on the plurality of locations marked in the map. And determining a circular area by taking each central position point as a center and taking a preset numerical value as a radius. And then, classifying the label images corresponding to the positions of the labels in each circular area into a classification set, so as to obtain a plurality of classification sets. At this time, the position information of the center position point corresponding to each classification set may be used as the position information of the corresponding classification set.
After the plurality of classification sets are obtained, because the sign types of the road signs in the plurality of sign images in each classification set may be different, for any classification set, for example, the first classification set, the server takes the plurality of sign images with the same sign type of the road signs included in the first classification set as the sign images of the same road sign, thereby obtaining one or more sign image sets of the road signs, and at this time, the sign type corresponding to each sign image set is the sign type of the road signs included in the sign images. In addition, the server may further use the position information of the first classification set, that is, the position information of the center position point corresponding to the first classification set as the position information corresponding to the sign image sets of the road signs.
By the method, the server classifies the sign images in each classification set again according to the sign types, so that a plurality of sign image sets of road signs are obtained.
Optionally, in some possible implementations, after obtaining the set of road sign images, the server may further manually check the sign images in the set of road sign images to avoid mixing of sign images other than the road sign in the set of road sign images.
For example, for any road sign image set, the server may provide a display interface, display the sign images included in the sign image set on the display interface, and delete the sign images other than the road sign from the sign images displayed on the display interface.
Optionally, the server may also classify the plurality of signage images according to a signage type of a road sign in each signage image in the plurality of signage images to obtain a plurality of classification sets, and then classify the plurality of signage images in the corresponding classification set again according to the position information of the road sign in the plurality of signage images included in each classification set, so as to obtain a signage image set corresponding to the plurality of road signs.
For example, the server may compare the sign types of the road signs included in each of the plurality of sign images, and then divide the plurality of sign images having the same sign type into one class, thereby obtaining a plurality of classification sets corresponding to the plurality of sign types one to one. Since each classification set may include road sign images of the same sign type at different positions, after a plurality of classification sets are obtained, the server reclassifies a plurality of sign images in the corresponding classification set according to the position information of the road signs included in the plurality of sign images in each classification set, and uses each image set obtained by reclassification as a sign image set of the road sign, where the implementation of reclassification according to the position information refers to the above-described implementation of classifying signs according to the position information of the signs, which is not described in detail herein. At this time, the position information corresponding to the placard image set of each road placard obtained by classification is the position information of the corresponding center position point in the classification.
Step 203: a similarity reference value for the set of placard images for each road sign is determined.
After obtaining the road sign image sets corresponding to the road signs, the server may determine a similarity reference value for each road sign image set of the road signs. Wherein, the similarity reference value can represent the similarity degree between each label image included in the label image set.
In a possible implementation manner, the server may obtain a similarity between each signage image in each signage image set and each signage image other than the signage image included in the signage image set, and further determine a similarity reference value of each road signage image set according to the similarity between each signage image and each signage image in each signage image other than the signage image.
For example, a road sign image set of any one of the plurality of road sign image sets is taken as an example, and is referred to as a third sign image set for convenience of description. Taking any one of the signage images included in the third signage image set as an example, for convenience of description, it is referred to as a signage image a, and the server compares the signage image a with each of the signage images included in the third signage image set except the signage image a, and further determines a similarity between the signage image a and each of the signage images included in the third signage image set except the signage image a.
For example, the similarity between the placard image A and the other placard images is P1、P2、P3……Pn-1. n is the number of signage images included in the third signage image set, the server may use the following formula (b)1) To calculate the similarity average D corresponding to the label image A1。
By the method, the server can obtain the similarity average value corresponding to each label image in the third label image set, and the similarity average values can be respectively marked as D1、D2、D3、……Dn. The similarity average value corresponding to the signage image is the similarity average value of the signage image and other signage images in the signage image set for any signage image in the signage image set, so the similarity average value corresponding to the signage image can indicate the similarity degree of the signage image and other signage images in the signage image set.
After calculating the similarity average value corresponding to each signage image in the third signage image set, the server determines the similarity average value corresponding to the third signage image set according to the similarity average value corresponding to each signage image in the third signage image set by the following formula (2).
D1~DnThe similarity average value corresponding to each placard image included in the third placard image set is W, and W is the similarity average value corresponding to the third placard image set. The average similarity value corresponding to the third signage image set is an average similarity value of the average similarity values of the signage images in the third signage image set, so the average similarity value corresponding to the third signage image set can reflect the average similarity degree of the signage images in the signage image set.
After calculating the similarity average value corresponding to the third placard image set, the server compares the similarity average value corresponding to each placard image in the third placard image set with the corresponding phase of the third placard image setThe similarity averages are compared. That is, the average value D of the similarity degree corresponding to each of the placard images included in the third placard image set1、D2、D3、……DnAnd comparing the similarity average value W corresponding to the third sign, and deleting the sign images of which the corresponding similarity average value is smaller than the similarity average value corresponding to the third sign image set. And then determining a similarity reference value of the third signage image set according to the similarity average value and the image quality score corresponding to each remaining signage image.
For example, assume that the average similarity values corresponding to the respective signage images included in the third signage image set are D1、D2、D3、……D10Wherein D is2And D3D is deleted if the similarity is less than the similarity average W corresponding to the third label image set2、D3The similarity reference value C of the third placard image set is then calculated by the following formula (3) according to the similarity average value corresponding to the remaining 8 placard images.
Wherein C is the calculated similarity reference value of the third placard image set.
