KR20220095675A - System and method of recogning airport baggage vision sorter - Google Patents

Abstract

Disclosed are an airport baggage vision recognition sorter system and method for improving baggage sorting accuracy and reaching an accurate drop-off point. According to an embodiment of the present invention, an airport baggage vision recognition sorter system comprises: a primary photographing unit which obtains feature data and barcode information from a first baggage image of the baggage photographed by a vision camera, and registers the feature data and barcode information in a server; and a secondary photographing unit which obtains feature data from a second baggage image of unclassified baggage photographed by the vision camera if the baggage is determined as unclassified baggage because the barcode information is not recognized in a barcode recognition system, and classifies and processes the unclassified baggage by extracting the registered barcode information corresponding to the obtained feature data from the server.

Description

Airport Baggage Vision Recognition Sorter System and Method

The present invention relates to an airport baggage vision recognition sorter system and method, which can classify and process baggage with higher accuracy by performing barcode recognition and vision recognition in parallel.

The sorter system, which is a baggage sorting system currently used in airports, recognizes barcode information attached to baggage and arrives the baggage to a designated place.

Although the sorter system currently has an accuracy of 95%, 5% of misclassified baggage that is not classified because it cannot read barcode information is still being reclassified in the form of cumbersome manual work.

That is, the current sorter system classifies luggage by reading only barcode information, so there has been a limit in that human resources must be input for the classification of luggage that does not read barcode information.

Accordingly, there is an urgent need for a new method of classifying luggage with higher accuracy by applying a vision recognition model in luggage classification.

In an embodiment of the present invention, an artificial intelligence camera is installed at the check-in desk to collect barcode information and the external shape of the luggage, and secondary pictures are taken for unrecognized luggage when passing through the barcode reader, thereby comparing the external shape. A task to solve is to provide an airport baggage vision recognition sorter system and method that improves the accuracy of baggage sorting by using the barcode information collected during the first shooting through

In addition, an embodiment of the present invention uploads feature data extracted through artificial intelligence from a camera and a PC installed at the check-in desk to a server, and then takes a second shot of unclassified baggage and uses artificial intelligence to transmit the feature data. After extraction, the purpose of distributing and reducing the overall load of the system is to download the feature data uploaded to the server and compare the feature data only for unclassified baggage.

According to an embodiment of the present invention, the airport baggage vision recognition sorter system acquires feature data and barcode information from a first baggage image of baggage photographed by a vision camera, and matches the feature data with the barcode information to a server the first recording unit to register; And if the baggage is determined as unclassified baggage because barcode information is not recognized by the barcode recognition system, feature data is obtained from the second baggage image of the unclassified baggage photographed by the vision camera, and registered corresponding to the acquired feature data By extracting barcode information from the server, it may include a secondary photographing unit that sorts and processes the unclassified baggage.

In addition, according to an embodiment of the present invention, the airport baggage vision recognition sorter method includes: obtaining, in a primary photographing unit, characteristic data and barcode information from a first baggage image of the baggage photographed by the vision camera; registering, in the primary photographing unit, the characteristic data and the barcode information in correspondence with the server; If the baggage is determined as unclassified baggage because barcode information is not recognized in the barcode recognition method, acquiring characteristic data from a second baggage image of the unclassified baggage photographed by a vision camera in a secondary photographing unit; and in the secondary photographing unit, by extracting barcode information registered in response to the acquired characteristic data from the server, classifying the unclassified baggage.

According to an embodiment of the present invention, an artificial intelligence camera is installed at the check-in desk to collect barcode information and the external shape of the luggage, and the external shape is taken by taking a secondary picture of the unrecognized luggage when passing through the barcode reader. Through the comparison, it is possible to provide an airport baggage vision recognition sorter system and method that improves the accuracy of the baggage sorting operation by using the barcode information collected during the first filming and reaches the correct drop-off point.

