KR102578727B1 - Method for tracking management and processing of logistics based on artificial intelligence - Google Patents

Abstract

The present invention relates to an artificial intelligence-based logistics tracking management and processing method, comprising: a first step of collecting and acquiring video and audio information using a device including a camera module and a microphone module attached to fixed equipment and mobile equipment; A second step of independently separating the acquired video and audio information to remove noise signals and automatically determining them into identifiable patterns through an artificial intelligence calculation process; A third step of specifying shape recognition information by analyzing logistics and worker shapes based on the determined pattern information; A fourth step of specifying relationship recognition information by analyzing the relationship between unit shapes classified as a result of the third step; A fifth step of correcting the final recognition information and combining the identification information by adding weights to the results of the third and fourth steps; And a sixth step of tracking and managing and processing logistics inside and outside the logistics processing site using the results of the fifth step.

Description

{Method for tracking management and processing of logistics based on artificial intelligence}

The present invention relates to an artificial intelligence-based logistics tracking management and processing method, and more specifically, to a logistics tracking management and processing method utilizing artificial intelligence that recognizes audio-visual and text information and operates for a set purpose.

With the development of artificial intelligence technology that can interpret various information according to specific purposes and robotics technology that can precisely process complex mechanical processes, the concept of artificial intelligence-based smart automation is being established in detail in various industrial fields. Recently, commercial artificial intelligence technology has mainly developed into a form that uses deep learning algorithms to recognize information such as video, voice, and text and convert it into an appropriate form according to the purpose. The added value created by this technology is self-driving cars, smart cars, etc. Commercial products such as factories and smart assistants can be found all around us. On the other hand, robotics technology has been developed to effectively perform tasks that require significant labor or are risky through machines such as two- or four-legged robots, and wearables to improve worker productivity, work efficiency, or protect safety. It has also been used in the manufacture of various types of machinery.

Meanwhile, in the current logistics industry, the work of recognizing objects, appropriately classifying, packaging, and moving logistics largely relies on human labor. When sorting objects that are difficult to transport in large quantities or small, detailed objects in a short period of time, damage to property and loss of goods occur due to limitations in manpower and human error, and the volume of goods transported relative to manpower increases every year, resulting in enormous volume. Economic and industrial losses are being caused.

The purpose of the present invention to solve the above problems is to recognize data of video, voice, and text information and utilize artificial intelligence for the set purpose to improve overall work efficiency and logistics processing flexibility and to improve comprehensiveness of handling logistics. The goal is to provide improved artificial intelligence-based logistics tracking management and processing methods.

The artificial intelligence-based logistics tracking management and processing method of the present invention to achieve the above objectives collects and acquires video and audio information using devices including camera modules and microphone modules attached to fixed equipment and mobile equipment. Step 1; A second step of independently separating the acquired video and audio information to remove noise signals and automatically determining them into identifiable patterns through an artificial intelligence calculation process; A third step of specifying shape recognition information by analyzing logistics and worker shapes based on the determined pattern information; A fourth step of specifying relationship recognition information by analyzing the relationship between unit shapes classified as a result of the third step; A fifth step of correcting the final recognition information and combining the identification information by adding weights to the results of the third and fourth steps; And a sixth step of tracking and managing and processing logistics inside and outside the logistics processing site using the results of the fifth step.

The recognition information is in the form of video or audio, and the artificial intelligence calculation process uses a filter involving convolutional operation or sets N unit information divided by input time into one sequential input unit. It involves a learning structure that undergoes transformation by a non-linear function and is characterized by estimating identifiable patterns by performing a classification or discrimination task.

The recognition information is processed by an application processor inside a device attached to fixed equipment or mobile equipment, or transmitted through a network and processed by a remote server.

The video and audio information acquired in the second step are independently divided, noise signals are removed, normalized, and converted into identifiable patterns through an artificial intelligence calculation process, and multiple object shapes are formed within one video and audio. It is characterized by being detected.

In the case of an image, the pattern information determined in the third step can be converted into R-dimensional embedding by designating an area specified by P pixel coordinates as one shape and matching it with a predefined identification label. On the other hand, in the case of voice, it is characterized by converting it into text based on a dictionary and using the word, phrase, or sentence of the information as a label.

In the fourth step, the relationship between unit shapes is analyzed based on the similarity between the identification label and the embedding of the shape.

In the fifth step, the shape recognition information in the third step and the relationship recognition information in the fourth step are related to the recognition information acquired by M fixed devices in the first step, and the recognition information acquired by L mobile devices in the first step. A preset or changeable reliability weight (weight range: 0 to 1, total weight) is given to the third stage shape recognition information and the fourth stage relationship recognition information, and then linearly added to the third stage shape recognition information and It is characterized by correcting the relationship recognition information in the fourth step to derive the final shape recognition information and final relationship recognition information to form a database, or by combining it with identification information of text recorded in an external logistics management DB, etc. to expand into a new database. .

Using the final shape recognition information and final relationship recognition information from the 6th to 5th steps, external inspection and logistics pattern analysis of incoming and outgoing logistics for logistics tracking management and processing inside and outside the work site, object location specification, and It is characterized by providing functions for quantity management, abnormal logistics detection, small object identification, and multi-product logistics classification.