It should be noted that, if the average value of the similarity corresponding to a certain signage image in the signage image set is smaller than the average value of the similarity of the signage image set, it indicates that the degree of similarity between the signage image and other signage images is lower than the average degree of similarity of the signage images in the signage image set, that is, the degree of similarity between the signage image and other signage images is lower, in this case, the average value of the similarity average values of the remaining signage images is calculated after the signage images are removed, and the average value is used as a similarity reference value, so that the degree of similarity of the signage images in the signage image set can be more accurately reflected.
Alternatively, in another possible implementation manner, the server may obtain an image quality score of each placard image in each placard image set of the plurality of placard image sets, and determine the similarity reference value of each placard image set according to the image quality score of each placard image in each placard image set, and the similarity between each placard image and each placard image in other placard images except for the placard image set.
Illustratively, still taking the third placard image set as an example, the server first obtains an image quality score for each placard image in the third placard image set; determining a similarity weighted average corresponding to each signage image according to the image quality score of each signage image in the third signage image set and the similarity between the corresponding signage image and each signage image in other signage images except the third signage image set; determining a similarity weighted average value corresponding to the third signage image set according to the similarity weighted average value and the image quality score corresponding to each signage image in the third signage image set; and deleting the signage images of which the similarity weighted average value corresponding to the third signage image set is smaller than the similarity weighted average value corresponding to the third signage image set, and determining the similarity reference value of the third signage image set according to the similarity weighted average value corresponding to each remaining signage image and the image quality score.
It can be known from the foregoing description that, when the image capturing device captures the signage image of the road signage, the image capturing device is affected by factors such as the shooting angle of view, the field of view, and the illumination intensity, and thus the image quality of the collected signage images of the road signage is different. Based on the image quality scores, the server can obtain the image quality scores of the label images, and further combine the image quality scores of the label images in the label image set and the similarity among the label images to determine the similarity reference value, so that the obtained similarity reference value considers the influence of the image quality on the image similarity, and can better reflect the similarity of the images in the label image set.
The server may present each signage image in the third set of signage images in a display interface, and receive an image quality score for each signage image presented in the display interface submitted by a user. It should be noted that the image quality score of the road sign image by the user may be in the interval range of (0,100), wherein the higher the image quality score is, the better and clearer the road sign in the sign image is.
After the image quality score of each placard image in the third placard image set is obtained, the server may determine a weighted average value of the similarity corresponding to each placard image in the third placard image set according to the image quality score of each placard image in the third placard image set and the similarity between the corresponding placard image and each of the placard images in other placard images except the server itself, where the weighted average value of the similarity is used to indicate the degree of similarity between the corresponding placard image and the other placard images.
Illustratively, still taking signage images A in the third set of signage images as an example, suppose that the image quality score of each signage image in the third set of signage images is S1、S2、S3……SnAs can be seen from the above description, the similarity between the placard image a and each of the placard images other than the placard image a included in the third placard image set is P1、P2、P3……Pn-1. The server may calculate a similarity weighted average D corresponding to the placard image a by the following equation (4)1。
By the method, the server can obtain the similarity weighted average value corresponding to each label image in the third label image set, and the similarity weighted average values can be respectively marked as D1、D2、D3、……Dn。
It should be noted that, as can be seen from the above formula (4), the weighted average value of the similarity of any one of the signage images a in the third signage image set is obtained by performing weighted average according to the similarity of the signage image a and each of the other signage images, wherein, during weighting, the weight of the similarity of the signage image a and each of the other signage images is the image quality score of the other signage images, so that the influence of the similarity of the signage image with higher image quality and the signage image a on the weighted average value of the similarity of the signage image a is larger, and the calculated weighted average value of the similarity of the signage image a can reflect the similarity of the signage image a and each of the other signage images in the signage image set more accurately.
After calculating the similarity weighted average value corresponding to each placard image in the third placard image set, the server determines the similarity weighted average value corresponding to the third placard image set according to the similarity weighted average value corresponding to each placard image in the third placard image set and the image quality score through the following formula (5), wherein the similarity weighted average value corresponding to the third placard image set is used for indicating the similarity degree between the placard images in the third placard image set.
Wherein D is1~DnA weighted average of the similarity corresponding to each placard image contained in the third placard image set, S1~SnAnd scoring the image quality of each placard image in the third placard image set, wherein W is the weighted average of the similarity corresponding to the third placard image set.
It can be seen from the above formula (5) that the weighted average of the similarity corresponding to the third signage image set is obtained by performing weighted average again according to the weighted average of the similarity of each signage image in the third signage image set, wherein the weight corresponding to the weighted average of the similarity of each signage image is the image quality score of itself when performing weighting, so that the influence of the corresponding weighted average of the similarity on the weighted average of the similarity corresponding to the signage image set is larger for signage images with higher image quality, and thus the finally calculated weighted average of the similarity of the third signage image set can reflect the average similarity of each signage image more accurately.
After the similarity weighted average value corresponding to the third placard image set is calculated, the server compares the similarity weighted average value corresponding to each placard image in the third placard image set with the similarity weighted average value corresponding to the third placard image set, and deletes the placard images of which the similarity weighted average values are smaller than the similarity weighted average value corresponding to the third placard image set. And then determining the similarity reference value of the third signage image set according to the similarity weighted average value and the image quality score corresponding to each remaining signage image.
Similarly, assume that D is the similarity weighted average corresponding to each of the placard images included in the third placard image set2And D3D is deleted if the similarity is less than the similarity weighted average W corresponding to the third signage image set2、D3The similarity reference value C of the third placard image set is then calculated by the following formula (6) according to the similarity weighted average and the image quality score corresponding to the remaining 8 placard images.
Wherein C is the calculated similarity reference value of the third placard image set.