In addition, according to an embodiment of the present invention, feature data extracted through artificial intelligence from a camera and a PC installed at the check-in desk are uploaded to the server, and unclassified baggage is captured secondly to be characterized by artificial intelligence. After extracting the data, it is possible to distribute and reduce the overall load of the system by downloading the feature data uploaded to the server and comparing the feature data only for unclassified baggage.

1 is a block diagram showing the configuration of an airport baggage vision recognition sorter system according to an embodiment of the present invention.
2 is a schematic diagram of a system according to an embodiment of the present invention;
3 is a diagram for illustrating a primary photographing unit installed at a check-in desk (counter).
4 is a diagram for explaining a change in the overall success rate according to the secondary vision recognition recognition rate.
5 is a flowchart illustrating an airport baggage vision recognition sorter method according to an embodiment of the present invention.

Hereinafter, embodiments will be described in detail with reference to the accompanying drawings. However, since various changes may be made to the embodiments, the scope of the patent application is not limited or limited by these embodiments. It should be understood that all modifications, equivalents and substitutes for the embodiments are included in the scope of the rights.

The terms used in the examples are used for the purpose of description only, and should not be construed as limiting. The singular expression includes the plural expression unless the context clearly dictates otherwise. In this specification, terms such as "comprise" or "have" are intended to designate that a feature, number, step, operation, component, part, or a combination thereof described in the specification exists, but one or more other features It should be understood that this does not preclude the existence or addition of numbers, steps, operations, components, parts, or combinations thereof.

Unless otherwise defined, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which the embodiment belongs. Terms such as those defined in commonly used dictionaries should be interpreted as having a meaning consistent with the meaning in the context of the related art, and should not be interpreted in an ideal or excessively formal meaning unless explicitly defined in the present application. does not

In addition, in the description with reference to the accompanying drawings, the same components are given the same reference numerals regardless of the reference numerals, and the overlapping description thereof will be omitted. In describing the embodiment, if it is determined that a detailed description of a related known technology may unnecessarily obscure the gist of the embodiment, the detailed description thereof will be omitted.

1 is a block diagram showing the configuration of an airport baggage vision recognition sorter system according to an embodiment of the present invention.

Referring to FIG. 1 , the airport baggage vision recognition sorter system 100 according to an embodiment of the present invention may include a primary photographing unit 110 and a secondary photographing unit 120 .

First, the primary photographing unit 110 acquires feature data and barcode information from the first baggage image of the baggage photographed by the vision camera. That is, the primary photographing unit 110 may perform a vision recognition for the first luggage image, and may serve to recognize characteristic data regarding the appearance of the luggage and barcode information tagged on the luggage.

Here, the vision camera may be a kind of vision machine that captures image data and generates an uncompressed image of the first luggage, so that it is possible to detect the shape of the luggage and read the information code.

In addition, the characteristic data is data regarding the external characteristics of the luggage, and may include, for example, external characteristics such as width/length/height length, shape, color, texture, and presence or absence of a handle/wheel of the luggage.

In addition, the barcode information may refer to combination information such as a bar-shaped line and a number that are tagged on the surface of a passenger's baggage upon check-in.

In addition, vision recognition may be a technology of extracting barcode information tagged on luggage as meaningful information from the first luggage image using AI time. Such vision recognition is being used in various industries such as autonomous vehicles, industrial robots, medical imaging technology, smart home, intelligent CCTV, access control, facial recognition, and factory automation as deep learning and deep view technologies are combined.

In other words, the primary photographing unit 110 may obtain the characteristic data of the appearance of the luggage and barcode information given at the time of check-in from the first luggage image obtained by photographing the luggage.

In addition, the primary photographing unit 110 registers the characteristic data and the barcode information in correspondence with the server. That is, the primary photographing unit 110 may play a role of maintaining the characteristic data and the barcode information in the server by matching them with each other so that the barcode information of the luggage can be secured by using the characteristic data.