As described above, according to the artificial intelligence-based logistics tracking management and processing method of the present invention, various types of information obtained in the work process are combined in logistics tracking management and processing, and the reproduction information analyzed according to the purpose is used to provide a wider range of types. It enables flexible handling and processing of logistics, improves operator convenience, and further allows some tasks to be replaced by machines, thereby increasing automation efficiency, reducing overall logistics damage, and increasing processing volume per unit time, creating economic benefits. Industrial effects can be expected.

Figure 1 is a configuration diagram of an artificial intelligence-based logistics tracking management and processing system according to the present invention.
Figure 2 is a flowchart of an artificial intelligence-based logistics tracking management and processing method according to the present invention.

Below, with reference to the attached drawings, embodiments of the present invention will be described in detail so that those skilled in the art can easily implement the present invention. However, the present invention may be implemented in many different forms and is not limited to the embodiments described herein.

Next, a preferred embodiment of the artificial intelligence-based logistics tracking management and processing system and method according to the present invention will be described in detail.

Figure 1 is a configuration diagram of an artificial intelligence-based logistics tracking management and processing system according to the present invention.

Referring to Figure 1, the artificial intelligence-based logistics tracking management and processing system according to the present invention includes a collection unit 100, a preprocessing unit 200, a determination unit 300, an analysis unit 400, a correction unit 500, It is comprised of a management and processing unit 600.

The collection unit 100 collects and acquires video and audio information as recognition information using devices including a camera module 110 and a microphone module 120 attached to fixed equipment and mobile equipment.

The recognition information is in the form of video or audio, and the artificial intelligence calculation process uses a filter involving convolutional calculation or sets N unit information divided by input time into one sequential input unit. It involves a learning structure that undergoes transformation by a non-linear function, and identifiable patterns can be estimated by performing classification or discrimination tasks.

Meanwhile, the recognition information can be processed by an application processor inside a device attached to M fixed devices or L mobile devices, or transmitted through a network and processed by a remote server.

The preprocessing unit 200 independently separates the acquired video and audio information and preprocesses them, such as removing noise signals. That is, the pre-processing unit 200 independently separates the acquired video and audio information, removes noise signals, normalizes them, and converts them into identifiable patterns through an artificial intelligence calculation process, within one video and audio. Multiple object shapes can be detected.

The determination unit 300 automatically determines the pattern into an identifiable pattern through an artificial intelligence calculation process.

The analysis unit 400 analyzes the shape of logistics and workers based on the determined pattern information to specify shape recognition information, and specifies relationship recognition information by analyzing the relationship between unit shapes divided by specific shape recognition information. Here, the determined pattern information can be converted into R-dimensional embedding by designating an area specified by P pixel coordinates as one shape in the case of an image, and matching it with a predefined identification label or using an arbitrary While an identification label is assigned, voice can be converted to text based on a dictionary and the word, phrase, or sentence of that information can be used as a label. Meanwhile, relationships between unit shapes are analyzed based on the similarity between identification labels and shape embeddings.

The correction unit 500 corrects the final recognition information and the final relationship recognition information by summing the weights of the shape recognition information and the relationship recognition information and combines the identification information. Here, the final shape recognition information and final relationship recognition information are used to inspect the appearance of incoming and outgoing logistics and analyze logistics patterns for logistics tracking management and processing inside and outside the work site, specifying object location and quantity management, and detecting abnormal logistics. It can identify small objects and differentiate logistics of various types.

The management and processing unit 600 corrects the final recognition information and the final relationship recognition information, uses the identification information to manage and process logistics tracking inside and outside the logistics processing site, and manages and processes logistics tracking outside the logistics processing site. can do.

Figure 2 is a flowchart of an artificial intelligence-based logistics tracking management and processing method according to the present invention.

Referring to Figure 2, the artificial intelligence-based logistics tracking management and processing method according to the present invention first collects and acquires video and audio information using a device including a camera module and microphone module attached to fixed equipment and mobile equipment. Do it (S110).

Here, the recognition information is in the form of video or audio, and the artificial intelligence calculation process uses a filter involving convolutional operation or sets N unit information divided by input time into one sequential input unit. Therefore, it involves a learning structure that undergoes transformation by a non-linear function, and an identifiable pattern can be estimated by performing a classification or discrimination task.

Meanwhile, the recognition information can be processed by an application processor inside a device attached to M fixed devices or L mobile devices, or transmitted through a network and processed by a remote server.

Next, the acquired video and audio information are independently classified and pre-processed to remove noise signals, and are automatically determined into identifiable patterns through an artificial intelligence calculation process (S120).

The video and audio information acquired in step S120 are independently divided, noise signals are removed, normalized, and converted into identifiable patterns through an artificial intelligence calculation process, and multiple object shapes are formed within one video and audio. This can be detected.

Next, the shape of the logistics and worker is analyzed based on the determined pattern information to specify shape recognition information (S130).