By the method, the server can calculate the similarity reference value of the sign image set of each road sign.
In summary, in this implementation manner, in the process of calculating the similarity reference value of the placard image set, the image quality score of each placard image is used as a weight, wherein the higher the image quality of the placard image, the greater the influence of the size of the similarity reference value of the placard image set, so that the calculated similarity reference value can more accurately reflect the similarity between the placard images in the placard image set, and has more referential property.
As can be seen from the foregoing description, the sign image set of each of the plurality of road signs corresponds to the sign type and the location information of the road sign, and therefore, after the similarity reference value of the sign image set of each road sign is calculated, the server may further store the sign image set of each road sign, the corresponding similarity reference value, and the sign type and the location information of the corresponding road sign, so as to obtain a road sign base, and provide a reference standard for performing subsequent damage identification on the road sign.
In the embodiment of the application, a plurality of sign images are collected in advance through image collection equipment, the sign type and the position information of the road sign in each sign image are obtained, then the plurality of sign images are classified according to the sign type and the position information of the road sign in each sign image, and a sign image set corresponding to the plurality of road signs is obtained. Since the signboard image set of each road signboard includes different images collected for the road signboard actually installed, the similarity reference value determined based on the similarity of the signboard images in the signboard image set is equivalent to comprehensively considering the presentation effect of the road signboard under different shooting conditions. The road sign board base is established by the plurality of sign image sets and each sign image set, and more accurate reference standards can be provided for the follow-up judgment of whether the to-be-detected sign is damaged, so that the accuracy of sign damage detection can be improved, and data support is provided for timely replacement of the road sign.
In addition, in the embodiment of the application, when the similarity reference value corresponding to the signage image set is determined, the similarity may be weighted and averaged based on the image quality score of the signage images in the image set, and then the signage images in the signage image set are screened according to the weighted average of the similarity, so that the better image quality and the higher similarity of the signage images in the signage image set are ensured, and a better reference standard is provided for subsequently judging whether the signage to be detected is damaged.
Next, a road sign breakage detection method provided in an embodiment of the present application will be described.
Fig. 3 is a method for detecting breakage of a road sign according to an embodiment of the present application. As shown in fig. 3, the method comprises the steps of:
step 301: and acquiring the label type and position information of the label to be detected contained in the first label image.
In the embodiment of the application, the image acquisition equipment acquires the image of the label to be detected to obtain the first label image. When the image is collected, the image collecting device can obtain the position information through the GPS device, and the position information is used as the position information of the to-be-detected label. And then, the image acquisition equipment identifies the label type of the label to be detected in the first label image through the AI model. And sending the first label image, the position information of the label to be detected and the label type of the label to be detected, which is obtained by identification, to a server. Accordingly, the server receives the first placard image and the placard type and location information of the placard to be detected included in the first placard image.
Optionally, in some possible implementations, after acquiring the first sign image and acquiring the position information of the sign to be detected, the image acquisition device may also directly send the first sign image and the position information to the server without identifying the sign type of the sign to be detected in the first sign image. And after receiving the first label image and the corresponding position information, the server identifies the type of the label to be detected in the first label image through the AI model.
Step 302: a first placard image set is obtained from a plurality of stored placard image sets according to placard type and position information of a placard to be detected, each placard image set including a plurality of placard images of a road placard collected in advance.
After the type and the position information of the label to be detected are obtained, the server can obtain candidate label image sets from the stored label image sets, and the distance between the position information corresponding to each candidate label image set and the position information of the label to be detected is within a reference threshold range; then, a second label image set is obtained from the obtained candidate label image set, and the label type corresponding to each second label image set is the same as the label type of the label to be detected; then, a first set of placard images is selected from the second set of placard images. The road sign image comprises a plurality of sign image sets, a plurality of sign image sets and a plurality of road sign images, wherein each sign image set in the plurality of sign image sets corresponds to one sign type and position information, and the position information corresponding to each sign image set is used for indicating the position of a road sign in the plurality of sign images included in the corresponding sign image set.
As can be seen from the foregoing description, the road sign base library stores a plurality of sign image sets, and each sign image set in the plurality of sign image sets corresponds to one sign type and position information, and based on this, the server may calculate a distance between the sign to be detected and the road sign included in each sign image set according to the position information of the sign to be detected and the position information corresponding to each sign image set. And then, taking the signage image set corresponding to the distance within the reference threshold range obtained by calculation as a candidate signage image set.
Alternatively, the server may also determine, according to the position information of the sign to be detected and the position information corresponding to each sign image set in the plurality of sign image sets, the position of the sign to be detected in the map and the position of the road sign corresponding to each sign image set in the plurality of sign image sets, then determine a circular area with the position of the sign to be detected in the map as the center and the upper limit value of the reference threshold range as the radius, and use, as the candidate sign image set, the sign image set including the road sign whose position is within the circular area.
For example, if the upper limit value of the reference threshold range is 10m and the lower limit value is 0, a circular area is determined by taking the position of the sign to be detected in the map as the center and taking 10m as the radius, and the sign image set corresponding to the road sign located in the circular area is taken as the candidate sign image set, where 10m is only an exemplary value and does not limit the embodiment of the present application.
It should be noted that there may be one or more candidate signage image sets obtained according to the position information, and when there is one candidate signage image set, the server may determine whether the signage type corresponding to the candidate signage image set is the same as the signage type of the sign to be detected, and if so, take the candidate signage image set as the first signage image set.
When the candidate sign image sets are multiple, the server may compare the sign type of the sign to be detected with the sign types corresponding to the candidate sign image sets, and further acquire, from the multiple candidate sign image sets, a sign image set whose corresponding sign type is the same as the sign type of the sign to be detected as a second sign image set. At this time, there may be one or more second placard image sets acquired.