When the baggage is determined as unclassified baggage because barcode information is not recognized by the barcode recognition system, the secondary photographing unit 120 acquires characteristic data from the second baggage image of the unclassified baggage photographed by the vision camera. That is, the secondary photographing unit 120 may perform a vision recognition for unclassified baggage for which barcode recognition is not performed, and may play a role of re-recognizing the characteristic data regarding the appearance of the baggage.

The barcode recognition system may be a device that reads the barcode itself tagged on the luggage, identifies a moving destination to which the corresponding baggage is to be moved, and transmits it to the moving device, so that the baggage is moved to and sorted by the identified moving destination.

The uncategorized luggage may refer to luggage whose moving destination cannot be identified because the barcode recognition system cannot accurately read the barcode according to obstacles such as the luggage being overturned or the barcode being damaged.

Thereafter, the secondary photographing unit 120 classifies the unclassified baggage by extracting barcode information registered in response to the acquired characteristic data from the server. That is, the secondary photographing unit 120 searches the server for the characteristic data of the first luggage image having the sameness as the characteristic data of the second luggage image within a selected range, and searches for barcode information stored in correspondence to the found characteristic data. Thus, it can serve to estimate the destination of unclassified baggage.

The selected range may be flexibly determined by the operator of the present invention through experience and experimentation, and for example, it may be determined that the same degree between the characteristic data of the second luggage image and the characteristic data of the first luggage image is 99%.

Ultimately, the secondary photographing unit 120 transmits the barcode information extracted from the server to the barcode recognition system, so that the barcode recognition system can support the movement and sorting of the unclassified baggage to the correct moving destination.

According to an embodiment, the airport baggage vision recognition sorter system 100 may specify the installation positions of the primary photographing unit 110 and the secondary photographing unit 120 that are components.

That is, the primary photographing unit 110 may be installed at the check-in counter, and the secondary photographing unit 120 may be installed at the rear end of the barcode recognition system. This installation position allows the barcode information of the luggage to be more clearly recognized by the primary photographing unit 110 at the time when the barcode is tagged, and also for unclassified luggage in which barcode recognition is not normally performed in the barcode recognition system. It may be a place where feature data can be quickly acquired by the secondary photographing unit 120 .

In addition, the airport baggage vision recognition sorter system 100 may specify a photographing direction in the vision camera attached to the primary photographing unit 110 and the secondary photographing unit 120 .

That is, the vision camera may photograph the first luggage image or the second luggage image in four different viewing directions from the four cameras installed obliquely from top to bottom. Through this, the vision camera can generate a luggage image that captures the front, back, left, right, and omnidirectional directions of the luggage, so that the outer shape of the luggage can be more accurately grasped.

According to an embodiment, the airport baggage vision recognition sorter system 100 may use various algorithms to obtain barcode information from the baggage image.

That is, the primary photographing unit 110 may obtain the barcode information from the first luggage image by applying at least one of YOLO Algorism, CNN and SIFT, and Template Matching Algorism.

YOLO (You Only Look Once) Algorism is designed to improve the slow performance of Region with Convolutional Neural Network (R-CNN), which generates hundreds of image candidates called region proposals and classifies them for each, while YOLO Algorism makes it possible to acquire barcode information, which is an object, at a high speed by dividing it into a grid grid to determine a class at a time and integrating it to classify the final object.

SIFT (Scale Invariant Feature Transform) is one of the representative algorithms for extracting features from images. It enables the acquisition of barcode information as a feature to be pulled from the position.

Template Matching Algorism is an algorithm that searches a reference image for a location that matches a template image. It performs template matching by converting it into a feature space such as edge, corner point, and frequency transformation, and the brightness of the image. Through a normalization process to be less sensitive, it is possible to obtain characteristic barcode information.

According to an embodiment, the airport baggage vision recognition sorter system 100 is passively classified by a classification expert for unclassified baggage that cannot extract barcode information according to the extraction of feature data by the secondary photographing unit 120 . It can induce work to be done.

