CN212418735U

CN212418735U - Automatic goods picking system based on visual identification

Info

Publication number: CN212418735U
Application number: CN202021750226.0U
Authority: CN
Inventors: 林子聪; 雷凯茵; 李迪晋
Original assignee: Hong Kong Productivity Council
Current assignee: Hong Kong Productivity Council
Priority date: 2020-08-20
Filing date: 2020-08-20
Publication date: 2021-01-29
Anticipated expiration: 2030-08-20

Abstract

The utility model provides an automatic goods picking system based on visual identification, which comprises an image acquisition device, an industrial PC and a goods transportation system; wherein, image acquisition device is used for gathering the image information of the goods of waiting to pick up goods, and industry PC is arranged in according to through the visual identification model image information obtains the goods category of goods in the image information, and goods transportation system is used for transporting goods to the processing platform that corresponds according to the goods category, the utility model discloses but the automatic identification goods, thereby help or replace staff's the work of picking up goods and can reduce staff's detection mistake, promote the efficiency of picking up goods.

Description

Automatic goods picking system based on visual identification

Technical Field

The utility model relates to a commodity circulation technical field especially relates to an automatic goods system of choosing based on visual identification.

Background

Many work sites, especially logistics centers, currently require a lot of inspection work. The picking work in the logistics center mainly depends on visual identification and inspection. The employee is not only required to pick up the order, but also to determine that the goods type and quantity are correct. Due to long and repetitive work, errors often occur when picking. In addition, the problem arises for the layman staff due to the lack of language. The wrong type or quantity of goods can lead to the incomplete quantity of the goods in stock and the resource is needed to correct the error, thus reducing the working efficiency.

SUMMERY OF THE UTILITY MODEL

An object of the utility model is to provide an automatic goods system of choosing based on visual identification, automatic identification goods, thereby help or replace staff's the work of choosing goods and can reduce staff's detection mistake, promote the efficiency of choosing goods.

In order to achieve the above purpose, the utility model discloses an automatic goods picking system based on visual identification, which comprises an image acquisition device, an industrial PC and a goods transportation system;

wherein, image acquisition device is used for gathering the image information of the goods of waiting to pick up goods, and industry PC is arranged in through the vision identification model basis the goods category of goods among the image information obtains image information, and goods transportation system is used for transporting goods to the processing platform that corresponds according to the goods category.

Preferably, the picking device further comprises a distance sensing device for collecting distance information of goods to be picked, and the industrial PC is used for obtaining the goods category of the goods in the image information according to the image information and the distance information.

Preferably, the defrosting device is further included;

the defrosting device is used for heating and defrosting the goods to be picked before the image acquisition device acquires the image information of the goods to be picked.

Preferably, the defrosting device is a hot air passage provided on the goods transportation system.

Preferably, the system further comprises a server for obtaining the visual recognition model according to a training data set through a machine learning principle and transmitting the visual recognition model to the industrial PC.

Preferably, the image acquisition device comprises a camera.

Preferably, the image acquisition device comprises a camera for acquiring image information of one angle of the goods or a plurality of cameras for acquiring image information of a plurality of angles of the goods.

Preferably, the distance sensing device comprises a 3D snapshot camera.

Preferably, the goods are placed on the transportation trolley in a layered mode, and the image acquisition device is arranged on the top of the trolley.

Preferably, the goods transport system comprises a controller and a transmission;

the transmission device comprises a main conveying channel provided with a plurality of turntables and branch channels connected with each turntable and the corresponding processing platform;

the controller is used for controlling the rotary table to rotate according to the goods type of the goods so that the goods are transported to the corresponding branch channel to be transported to the corresponding processing platform.

The utility model provides a goods system is picked in automation based on visual identification that one set of suitable goods of picking use, through the quick accurate identification goods under the real-time condition of the visual identification mould that trains, then the goods conveying system through the commodity circulation can be with the different types of goods transportation that the identification obtained to the processing platform that corresponds. When the automatic identification device is used, equipment such as the image acquisition device and the like can be arranged on a goods picking transport line or a trolley, and the automatic identification function assists or replaces the goods picking work of staff, so that the detection errors of the staff can be reduced, the manual inspection is saved, and the working efficiency is improved.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.

FIG. 1 illustrates one of the block diagrams of one embodiment of the vision based automated picking system of the present invention;

fig. 2 is a schematic diagram of a transportation system in an embodiment of the automatic pick-up system based on visual recognition;

figure 3 illustrates a second block diagram of an embodiment of the vision recognition based automated picking system of the present invention;

figure 4 is a schematic diagram of the range of the trolley goods in one embodiment of the automatic goods picking system based on visual identification of the invention;

FIG. 5 is a schematic diagram of a defroster in one embodiment of an automated visual identification based picking system of the present invention;

fig. 6 shows a third block diagram of an embodiment of the automatic picking system based on visual recognition;

fig. 7 is a schematic diagram of an image capturing device in an embodiment of the automatic pick-up system based on visual recognition according to the present invention;

fig. 8 is a schematic diagram of another image capturing device in an embodiment of the automatic picking system based on visual recognition according to the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be described clearly and completely with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments in the present invention, all other embodiments obtained by a person skilled in the art without creative work belong to the protection scope of the present invention.