Optionally, the server may also obtain, according to the type of the sign to be detected, a sign image set, in which the corresponding sign type is the same as the sign type of the sign to be detected, from the plurality of sign image sets stored in the road sign base as a candidate sign image set. And then according to the position information corresponding to the candidate sign image set and the position information of the sign to be detected, acquiring a sign image set as a second sign image set, wherein the distance between the corresponding position information and the position information of the sign to be detected is within a reference threshold range from the candidate sign image set. At this time, there may be one or more second placard image sets acquired.
In the process of acquiring the second signage image set from the acquired candidate signage image set, the server may refer to the implementation manner described above for determining the candidate signage image set according to the signage image set and the position information of the to-be-detected signage, which is not described herein again in this embodiment of the present application.
After obtaining the second signage image set, the server may select one signage image set from the second signage image set that matches the first signage image most as the first signage image set.
As can be seen from the above description, there may be one or more second signage image sets obtained by the server that match the signage to be detected. Based on this, the implementation of selecting one of the signage image sets that matches the first signage image most closely from the second signage image set as the first signage image set will be divided into the following two cases.
In the first case: when the number of the acquired second signage image sets is plural, the server may select one signage image set having the highest matching degree with the first signage image from the plural second signage image sets as the first signage image set.
In an embodiment of the present application, the server may determine a similarity of the first signage image and each signage image in each second signage image set; then determining the similarity average value of the first label image and the label images in the corresponding second label image set according to the similarity of the first label image and each label image in each second label image set; then, the second placard image set having the largest corresponding similarity average is selected from the plurality of second placard image sets as the first placard image set.
For example, taking any one of the plurality of second signage image sets as an example, if the second signage image set includes 10 signage images, the server may compare the similarity between the first signage image and each of the 10 signage images included in the second signage image set, so as to obtain the similarity between the first signage image and each of the 10 signage images. The server may note the determined similarity of the first placard image to the 10 placard images as p1、p2、p3……p10. Similarity of the first placard image to the respective placard images in each of the plurality of second placard image sets may be determined according to the same method.
After calculating the similarity of the first signage image and each signage image in each signage image set in the plurality of second signage image sets, for any second signage image set, the server calculates the average value of the similarity of the first signage image and each signage image in the second signage image set, so as to obtain the average value of the similarity of the first signage image and each signage image in the corresponding second signage image set.
Continuing with the foregoing example, when the similarity between the 10 placard images included in the first placard image set and the second placard image set is p1、p2、p3……p10The server may then calculate the first signage image and the second signage imageSimilarity p of 10 placard images included in two placard image set1、p2、p3……p10To obtain an average value of the similarity of the first signage image and the signage images in the corresponding second signage image set, which may be denoted as E. The calculation formula of the similarity average value may refer to the following formula (7).
Where n is the number of signage images in the signage image set.
According to the same calculation method, the average value of the similarity of the first placard image and the placard image in each of the plurality of second placard image sets can be obtained. As can be seen from the above description, the average value of the similarity between the first signage image and the signage image in each second signage image set is actually the average value of the similarity between the first signage image and each signage image in the corresponding signage image set, so that the average value of the similarity actually represents the similarity between the first signage image and the signage image in the signage image set, in other words, the similarity between the signage to be detected in the first signage image and the road signage in the signage image set.
Since the average value of the similarity between the first signage image and the signage images in the signage image set actually represents the similarity between the first signage image and the signage images in the signage image set, when the average value of the similarity is higher, it indicates that the similarity between the first signage image and the signage images in the signage image set is higher, that is, the matching degree between the first signage image and the signage image set is higher. Based on this, after obtaining the similarity average value of the first signage image and the signage image in each of the plurality of second signage image sets, the maximum similarity average value may be determined from the similarity average values corresponding to the plurality of second signage image sets, and the second signage image set corresponding to the maximum similarity average value may be used as the first signage image set having the highest matching degree with the first signage image.
In the second case: when the number of the acquired second signage image sets is 1, the server may regard the second signage image set as the first signage image set. In this case, the server may further determine a similarity between the first placard image and each of the placard images in the first placard image set, and determine an average value of the similarities between the first placard image and the placard images in the first placard image set based on the similarities between the first placard image and each of the placard images in the first placard image set.
The method for calculating the average value of the similarity between the first signage image and the signage images in the first signage image set refers to the method for calculating the average value of the similarity described above, and details of the embodiment of the present application are omitted.
Step 303: and carrying out damage detection on the label to be detected according to the similarity between the first label image and the label image in the first label image set.
As can be seen from the foregoing description, each signage image set corresponds to a similarity reference value, and the similarity reference value can represent the similarity between the signage images included in the corresponding signage image set. Based on this, after the first signage image set is determined, the server may obtain the similarity reference value corresponding to the first signage image set, and perform damage detection on the signage to be detected included in the first signage image according to each similarity and similarity reference value of the first signage image and each signage image in the first signage image set.
Illustratively, the server compares the average of the similarity of the first placard image and the first placard image set determined in step 302 above with the similarity reference value corresponding to the first placard image set. And if the similarity average value is greater than or equal to the similarity reference value, determining that the sign to be detected contained in the first sign image is not broken. And if the similarity average value is smaller than the similarity reference value, determining that the sign to be detected contained in the first sign image is damaged.
Optionally, in some possible implementations, if the average similarity value is smaller than the reference similarity value corresponding to the first signage image set, the server may further calculate a difference between the reference similarity value and the average similarity value, and if the difference is within a preset threshold range, determine that the signage to be detected included in the first signage image is damaged.