To this end, the secondary photographing unit 120 guides the manual classification process for the unclassified baggage if the feature data that matches the feature data obtained from the second baggage image within a predetermined ratio is not registered in the server. can do.

The guidance on the manual sorting process may include transmitting information related to the generation of unclassified baggage to a terminal possessed by a sorting expert or notifying it by voice.

That is, when the secondary photographing unit 120 cannot find the characteristic data of the first luggage image having the sameness with the characteristic data of the second luggage image in the selected range in the server, the unclassified luggage is manually classified. In other words, it can support the movement of unclassified baggage to an appropriate place of movement.

According to an embodiment of the present invention, an artificial intelligence camera is installed at the check-in desk to collect barcode information and the external shape of the luggage, and the external shape is taken by taking a secondary picture of the unrecognized luggage when passing through the barcode reader. Through the comparison, it is possible to provide an airport baggage vision recognition sorter system and method that improves the accuracy of the baggage sorting operation by using the barcode information collected during the first filming and reaches the correct drop-off point.

In addition, according to an embodiment of the present invention, feature data extracted through artificial intelligence from a camera and a PC installed at the check-in desk are uploaded to the server, and unclassified baggage is captured secondly to be characterized by artificial intelligence. After extracting the data, it is possible to distribute and reduce the overall load of the system by downloading the feature data uploaded to the server and comparing the feature data only for unclassified baggage.

In the case of baggage classification, conventionally, it consists of only a system with a primary classification function that reads and sorts barcodes attached to goods using barcode recognition automatic goods classification technology. There has been inconvenience to classify.

This manual reclassification operation is a cause of aggravating the human/material burden in airports that handle a large amount of baggage.

The pre-installed sorter system, which is an automatic sorting system for barcode recognition, costs a lot to install, and has about 95% of sorting performance.

In the present invention, the airport baggage vision, which can improve the current 95% sorting rate to 98% or more, by building a total two-stage product classification system through the installation of an additional system while maintaining the existing installed sorter system A recognition sorter system and method are provided.

Accordingly, according to the present invention, it is possible to develop and disseminate a low-cost and high-efficiency system to improve the working environment of field workers and increase customer satisfaction due to increased classification accuracy, rather than the costly improvement of the entire system.

2 is a schematic diagram of a system according to an embodiment of the present invention;

As shown in FIG. 2 , the airport baggage vision recognition sorter system 100 installs an artificial intelligence camera at the check-in desk (counter) to first photograph the baggage. That is, the airport baggage vision recognition sorter system 100 may acquire the appearance and barcode information of the baggage by performing product vision recognition at the starting point of baggage classification.

The airport baggage vision recognition sorter system 100 may perform (primary) barcode recognition through the barcode recognition system.

In addition, the airport baggage vision recognition sorter system 100, by arranging an artificial intelligence camera at the rear end of the barcode recognition system, may take a secondary photograph for baggage whose barcode recognition is not properly performed in the barcode recognition system.

Then, the airport baggage vision recognition sorter system 100 compares the images of the first and second photographed baggage, and obtains barcode information about the baggage for which barcode recognition is not performed, from the first photographed image, By performing recognition data information-based recognition, baggage can be classified correctly.

The airport baggage vision recognition sorter system 100 may apply the latest artificial intelligence image recognition technology to support correctly finding baggage with better recognition performance than the human eye.

3 is a diagram for illustrating a primary photographing unit installed at a check-in desk (counter).

As shown in FIG. 3 , the primary photographing unit 310 is configured by installing four vision cameras at the check-in desk so that the luggage can be photographed obliquely from the top to the bottom in four directions. 4 cameras are connected to one PC, which is connected to the network

The primary photographing unit 310 cross-checks the AI camera, the check-in date, and the pre-classification camera to recognize the shape of the baggage.

As a method of recognizing the shape of a luggage, for example, YOLO Algorism, CNN and SIFT, Template Matching Algorism, etc. can be used.