According to the utility model discloses, this embodiment discloses an automatic goods system of choosing based on visual identification. As shown in fig. 1 and 2, the system includes an image pickup device 1, an industrial PC3, and a goods transportation system 2.

The image acquisition device 1 is used for acquiring image information of goods 4 to be picked. It is understood that the image capturing apparatus 1 may be a camera, a video camera, or the like having an image or video capturing function. Further, when the image capturing device 1 captures a video including the goods 4 to be picked, it is necessary to extract video frames in the video to obtain image information including the goods 4.

The industrial PC3 is used to recognize the article type of the article 4 in the image information by a preset visual recognition model. Wherein, the image information can be gathered through the image acquisition device 1 who adopts this embodiment to predetermined visual model, forms the training data set through the mode of artifical or machine mark goods category, then obtains predetermined visual identification model through machine learning technique, then this visual identification model can carry out automatic, accurate discernment to the goods category in the image information who gathers, guarantees the recognition accuracy degree, can replace and assist manual operation, can realize according to machine learning technique for technical staff in the field.

The goods transportation system 2 is used to transport the goods 4 to the corresponding processing platform according to the goods category. It can be understood that, in this embodiment, the goods transportation system 2 has a function of transferring the classified goods 4 to the corresponding processing platform, that is, after the goods category of the goods 4 is obtained through the visual recognition model, the transportation system 2 is controlled according to the goods category to adjust the transportation direction of the goods 4, so that the goods 4 reach the corresponding processing platform, and then the goods are carried by hands or a robot and placed on a designated container or a shelf.

The utility model provides a one set is fit for picking the goods system of choosing based on visual identification who uses of picking up goods, through the quick accurate identification goods 4 under the real-time condition of the visual identification mould that trains, then the goods conveying system 2 through the commodity circulation can be with the different types of goods 4 transportation that the identification obtained to the processing platform that corresponds. When the automatic identification device is used, equipment such as the image acquisition device 1 and the like can be arranged on a goods picking transport line or a trolley, and the automatic identification function assists or replaces the goods picking work of staff, so that the detection errors of the staff can be reduced, the manual inspection is saved, and the working efficiency is improved.

In a preferred embodiment, as shown in fig. 3, the system further comprises a distance sensing means 5. Wherein the distance sensing device 5 is used for collecting distance information of the goods 4 to be picked.

The industrial PC3 is used to obtain the goods category of the goods in the image information based on the image information and the distance information. Before the goods category of the goods 4 in the image information is identified through a preset visual identification model, the identification range of the image information is determined according to the distance information so that the preset visual identification model identifies the goods category of the goods 4 in the identification range of the image information.

The distance sensing device 5 may be a 3D snapshot camera, for example, a laser snapshot type 3D camera, but in other embodiments, other distance sensing devices 5 with distance measurement functions may also be used. It can be understood that the 3D snapshot camera is adopted to shoot the goods 4, so that the depth information of the goods 4 and the environment can be obtained, that is, the distance information between the goods 4 and the distance sensing device 5 can be obtained, and the recognition range of the goods 4 far away from the image acquisition device 1 is relatively small; conversely, the closer the article 4 is to the image pickup device 1, the larger the recognition range. According to the corresponding relation between the range of the goods 4 on the image information of the goods 4 acquired by the image acquisition device 1 and the goods 4 and the distance information, the range of the goods 4 on the image information acquired in real time can be obtained by combining the real-time distance information, so that the identification range of the visual identification model can be limited, and the identification efficiency of the visual identification model can be improved.

In one embodiment of the application, as shown in fig. 4, the goods 4 are arranged in layers in the trolley, the image pickup device 1 is arranged on the top of the trolley, and when the distance sensing device 5 is further arranged, the distance sensing device 5 is preferably also arranged on the top of the trolley at the same height as the image pickup device 1. The visual recognition model may be set on an industrial PC, which is connected to the image acquisition device 1 and the distance sensing device 5, respectively. The image acquisition device 1 acquires the image information of the goods 4 in the trolley and the distance sensing device 5 acquires the distance information of the goods 4 in the trolley, and then the image information and the distance information are respectively transmitted to the working PC for data processing and goods category identification.

The image processing model preset in the working PC determines the distance between the goods 4 and the distance acquisition device according to the acquired distance information, so that the distance between the goods 4 and the image acquisition device 1 can be obtained, and the specific layer number of the goods 4 positioned on the trolley can be determined. Furthermore, the range of the goods on the acquired image information can be determined according to the distance information, namely the identification range is obtained, so that the visual identification model carries out image identification in the identification range, the image identification process of useless areas can be reduced, and the identification efficiency of the goods 4 is improved.