Optionally, after determining that the to-be-detected sign is damaged, the server may further determine the damage degree of the to-be-detected sign according to the similarity reference value corresponding to the first sign image set and the average value of the similarities between the first sign image and the sign images in the first sign image set.
For example, the server may determine an absolute value of a difference between an average value of the similarity of the first placard image and the placard images in the first placard image set and a similarity reference value corresponding to the first placard image set. And then, determining the damage degree of the label to be detected according to the reference difference interval in which the absolute value of the difference is positioned.
It should be noted that the server may store mapping relationships between different reference difference intervals and corresponding damage degrees. Based on the above, the server determines a reference difference interval in which the absolute value of the calculated difference is located, and takes the damage degree corresponding to the determined reference difference interval as the damage degree of the to-be-detected label. The damage degree of the label can be represented by a percentage or other parameters such as damage levels, which is not limited in the embodiment of the present application.
For example, the server stores mapping relationships between 3 different reference difference intervals and corresponding damage degrees, where the 3 different reference difference intervals may be (1, 3), (4, 7), and (8, 10), where the reference difference interval (1, 3) represents a slight damage, the reference difference interval (4, 7) represents a moderate damage, and the reference difference interval (8, 10) represents a severe damage. Based on the above, when the absolute value of the difference obtained by the calculation of the server is 2, the to-be-detected label is considered to be slightly damaged; when the absolute value of the difference calculated by the server is 5, the to-be-detected label is considered to be moderately damaged; and when the absolute value of the difference calculated by the server is 9, the to-be-detected label is considered to be severely damaged. It should be noted that the reference difference interval and the absolute value of the difference are exemplary values, and do not limit the embodiments of the present application.
In the embodiment of the application, a plurality of label image sets are stored in advance, and the label image set matched with the type and the position information of the label to be detected is acquired from the plurality of label image sets stored in advance according to the type and the position information of the label to be detected, so that the probability that the acquired label image set is the image set containing the label to be detected can be improved. Moreover, each of the plurality of signboard image sets comprises a plurality of different signboard images which are acquired in advance and are actually provided with the road signboard, so that whether the to-be-detected signboard is damaged or not is judged by taking the acquired signboard image set as a reference through the similarity between the image of the to-be-detected signboard and the acquired signboard image in the signboard image set, whether the to-be-detected signboard is damaged or not can be judged more accurately, and data support can be provided for timely replacement of the road signboard.
In addition, since the sign image set of each road sign includes different images collected for the road sign actually mounted, the similarity reference value determined based on the similarity of the sign images in the sign image set is equivalent to comprehensively considering the presentation effect of the road sign under different shooting conditions. On the basis, the similarity reference value is used as a reference standard for judging whether the to-be-detected label in the collected first label image is damaged or not, and the method is more accurate.
Finally, the embodiment of the application determines the damage degree of the sign to be detected by setting the mapping relation between the reference difference value interval and the corresponding damage degree, so that the damage degree can be quantized, and data support can be better provided for replacing the road sign.
Next, a road sign breakage detection device provided in an embodiment of the present application will be described.
Referring to fig. 4, the present application provides a road sign breakage detection apparatus 400, the apparatus 400 includes: a first obtaining module 401, a second obtaining module 402, and a detecting module 403.
A first obtaining module 401, configured to obtain a label type and position information of a to-be-detected label included in the first label image;
a second obtaining module 402, configured to obtain a first signage image set from the stored plurality of signage image sets according to a signage type and position information of the signage to be detected, where each signage image set includes a plurality of signage images of one road signage acquired in advance;
the detecting module 403 is configured to perform damage detection on the to-be-detected signage according to similarity between the first signage image and the signage image in the first signage image set.
Optionally, each signage image set of the plurality of signage image sets corresponds to one signage type and location information;
a second obtaining module 402, comprising:
the first acquisition submodule is used for acquiring candidate sign image sets from the plurality of sign image sets, and the distance between the position information corresponding to each candidate sign image set and the position information of the sign to be detected is within a reference threshold range;
the second acquisition sub-module is used for acquiring a second label image set from the acquired candidate label image set, and the label type corresponding to each second label image set is the same as the label type of the label to be detected;
a selection sub-module for selecting the first set of signage images from the second set of signage images.
Optionally, a selection submodule for:
determining similarity of the first signage image and each signage image in each second signage image set;
determining the similarity average value of the first label image and the label images in the corresponding second label image set according to the similarity of the first label image and each label image in each second label image set;
and selecting the second label image set with the maximum corresponding similarity average value from the plurality of second label image sets as the first label image set.
Optionally, the detection module 403 includes:
the third acquisition submodule is used for acquiring the similarity reference value of the first label image set;
and the detection submodule is used for carrying out damage detection on the label to be detected according to the similarity and the similarity reference value of the first label image and each label image in the first label image set.
Optionally, a detection submodule for:
determining the similarity average value of the first label image and the first label image set according to the similarity of the first label image and each label image in the first label image set;
and if the average value of the similarity of the first label image and the first label image set is smaller than the similarity reference value, determining that the label to be detected is damaged.
Optionally, the detection submodule is further configured to:
determining an absolute value of a difference between the average value of the similarity of the first placard image and the first placard image set and the reference value of the similarity;
and determining the damage degree of the label to be detected according to the reference difference interval in which the absolute value of the difference is positioned.
Optionally, the apparatus 400 further comprises:
a third obtaining module 404, configured to obtain a plurality of sign images collected in advance and sign types and position information of road signs in each sign image;
the classification module 405 is configured to classify the plurality of signage images according to signage types and position information of road signs in each signage image to obtain a plurality of signage image sets, where each signage image set corresponds to one signage type and position information;
a determining module 406 for determining a similarity reference value for each signage image set.