YOLO Algorism recognizes the brand and characteristics of luggage, checks the boundaries of luggage through ROI recognition, and has a very fast operation speed.

CNN is a type of artificial neural network used to analyze visual images, and SIFT (Scale Invariant Feature Transform) may be a method for extracting feature points.

The primary photographing unit 310 may track and classify detailed features of luggage using CNN and SIFT.

The Template Matching Algorism combines the information extracted from the previous steps and matches the information stored in the BASS system to check baggage.

The secondary photographing unit 320 is installed at the rear end of the barcode recognition system, and similarly to the primary photographing unit 310, four cameras are installed so as to be able to photograph the baggage obliquely from top to bottom in 4 directions. Four cameras are connected to one PC, which is connected to the network and barcode recognition system.

When the baggage is moved from the check-in desk to the conveyor, the primary photographing unit 310 takes photos of the baggage in four directions with four installed vision cameras, processes it with artificial intelligence through a connected PC, extracts characteristic data, and Feature data can be uploaded to the server along with barcode information.

The secondary photographing unit 320, when unsorted baggage that the barcode system cannot classify occurs, takes photos of the baggage again in four directions with four installed vision cameras, and processes the characteristic data with artificial intelligence through a connected PC. After extraction, it is compared with the feature data uploaded to the server. As a result of the comparison, the secondary photographing unit 320 may check the barcode information of the baggage that matches the reliable score, and automatically classify the baggage according to the checked barcode information.

The primary photographing unit 310 does not miss the overall characteristics of the baggage as it shoots the baggage in four directions, and the secondary photographing unit 320 also takes the baggage in four directions, so it is strong against rotations that occur during the movement process. characteristics can be matched.

The airport baggage vision recognition sorter system 100 shoots a four-way image from the primary photographing unit 310 and the secondary photographing unit 320 for a large amount of actual baggage moved at the actual airport, and By matching using the barcode system, it can be used as learning data for artificial intelligence.

The airport baggage vision recognition sorter system 100 applies a CNN-based feature extraction network using artificial intelligence and actually learns about various and large numbers of baggage, so that the color, shape, handle shape, and characteristic features of the baggage It is automatically learned to find patterns, etc., and can classify real baggage with high accuracy.

The airport baggage vision recognition sorter system 100 uploads only the feature data extracted from the images captured by the camera and PC installed at the check-in desk through artificial intelligence to the server, and the secondary photographing unit 320 takes the baggage and uses it as artificial intelligence. After extracting the feature data, download the feature data uploaded to the server and compare the feature data only for unclassified baggage to distribute and reduce the overall load of the system.

4 is a diagram for explaining a change in the overall success rate according to the secondary vision recognition recognition rate.

The airport baggage vision recognition sorter system 100 can improve the accuracy and safety of the airport sorter system based on artificial intelligence vision recognition. The airport baggage vision recognition sorter system 100 can increase the recognition rate by more than 98% by adding the primary barcode recognition of the existing sorter system and the secondary object precision recognition of the artificial intelligence vision recognition system.

For example, if about 12.5 million customers used baggage at OO Airport for one year, and a simple calculation with 95% accuracy and 15 million pieces of baggage, it can be estimated that 750,000 pieces of unclassified baggage were generated.

As shown in Figure 4, if the secondary vision recognition recognition rate is 50%, the airport baggage vision recognition sorter system 100 can reduce the number of unsorted baggage for 15 million pieces of baggage to 375,000 with an accuracy of 97.5%. .

If the vision recognition recognition rate is improved to 95%, the airport baggage vision recognition sorter system 100 can dramatically reduce the number of unsorted baggage for 15 million pieces of baggage to 375 thousand with an accuracy of 99.75%.

The airport baggage vision recognition sorter system 100 according to the present invention aims to reduce unsorted items to less than 50,000 items per year with an accuracy of 98% or more when the first and second sorting tasks are completed.