In the case of an image, the pattern information determined in step S130 can be converted into R-dimensional embedding by designating an area specified by P pixel coordinates as one shape, and matching it with a predefined identification label. While an arbitrary identification label can be assigned, voice can be converted to text based on a dictionary and the word, phrase, or sentence of that information can be used as a label.

Next, the relationship between the unit shapes classified as a result of step S130 is analyzed to specify relationship recognition information (S140).

In the fourth step, the relationship between unit shapes can be analyzed based on the similarity between the identification label and shape embedding.

Next, the results of steps S130 and S140 are weighted to correct the final recognition information and the identification information is combined (S150).

Here, the shape recognition information in step S130 and the relationship recognition information in step S140 for the recognition information acquired by M fixed devices in steps S150 to S110, and the step S130 for recognition information acquired by L mobile devices in step S110. A preset or changeable reliability weight (weight range: 0 to 1, total weight) is given to the shape recognition information and the relationship recognition information of step S140, and then linearly added to obtain the shape recognition information of step S130 and the relationship recognition information of step S140. can be corrected to derive the final shape recognition information and final relationship recognition information to form a database, or it can be expanded into a new database by combining it with the identification information of text recorded in an external logistics management DB, etc.

Next, the results of step S150 are used to track and manage logistics inside and outside the logistics processing site (S160).

Using the final shape recognition information and final relationship recognition information from steps S160 to S150, external appearance inspection and logistics pattern analysis of incoming and outgoing logistics, specification of object location, and quantity management for logistics tracking management and processing inside and outside the work site. It can provide functions for detecting abnormal logistics, identifying small objects, and distinguishing between various types of logistics.

Although the embodiments of the present invention have been described in detail above, the scope of the present invention is not limited thereto, and various modifications and improvements made by those skilled in the art using the basic concept of the present invention defined in the following claims are also possible. It falls within the scope of rights.

100: collection unit 110: camera module
120: microphone module 200: preprocessing unit
300: Determination unit 400: Analysis unit
500: Correction Department 600: Management and Processing Department

Claims (8)

A first step of collecting and acquiring video and audio information using devices including camera modules and microphone modules attached to fixed equipment and mobile equipment;
A second step of independently separating the acquired video and audio information to remove noise signals and automatically determining them into identifiable patterns through an artificial intelligence calculation process;
A third step of specifying shape recognition information by analyzing logistics and worker shapes based on the determined pattern information;
A fourth step of specifying relationship recognition information by analyzing the relationship between unit shapes classified as a result of the third step;
A fifth step of correcting the final recognition information and combining the identification information by adding weights to the results of the third and fourth steps; and
A sixth step of tracking and managing logistics inside and outside the logistics processing site using the results of the fifth step,
In the fifth step, the shape recognition information in the third step and the relationship recognition information in the fourth step are related to the recognition information acquired by M fixed devices in the first step, and the recognition information acquired by L mobile devices in the first step. A preset or changeable reliability weight (weight range: 0 to 1, total weight) is given to the third stage shape recognition information and the fourth stage relationship recognition information, and then linearly added to the third stage shape recognition information and Characterized by correcting the relationship recognition information in the fourth step to derive the final shape recognition information and final relationship recognition information to form a database or combining it with identification information of text recorded in an external logistics management DB, etc. to expand into a new database. Artificial intelligence-based logistics tracking management and processing method. According to paragraph 1,
The recognition information is in the form of video or audio, and the artificial intelligence calculation process uses a filter involving convolutional operation or sets N unit information divided by input time into one sequential input unit. An artificial intelligence-based logistics tracking management and processing method that involves a learning structure that undergoes transformation by a non-linear function and performs a classification or discrimination task to estimate identifiable patterns. According to paragraph 1,
An artificial intelligence-based logistics tracking management and processing method, characterized in that the recognition information is processed by an application processor inside a device attached to fixed equipment and mobile equipment, or transmitted through a network and processed by a remote server. According to paragraph 1,
The video and audio information acquired in the second step are independently divided, noise signals are removed, normalized, and converted into identifiable patterns through an artificial intelligence calculation process, and multiple object shapes are formed within one video and audio. An artificial intelligence-based logistics tracking management and processing method characterized by detection. According to paragraph 1,
In the case of an image, the pattern information determined in the third step can be converted into R-dimensional embedding by designating an area specified by P pixel coordinates as one shape and matching it with a predefined identification label. Artificial intelligence is characterized by converting voice into text based on a dictionary and using the word, phrase, or sentence of that information as a label. Based logistics tracking management and processing methods. According to paragraph 1,
An artificial intelligence-based logistics tracking management and processing method, characterized in that in the fourth step, the relationship between unit shapes is analyzed based on the similarity between the identification label and the embedding of the shape. delete According to paragraph 1,
Using the final shape recognition information and final relationship recognition information from the 6th to 5th steps, external inspection and logistics pattern analysis of incoming and outgoing logistics for logistics tracking management and processing inside and outside the work site, object location specification, and An artificial intelligence-based logistics tracking management and processing method that provides functions for quantity management, abnormal logistics detection, small object identification, and multi-product logistics classification.