Optionally, the determining module 406 includes:
the fourth acquisition submodule is used for acquiring the image quality score of each sign image in the third sign image set, and the third sign image set is a sign image set of any road sign;
and the determining submodule is used for determining a similarity reference value of the third label image set according to the image quality score of each label image in the third label image set and the similarity between each label image and each label image in other label images except the third label image set.
Optionally, a determination submodule for:
determining a similarity weighted average corresponding to each signage image according to the image quality score of each signage image in the third signage image set and the similarity between the corresponding signage image and each signage image in other signage images except the third signage image set;
determining a similarity weighted average value corresponding to the third signage image set according to the similarity weighted average value and the image quality score corresponding to each signage image in the third signage image set;
and deleting the signage images of which the similarity weighted average value corresponding to the third signage image set is smaller than the similarity weighted average value corresponding to the third signage image set, and determining the similarity reference value of the third signage image set according to the similarity weighted average value corresponding to each remaining signage image and the image quality score.
In summary, in the embodiment of the present application, a plurality of signage image sets are stored in advance, and a first signage image set is obtained from the plurality of signage image sets stored in advance according to the type and the position information of the signage to be detected, so that the probability that the obtained signage image set is the image set including the signage to be detected can be improved. Moreover, each of the plurality of signboard image sets comprises a plurality of signboard images which are acquired in advance and are actually installed on the road signboard, so that whether the to-be-detected signboard is damaged or not is judged by taking the acquired signboard image set as a reference according to the similarity between the image of the to-be-detected signboard and the acquired signboard image in the signboard image set, whether the to-be-detected signboard is damaged or not can be judged more accurately, and data support can be provided for timely replacement of the road signboard.
It should be noted that, when the road sign breakage detection apparatus provided in the above embodiment detects a road sign, the division of the above function modules is merely exemplified, and in practical applications, the above functions may be distributed to different function modules according to needs, that is, the internal structure of the apparatus may be divided into different function modules to complete all or part of the above described functions. In addition, the road sign breakage detection device provided by the embodiment and the road sign breakage detection method embodiment belong to the same concept, and the specific implementation process is detailed in the method embodiment and is not described again.
Fig. 5 is a schematic diagram illustrating a server architecture in accordance with an example embodiment. The function of the road sign breakage detection in the above embodiment may be implemented by the server shown in fig. 5. The server may be a server in a cluster of background servers. Specifically, the method comprises the following steps:
the server 500 includes a Central Processing Unit (CPU) 501, a system Memory 504 including a Random Access Memory (RAM) 502 and a Read-Only Memory (ROM) 503, and a system bus 505 connecting the system Memory 504 and the CPU 501. The server 500 also includes a basic Input/Output system (I/O system) 506, which facilitates information transfer between devices within the computer, and a mass storage device 507, which stores an operating system 513, application programs 514, and other program modules 515.
The basic input/output system 506 comprises a display 508 for displaying information and an input device 509, such as a mouse, keyboard, etc., for user input of information. Wherein a display 508 and an input device 509 are connected to the central processing unit 501 through an input output controller 510 connected to the system bus 505. The basic input/output system 506 may also include an input/output controller 510 for receiving and processing input from a number of other devices, such as a keyboard, mouse, or electronic stylus. Similarly, input-output controller 510 also provides output to a display screen, a printer, or other type of output device.
The mass storage device 507 is connected to the central processing unit 501 through a mass storage controller (not shown) connected to the system bus 505. The mass storage device 507 and its associated computer-readable media provide non-volatile storage for the server 500. That is, the mass storage device 507 may include a computer-readable medium (not shown) such as a hard disk or a CD-ROM (Compact disk Read-Only Memory) drive.
Without loss of generality, computer readable media may comprise computer storage media and communication media. Computer storage media includes volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Computer storage media includes RAM, ROM, EPROM (Erasable Programmable Read-Only Memory), EEPROM (Electrically Erasable Programmable Read-Only Memory), flash Memory or other solid state Memory device, CD-ROM, DVD (Digital Versatile disk), or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices. Of course, those skilled in the art will appreciate that computer storage media is not limited to the foregoing. The system memory 504 and mass storage device 507 described above may be collectively referred to as memory.
According to various embodiments of the present application, server 500 may also operate as a remote computer connected to a network through a network, such as the Internet. That is, the server 500 may be connected to the network 512 through the network interface unit 511 connected to the system bus 505, or may be connected to other types of networks or remote computer systems (not shown) using the network interface unit 511.
The memory further includes one or more programs, and the one or more programs are stored in the memory and configured to be executed by the CPU. The one or more programs include instructions for performing the road sign breakage detection method provided by the embodiments of the present application.
Embodiments of the present application also provide a computer-readable storage medium, where instructions are executed by a processor of a server, so that the server can execute the road sign breakage detection method provided in the above embodiments. For example, the computer readable storage medium may be a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical data storage device, and the like. It is noted that the computer-readable storage medium referred to in the embodiments of the present application may be a non-volatile storage medium, in other words, a non-transitory storage medium.
It should be understood that all or part of the steps for implementing the above embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. The computer instructions may be stored in the computer-readable storage medium described above.
That is, in some embodiments, there is also provided a computer program product containing instructions which, when run on a computer, cause the computer to perform the road sign breakage detection method provided by the above-described embodiments.
The above description should not be taken as limiting the embodiments of the present application, and any modifications, equivalents, improvements, etc. made within the spirit and principle of the embodiments of the present application should be included in the scope of the embodiments of the present application.