When the airport baggage vision recognition sorter system 100 increases the accuracy through deep learning, it reduces the steps of work by applying a method of using an artificial intelligence vision recognition camera alone at check-in without a separate barcode at the check-in counter, thereby reducing the steps of the barcode It is possible to reduce misclassification caused by loss.

The airport baggage vision recognition sorter system 100 records the status of moving goods at check-in through big data-based operation, so that customer service can be strengthened by promptly responding to problems with goods in transit.

The airport baggage vision recognition sorter system 100 is a high value-added product based on AI for the 4th industry, which helps in the development of smart airports and can prevent overload of baggage handling facilities at airports during peak season.

Hereinafter, a work flow of the airport baggage vision recognition sorter system 100 according to embodiments of the present invention will be described in detail with reference to FIG. 5 .

The airport baggage vision recognition sorter method according to the present embodiment may be performed by the airport baggage vision recognition sorter system 100 .

5 is a flowchart illustrating an airport baggage vision recognition sorter method according to an embodiment of the present invention.

First, the primary photographing unit of the airport baggage vision recognition sorter system 100 acquires feature data and barcode information from the first baggage image of the baggage photographed by the vision camera ( 510 ). Step 510 may be a process of recognizing the characteristic data regarding the appearance of the luggage and barcode information tagged on the luggage by performing vision recognition on the first luggage image.

Here, the vision camera may be a kind of vision machine that captures image data and generates an uncompressed image of the first luggage, so that it is possible to detect the shape of the luggage and read the information code.

In addition, the characteristic data is data regarding the external characteristics of the luggage, and may include, for example, external characteristics such as width/length/height length, shape, color, texture, and presence or absence of a handle/wheel of the luggage.

In addition, the barcode information may refer to combination information such as a bar-shaped line and a number that are tagged on the surface of a passenger's baggage upon check-in.

In addition, vision recognition may be a technology of extracting barcode information tagged on luggage as meaningful information from the first luggage image using AI time. Such vision recognition is being used in various industries such as autonomous vehicles, industrial robots, medical imaging technology, smart home, intelligent CCTV, access control, facial recognition, and factory automation as deep learning and deep view technologies are combined.

In other words, the primary photographing unit may obtain the characteristic data of the appearance of the luggage and barcode information provided at check-in from the first luggage image obtained by photographing the luggage.

In addition, the primary photographing unit of the airport baggage vision recognition sorter system 100 registers the characteristic data and the barcode information in correspondence with the server ( 520 ). Step 520 may be a process of maintaining the feature data and barcode information in a server by matching them with each other so that the barcode information of the luggage can be secured using the feature data.

Subsequently, the secondary photographing unit 120 of the airport baggage vision recognition sorter system 100 determines that the baggage is unclassified baggage because barcode information is not recognized by the barcode recognition system. 2 Acquire feature data from the hydrate image ( 530 ). Step 530 may be a process of re-recognizing the characteristic data regarding the appearance of the baggage by performing vision recognition on unclassified baggage for which barcode recognition is not performed.

The barcode recognition system may be a device that reads the barcode itself tagged on the luggage, identifies a moving destination to which the corresponding baggage is to be moved, and transmits it to the moving device, so that the baggage is moved to and sorted by the identified moving destination.

The uncategorized luggage may refer to luggage whose moving destination cannot be identified because the barcode recognition system cannot accurately read the barcode according to obstacles such as the luggage being overturned or the barcode being damaged.

Thereafter, the secondary photographing unit of the airport baggage vision recognition sorter system 100 extracts barcode information registered in response to the acquired characteristic data from the server, thereby classifying the unclassified baggage (540). In step 540, the server finds the characteristic data of the first luggage image having the sameness as the characteristic data of the second luggage image in a selected range, and retrieves barcode information stored in correspondence with the found characteristic data, It may be a process of estimating a moving destination.