In a preferred embodiment, referring to fig. 5, the system further comprises a defrost device 6. The defrosting device 6 is used to perform heating defrosting processing on the goods 4 to be picked before the image information of the goods 4 to be picked is acquired by the image acquisition device 1. It will be appreciated that since the present application identifies the type of the goods 4 by means of image information, when some refrigerated products are initially transported, a layer of frost is generally formed on the surface thereof. Therefore, in the preferred embodiment, the goods 4 to be picked may be subjected to a heating defrosting process, for example, a hot air channel may be provided on the goods transportation system, so that the goods 4 are transported through the hot air channel, and frost on the surface of the goods 4 is removed by the hot air process, so that the image capturing device 1 can clearly capture the surface characteristics of the goods 4 for visual recognition.

In a preferred embodiment, the system further comprises a server 7, as shown in fig. 6. The server 7 is used for obtaining the visual recognition model according to the training data set through a machine learning principle and transmitting the visual recognition model to the industrial PC. It is understood that the server may be a single server set by a computer, or may be a server cluster composed of a plurality of the aforementioned servers.

In one embodiment, the training data set is first uploaded to the server 7 before training the model. The training data set is typically a processed pick video, which is a frame of annotated video. The annotation is performed manually, and an identification range is set for circling the goods 4 on each video frame. The annotation data is converted into a text file and uploaded to the server together with the video frame images. The data are processed by an example segmentation model convolutional neural network (Mask R-CNN) method, characteristics are extracted from RGB data in a circling range, and the extracted characteristics correspond to manually marked goods types to obtain a visual identification model. In practical application, the visual recognition model can read color data (RGB) in image information to be recognized, and then process the data by using a Mask R-CNN (Mask R-CNN) mode of an example segmentation model to obtain features within a range so as to recognize goods 4 in the image.

In this embodiment, referring to fig. 7, the image capturing device 1 may capture image information of an angle of the goods 4, and identify characteristics of the angle of the goods 4 through the visual recognition model to obtain the goods category. In other preferred embodiments, referring to fig. 8, the image capturing device 1 may also capture image information of multiple angles of the goods 4, obtain a visual identification model capable of identifying the goods category at multiple angles through a machine learning technique after manual labeling, and capture the image information of the goods 4 at multiple angles to identify the goods 4 at multiple angles, so as to improve the accuracy of the goods category identification. Wherein, can set up image acquisition devices 1 such as camera in the different position of goods 4 to gather the image information of 4 a plurality of angles of goods.

In a preferred embodiment, the goods transport system comprises a controller and a transmission.

It is understood that the controller of the goods transportation system may receive the goods category of the goods 4 transferred by the industrial PC3, determine the processing platform to which the currently received goods category should be transported according to the correspondence between the goods category and the processing platform, and further determine the turntable corresponding to the processing platform. The position of goods 4 can be confirmed through transmission's transportation, and when goods 4 transported to the revolving stage that corresponds, the control revolving stage rotated and conveyed goods 4 to the branch passageway that corresponds, and then transported to the processing platform that corresponds and supply manual work or robot to handle.

The utility model discloses a visual identification technique brings a low-cost automatic goods system of choosing for the logistics industry, realizes easily, only needs to let software train just can begin work before the use. Compared with the robot automation, the traditional logistics center can enjoy improved work efficiency without additionally modifying the operation mode. The automatic picking system can improve the working efficiency for a plurality of industries. In the aspect of the inspection, the utility model discloses the kind and the quantity of identification article that can be quick and accurate. Compared with the manpower inspection, the utility model discloses can reduce the mistake of checking. Typical identification systems rely on network connections to transmit data to an image recognition server for processing of the image data. The utility model discloses can the independent operation, as long as download trained visual identification model on the device, just need not have network connection and operate. Therefore, the utility model discloses do not need high bandwidth internet infrastructure also can quick operation.

The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments.

The above description is only an example of the present application and is not intended to limit the present application. Various modifications and changes may occur to those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. An automatic goods picking system based on visual identification is characterized by comprising an image acquisition device, an industrial PC and a goods transportation system;

2. The automatic picking system based on visual identification according to claim 1, characterized in that further comprises a distance sensing device for collecting distance information of goods to be picked, the industrial PC is used for obtaining the goods category of the goods in the image information according to the image information and the distance information.

3. The automatic picking system based on visual recognition of claim 1, further comprising a defrosting device;

4. The automatic picking system based on visual recognition of claim 3, wherein the defrosting device is a hot air channel provided on a goods transportation system.

5. The automatic picking system based on visual recognition of claim 1, further comprising a server for obtaining the visual recognition model through machine learning principles from a training data set and transmitting the visual recognition model to an industrial PC.

6. The automated visual identification-based order picking system of claim 1, wherein the image capture device comprises a camera.

7. The automatic picking system based on visual identification according to claim 6, characterized in that the image collecting device comprises a camera collecting image information of an angle of goods or a plurality of cameras collecting image information of a plurality of angles of goods.

8. The automated visual recognition-based order picking system of claim 2, wherein the distance sensing device comprises a 3D snapshot camera.

9. The automatic picking system based on visual recognition of claim 1, wherein the goods are layered on a transportation trolley and an image capturing device is provided on the top of the trolley.

10. The automatic pick-up system based on visual identification as claimed in claim 1, wherein the goods transportation system comprises a controller and a transmission;