Claims (13)
1. A method of detecting breakage of a pavement marker, the method comprising:
acquiring the label type and position information of a label to be detected contained in the first label image;
acquiring a first sign image set from a plurality of stored sign image sets according to the sign type and the position information of the sign to be detected, wherein each sign image set comprises a plurality of sign images of a road sign acquired in advance;
and carrying out damage detection on the to-be-detected label according to the similarity between the first label image and the label image in the first label image set.
2. The method of claim 1, wherein each signage image set of the plurality of signage image sets corresponds to a signage type and location information;
the acquiring a first signage image set from a plurality of stored signage image sets according to the signage type and the position information of the signage to be detected comprises:
acquiring candidate sign image sets from the plurality of sign image sets, wherein the distance between the position information corresponding to each candidate sign image set and the position information of the sign to be detected is within a reference threshold range;
acquiring a second label image set from the acquired candidate label image set, wherein the label type corresponding to each second label image set is the same as the label type of the label to be detected;
a first set of placard images is selected from the second set of placard images.
3. The method of claim 2, wherein selecting the first set of placard images from the second set of placard images comprises:
determining a similarity of the first signage image and each signage image in each second signage image set;
determining the similarity average value of the first label image and the label images in the corresponding second label image set according to the similarity of the first label image and each label image in each second label image set;
and selecting the second label image set with the maximum corresponding similarity average value from the plurality of second label image sets as the first label image set.
4. The method of claim 1, wherein the detecting breakage of the signage based on the similarity of the first signage image and the signage images in the first signage image set comprises:
acquiring a similarity reference value of the first label image set;
and carrying out damage detection on the to-be-detected label according to the similarity of the first label image and each label image in the first label image set and the similarity reference value.
5. The method of claim 4, wherein the detecting the breakage of the sign to be detected according to the similarity of the first sign image and the similarity reference value, comprises:
determining an average value of the similarity of the first placard image and the first placard image set according to the similarity of the first placard image and the placard image in the first placard image set;
and if the average value of the similarity of the first label image and the first label image set is smaller than the reference value of the similarity, determining that the label to be detected is damaged.
6. The method of claim 5, wherein the detecting breakage of the signage to be detected is performed according to the similarity of the first signage image and the similarity reference value, and further comprising:
determining an absolute value of a difference between the average of the similarity of the first placard image and the first placard image set and the similarity reference value;
and determining the damage degree of the to-be-detected label according to the reference difference interval in which the absolute value of the difference is positioned.
7. The method according to any one of claims 1-6, further comprising:
acquiring a plurality of sign images collected in advance and sign type and position information of a road sign in each sign image;
classifying the plurality of sign images according to the sign types and the position information of the road signs in each sign image to obtain a plurality of sign image sets, wherein each sign image set corresponds to one sign type and position information;
a similarity reference value is determined for each signage image set.
8. The method of claim 7, wherein determining the similarity reference value for each placard image set comprises:
acquiring an image quality score of each sign image in a third sign image set, wherein the third sign image set is a sign image set of any road sign;
and determining a similarity reference value of the third label image set according to the image quality score of each label image in the third label image set and the similarity between each label image and each label image in other label images except the third label image set.
9. The method of claim 8, wherein determining the similarity reference value for the third set of placard images based on the image quality score for each placard image in the third set of placard images, the similarity between each placard image and each of the other placard images except for the third placard image comprises:
determining a similarity weighted average corresponding to each signage image according to the image quality score of each signage image in the third signage image set and the similarity between the corresponding signage image and each signage image in other signage images except the third signage image set;
determining a similarity weighted average value corresponding to the third placard image set according to the similarity weighted average value and the image quality score corresponding to each placard image in the third placard image set;
and deleting the signage images of which the similarity weighted average value corresponding to the third signage image set is smaller than the similarity weighted average value corresponding to the third signage image set, and determining the similarity reference value of the third signage image set according to the similarity weighted average value corresponding to each remaining signage image and the image quality score.
10. The utility model provides a damaged detection device of road sign board which characterized in that, the device includes:
the first acquisition module is used for acquiring the label type and the position information of the label to be detected contained in the first label image;
the second acquisition module is used for acquiring a first sign image set from a plurality of stored sign image sets according to the sign type and the position information of the sign to be detected, wherein each sign image set comprises a plurality of pre-acquired sign images of one road sign;
and the detection module is used for carrying out damage detection on the to-be-detected label according to the similarity between the first label image and the label image in the first label image set.