The selected range may be flexibly determined by the operator of the present invention through experience and experimentation, and for example, it may be determined that the same degree between the characteristic data of the second luggage image and the characteristic data of the first luggage image is 99%.

Ultimately, the secondary photographing unit transmits the barcode information extracted from the server to the barcode recognition system, so that the barcode recognition system can support the unclassified luggage to be moved and sorted to the correct moving destination.

According to an embodiment, the airport baggage vision recognition sorter system 100 may specify the installation positions of the primary photographing unit and the secondary photographing unit, which are components.

That is, the primary photographing unit may be installed at the check-in counter, and the secondary photographing unit may be installed at the rear end of the barcode recognition system. This installation location allows the barcode information of the luggage to be more clearly recognized by the primary photographing unit at the time when the barcode is tagged, and also secondary photographing for unclassified luggage where barcode recognition is not normally performed in the barcode recognition system It may be a place where feature data can be quickly acquired by the department.

In addition, the airport baggage vision recognition sorter system 100 may specify a photographing direction in the vision camera attached to the primary photographing unit and the secondary photographing unit.

That is, the vision camera may photograph the first luggage image or the second luggage image in four different viewing directions from the four cameras installed obliquely from top to bottom. Through this, the vision camera can generate a luggage image that captures the front, back, left, right, and omnidirectional directions of the luggage, so that the outer shape of the luggage can be more accurately grasped.

According to an embodiment, the airport baggage vision recognition sorter system 100 may use various algorithms to obtain barcode information from the baggage image.

That is, the primary photographing unit may obtain the barcode information from the first luggage image by applying at least one of YOLO Algorism, CNN and SIFT, and Template Matching Algorism.

YOLO (You Only Look Once) Algorism is designed to improve the slow performance of Region with Convolutional Neural Network (R-CNN), which generates hundreds of image candidates called region proposals and classifies them for each, while YOLO Algorism makes it possible to acquire barcode information, which is an object, at a high speed by dividing it into a grid grid to determine a class at a time and integrating it to classify the final object.

SIFT (Scale Invariant Feature Transform) is one of the representative algorithms for extracting features from images. It enables the acquisition of barcode information as a feature to be pulled from the position.

Template Matching Algorism is an algorithm that searches a reference image for a location that matches a template image. It performs template matching by converting it into a feature space such as edge, corner point, and frequency transformation, and the brightness of the image. Through a normalization process to be less sensitive, it is possible to obtain characteristic barcode information.

According to the embodiment, the airport baggage vision recognition sorter system 100 performs a passive classification operation by a classification expert for unclassified baggage that cannot extract barcode information according to the extraction of feature data by the secondary photographing unit. can be induced.

To this end, the secondary photographing unit may guide the manual classification process for the unclassified baggage if the feature data that matches the feature data obtained from the second baggage image within a predetermined ratio is not registered in the server.

The guidance on the manual sorting process may include transmitting information related to the generation of unclassified baggage to a terminal possessed by a sorting expert or notifying it by voice.

That is, if the secondary photographing unit cannot find the characteristic data of the first luggage image having the sameness as the characteristic data of the second luggage image in a selected range on the server, the unclassified luggage is manually sorted, so that the unclassified luggage is We can support you to move to an appropriate place of movement.

According to an embodiment of the present invention, an artificial intelligence camera is installed at the check-in desk to collect barcode information and the external shape of the luggage, and the external shape is taken by taking a secondary picture of the unrecognized luggage when passing through the barcode reader. Through the comparison, it is possible to provide an airport baggage vision recognition sorter system and method that improves the accuracy of the baggage sorting operation by using the barcode information collected during the first filming and reaches the correct drop-off point.

In addition, according to an embodiment of the present invention, feature data extracted through artificial intelligence from a camera and a PC installed at the check-in desk are uploaded to the server, and unclassified baggage is captured secondly to be characterized by artificial intelligence. After extracting the data, it is possible to distribute and reduce the overall load of the system by downloading the feature data uploaded to the server and comparing the feature data only for unclassified baggage.

The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

Software may comprise a computer program, code, instructions, or a combination of one or more thereof, which configures a processing device to operate as desired or is independently or collectively processed You can command the device. The software and/or data may be any kind of machine, component, physical device, virtual equipment, computer storage medium or apparatus, to be interpreted by or to provide instructions or data to the processing device. , or may be permanently or temporarily embody in a transmitted signal wave. The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

As described above, although the embodiments have been described with reference to the limited drawings, those skilled in the art may apply various technical modifications and variations based on the above. For example, the described techniques are performed in an order different from the described method, and/or the described components of the system, structure, apparatus, circuit, etc. are combined or combined in a different form than the described method, or other components Or substituted or substituted by equivalents may achieve an appropriate result.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

100: Airport baggage vision recognition sorter system
110: 1st filming unit
120: 2nd filming unit

Claims (11)

a primary photographing unit that acquires characteristic data and barcode information from a first luggage image of the luggage photographed by the vision camera, and registers the characteristic data and the barcode information in correspondence with the server; and
If the baggage is determined as unclassified baggage because barcode information is not recognized by the barcode recognition system,
Secondary shooting of classifying and processing the unclassified luggage by obtaining characteristic data from the second luggage image of the unclassified luggage photographed by the vision camera, and extracting barcode information registered in response to the acquired characteristic data from the server wealth
Airport baggage vision recognition sorter system that includes. According to claim 1,
The first photographing unit,
installed at the check-in counter,
The secondary photographing unit,
installed at the rear end of the barcode recognition system
Airport Baggage Vision Recognition Sorter System. According to claim 1,
The vision camera is
In four cameras installed obliquely from top to bottom, in four different viewing directions, the first luggage image or the second luggage image is taken
Airport Baggage Vision Recognition Sorter System. According to claim 1,
The first photographing unit,
Applying at least one of YOLO Algorism, CNN and SIFT, and Template Matching Algorism to obtain the barcode information from the first luggage image
Airport Baggage Vision Recognition Sorter System. According to claim 1,
The secondary photographing unit,
If the characteristic data obtained from the second luggage image and the characteristic data matching within a predetermined ratio are not registered in the server,
Guide to the manual sorting process for the unsorted baggage
Airport Baggage Vision Recognition Sorter System. obtaining, in the primary photographing unit, characteristic data and barcode information from a first baggage image of the baggage photographed by the vision camera;
registering, in the primary photographing unit, the characteristic data and the barcode information in correspondence with the server;
If the baggage is determined as unclassified baggage because barcode information is not recognized in the barcode recognition method,
obtaining, in the secondary photographing unit, characteristic data from a second baggage image of the unclassified baggage photographed by a vision camera; and
In the secondary photographing unit, by extracting barcode information registered in response to the acquired characteristic data from the server, sorting processing for the unclassified luggage
Airport baggage vision recognition sorter method comprising a. 7. The method of claim 6,
The first photographing unit,
installed at the check-in counter,
The secondary photographing unit,
installed at the rear end of the barcode recognition method
Airport Baggage Vision Recognition Sorter Method. 7. The method of claim 6,
The vision camera is
In four cameras installed obliquely from top to bottom, in four different viewing directions, the first luggage image or the second luggage image is taken
Airport Baggage Vision Recognition Sorter Method. 7. The method of claim 6,
Obtaining the barcode information from the first luggage image by applying at least one of YOLO Algorism, CNN and SIFT, and Template Matching Algorism in the first photographing unit
Airport baggage vision recognition sorter method further comprising. 7. The method of claim 6,
If the characteristic data obtained from the second luggage image and the characteristic data matching within a predetermined ratio are not registered in the server,
In the secondary photographing unit, guiding the manual sorting process for the unsorted baggage
Airport baggage vision recognition sorter method further comprising. A computer-readable recording medium recording a program for executing the method of any one of claims 6 to 10.

Images