11. The apparatus of claim 10, wherein each signage image set of the plurality of signage image sets corresponds to a signage type and location information;
the second obtaining module includes:
the first acquisition submodule is used for acquiring candidate sign image sets from the plurality of sign image sets, and the distance between the position information corresponding to each candidate sign image set and the position information of the sign to be detected is within a reference threshold range;
the second acquisition sub-module is used for acquiring a second label image set from the acquired candidate label image set, and the label type corresponding to each second label image set is the same as the label type of the label to be detected;
a selection sub-module for selecting a first set of signage images from the second set of signage images;
wherein the selection submodule is mainly used for:
determining a similarity of the first signage image and each signage image in each second signage image set;
determining the similarity average value of the first label image and the label images in the corresponding second label image set according to the similarity of the first label image and each label image in each second label image set;
selecting a second placard image set with the largest corresponding similarity average value from the plurality of second placard image sets as the first placard image set;
wherein, the detection module includes:
a third obtaining submodule, configured to obtain a similarity reference value of the first signage image set;
the detection submodule is used for carrying out damage detection on the label to be detected according to the similarity of the first label image and each label image in the first label image set and the similarity reference value;
wherein, the detection submodule is mainly used for:
determining an average value of the similarity of the first placard image and the first placard image set according to the similarity of the first placard image and the placard image in the first placard image set;
if the average value of the similarity of the first label image and the first label image set is smaller than the reference value of the similarity, determining that the label to be detected is damaged;
wherein the detection submodule is further configured to:
determining an absolute value of a difference between the average of the similarity of the first placard image and the first placard image set and the similarity reference value;
determining the damage degree of the to-be-detected label according to the reference difference interval in which the absolute value of the difference is located;
wherein the apparatus further comprises:
the third acquisition module is used for acquiring a plurality of sign images acquired in advance and sign types and position information of the road signs in each sign image;
the classification module is used for classifying the plurality of sign images according to the sign types and the position information of the road signs in each sign image to obtain a plurality of sign image sets, and each sign image set corresponds to one sign type and position information;
the determining module is used for determining a similarity reference value of each label image set;
wherein the determining module comprises:
a fourth obtaining submodule, configured to obtain an image quality score of each signage image in a third signage image set, where the third signage image set is a signage image set of any road signage;
the determining submodule is used for determining a similarity reference value of the third label image set according to the image quality score of each label image in the third label image set and the similarity between each label image and each label image in other label images except the third label image set;
wherein the determining submodule is mainly used for:
determining a similarity weighted average corresponding to each signage image according to the image quality score of each signage image in the third signage image set and the similarity between the corresponding signage image and each signage image in other signage images except the third signage image set;
determining a similarity weighted average value corresponding to the third placard image set according to the similarity weighted average value and the image quality score corresponding to each placard image in the third placard image set;
and deleting the signage images of which the similarity weighted average value corresponding to the third signage image set is smaller than the similarity weighted average value corresponding to the third signage image set, and determining the similarity reference value of the third signage image set according to the similarity weighted average value corresponding to each remaining signage image and the image quality score.
12. A computer device, characterized in that the computer device comprises a processor and a memory, the memory being used for storing a computer program, the processor being used for executing the computer program stored on the memory to implement the road sign breakage detection method according to any one of claims 1 to 9.
13. A computer-readable storage medium, in which a computer program is stored, the computer program, when executed by a computer, implementing the road sign breakage detection method according to any one of claims 1 to 9.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111530371.7A CN114187568A (en) | 2021-12-14 | 2021-12-14 | Road sign breakage detection method, apparatus and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111530371.7A CN114187568A (en) | 2021-12-14 | 2021-12-14 | Road sign breakage detection method, apparatus and storage medium |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114187568A true CN114187568A (en) | 2022-03-15 |
Family
ID=80543826
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111530371.7A Pending CN114187568A (en) | 2021-12-14 | 2021-12-14 | Road sign breakage detection method, apparatus and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114187568A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116567405A (en) * | 2023-07-05 | 2023-08-08 | 北京华盛恒辉科技有限公司 | Picture change identification and shooting method, equipment and medium |
-
2021
- 2021-12-14 CN CN202111530371.7A patent/CN114187568A/en active Pending
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN116567405A (en) * | 2023-07-05 | 2023-08-08 | 北京华盛恒辉科技有限公司 | Picture change identification and shooting method, equipment and medium |
CN116567405B (en) * | 2023-07-05 | 2023-09-12 | 北京华盛恒辉科技有限公司 | Picture change identification and shooting method, equipment and medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110148196B (en) | Image processing method and device and related equipment | |
CN110705405B (en) | Target labeling method and device | |
CN110146097B (en) | Method and system for generating automatic driving navigation map, vehicle-mounted terminal and server | |
CN110569856B (en) | Sample labeling method and device, and damage category identification method and device | |
CN112541372B (en) | Difficult sample screening method and device | |
CN105321350A (en) | Method and device for detection of fake plate vehicles | |
CN111898581A (en) | Animal detection method, device, electronic equipment and readable storage medium | |
CN111435446A (en) | License plate identification method and device based on L eNet | |
CN113240009A (en) | Point cloud data labeling method and device, storage medium and electronic equipment | |
RU2019136604A (en) | PLATFORM FOR MANAGING AND VERIFICATION OF VIDEO CONTENT, PHOTOS OR SIMILAR CONTENT GENERATED BY DIFFERENT DEVICES | |
CN112396044A (en) | Method for training lane line attribute information detection model and detecting lane line attribute information | |
CN115810134A (en) | Image acquisition quality inspection method, system and device for preventing car insurance from cheating | |
CN114187568A (en) | Road sign breakage detection method, apparatus and storage medium | |
CN112528903B (en) | Face image acquisition method and device, electronic equipment and medium | |
CN113505653A (en) | Object detection method, device, equipment, medium and program product | |
CN114821658B (en) | Face recognition method, operation control device, electronic equipment and storage medium | |
CN116580230A (en) | Target detection method and training method of classification model | |
CN115331151A (en) | Video speed measuring method and device, electronic equipment and storage medium | |
CN110555344A (en) | Lane line recognition method, lane line recognition device, electronic device, and storage medium | |
CN114241195A (en) | Target identification method and device, electronic equipment and storage medium | |
CN114373081A (en) | Image processing method and device, electronic device and storage medium | |
Kaimkhani et al. | UAV with Vision to Recognise Vehicle Number Plates | |
CN112069971A (en) | Video-based highway sign line identification method and identification system | |
CN111639640A (en) | License plate recognition method, device and equipment based on artificial intelligence | |
CN112232295A (en) | Method and device for confirming newly-added target ship and electronic equipment |
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 |