Video data analysis system using AI

Abstract

The present invention relates to an AI-based image data analysis system implemented to automatically analyze and process image data using artificial intelligence (AI) technology, comprising: an image capturing device that captures an image analysis target from multiple directions to acquire at least one image; a media distribution unit that transmits an image received from the image capturing device; and an image processing unit that uses a deep learning model with a GPU embedded in a processor to estimate metadata from an image transmitted through the media distribution unit, stores the estimated metadata in a database, and decodes and processes an image received from the media distribution unit into an outputtable form.

Description

Technology Field

[0001] The present invention relates to an image data analysis system using AI, and more specifically, to an image data analysis system using AI implemented to automatically analyze and process image data by utilizing artificial intelligence (AI) technology. Background Technology

[0003] Target AI-based object detection technologies include R-CNN (Region-based Convolutional Neural Network), Faster R-CNN, and YOLO. R-CNN consists of a region proposal stage that estimates the location of objects in an image and a classification stage that classifies object types, while Faster R-CNN improves processing speed by integrating the region proposal and classification stages into a single network. Additionally, YOLO, released in 2016, is a model that detects objects by processing the entire image at once, contributing to increased processing speed and implementation efficiency.

[0004] While YOLO has continuously improved its performance through version updates, the complexity of the network has increased along with this improvement. Training complex networks requires a large amount of training data, but in real-world applications, it is difficult to obtain such data, and there are often cases where simple and inexpensive networks are required.

[0005] Therefore, in order to effectively perform image analysis using artificial intelligence models in various application fields, it is crucial to easily generate training data for each field.

[0006] Meanwhile, the aforementioned background technology is technical information that the inventor possessed for the derivation of the present invention or acquired during the process of deriving the present invention, and it cannot necessarily be considered publicly known technology disclosed to the general public prior to the filing of the present invention. Prior art literature

[0008] Korean Registered Patent No. 10-2588729 (Published on Oct. 16, 2023) The problem to be solved

[0009] One aspect of the present invention provides an AI-based image data analysis system implemented to automatically analyze and process image data using artificial intelligence (AI) technology.

[0010] The technical problems of the present invention are not limited to those mentioned above, and other unmentioned technical problems will be clearly understood by those skilled in the art from the description below. means of solving the problem

[0012] An image data analysis system using AI according to one embodiment of the present invention comprises: an image capturing device that captures an image analysis target from multiple directions to acquire at least one image; a media distribution unit that transmits an image received from the image capturing device; and an image processing unit that uses a deep learning model with a GPU embedded in a processor to estimate metadata from an image transmitted through the media distribution unit, stores the estimated metadata in a database, and decodes and processes an image received from the media distribution unit into an outputtable form.

[0013] In one embodiment, the image capturing device may include: a mounting table that rotates the image analysis target when the image analysis target is mounted thereon; a moving rail that extends in a curved shape along the periphery of the image analysis target mounted on the mounting table; and at least one capturing module that is connected and installed to enable sliding movement on the moving rail, moves along the moving rail, captures an image of the image analysis target mounted on the mounting table, and then transmits it to the media distribution unit.

[0014] In one embodiment, the mounting table may include: a table body; a rotating shelf formed in the shape of a disc and rotatably connected and installed on the top of the table body, which rotates when the image analysis target is placed thereon; and an object cleaning unit installed along the circumference of the rotating shelf, which is driven when the image analysis target is placed on the rotating shelf to remove foreign matter attached to the image analysis target.

[0015] In one embodiment, the object cleaning unit may include: a plurality of support frames that are connected in an interlocking manner so as to be upright and vertically movable around the circumference of the rotating shelf; a plurality of lifting drive gears that are rotatably installed inside the rotating shelf and are driven to rotate in a forward or reverse direction by engaging with gear teeth formed at the rear end of the support frames to move the support frames up and down; a cylindrical cover formed in the shape of a cylindrical pipe, supported by the plurality of support frames installed on an inward surface, and moves up and down together with the support frames as they move up and down; and a plurality of air injection units installed along the inner circumference of the cylindrical cover, which spray compressed air toward the image analysis target to remove foreign matter when the image analysis target seated on the rotating shelf is placed inside the cylindrical cover.

[0016] In one embodiment, the support frame may include: a frame body installed to be upright around the circumference of the rotating shelf and capable of moving up and down in the vertical direction, and connected by gear coupling with the lifting drive gear; a plurality of support sliders installed at regular intervals along the front end of the frame body facing the inward surface of the cylindrical cover, capable of moving horizontally by engaging with a horizontal movement groove formed extending horizontally along the inward surface of the cylindrical cover, and supporting the cylindrical cover; and a rotary drive gear rotatably installed on the support slider, engaged by gear coupling with a gear tooth formed on the inner side of the horizontal movement groove, and driven to rotate in a forward or reverse direction to rotate the cylindrical cover.

[0017] In one embodiment, the air injection unit may include: a cone-shaped nozzle formed by being recessed into the inner circumference of the cylindrical cover, with the inner diameter gradually decreasing towards the inside; a spray nozzle installed inside the nozzle to spray compressed air through the nozzle; and a variable discharge nozzle installed around the inlet of the nozzle to induce the compressed air discharged through the nozzle to diffuse or concentrate.

[0018] In one embodiment, the variable discharge nozzle comprises: a wing seating groove formed by recessing along the circumference of the inlet of the nozzle into a conical shape corresponding to the inclined surface of the nozzle; a plurality of first guide wings spaced apart at regular intervals along the inner side of the wing seating groove; a first gap cover made of a thin film of an elastic material capable of expansion and contraction and installed between the plurality of first guide wings; a plurality of second guide wings spaced apart at regular intervals along the wing seating groove, connected and installed to be rotatable toward the nozzle at the front end of the first guide wings; a second gap cover made of a thin film of an elastic material capable of expansion and contraction and installed between the plurality of second guide wings; a first driving screw thread installed in the front-rear length direction along each inward surface of the plurality of first guide wings; and a plurality of second driving screw threads connected to the first driving screw thread and installed in the front-rear length direction along each inward surface of the plurality of second guide wings. It may include a wing drive gear that is connected to the inward-facing surface of the wing mounting groove on the entrance side, facing each of the plurality of second drive screw threads, so as to enable rotational drive, and is connected to the second drive screw threads by gear coupling, thereby exposing the second guide wing from the wing mounting groove as it is rotated in the forward direction, then exposing the first guide wing from the wing mounting groove by engaging with the first drive screw threads by gear coupling, and sequentially storing the first guide wing and the second guide wing into the wing mounting groove as it is rotated in the reverse direction; and a wing support ledge that is installed on one side and the other side of the inward-facing surface of the first guide wing, respectively, and supports the second guide wing so that it cannot rotate further after being exposed from the wing mounting groove and rotating due to the tension of the second gap cover.

[0019] In one embodiment, the frame body may have a wing movement groove formed along the contact surface with the cylindrical cover after the air injection part moves to move the first guide wing and the second guide wing when the cylindrical cover rotates after the first guide wing and the second guide wing are exposed from the wing seating groove. Effects of the invention

[0021] According to one aspect of the present invention described above, first, by utilizing AI through automated image processing and analysis, large-scale image data can be analyzed quickly and efficiently, and this can be performed much faster and more accurately than by manual processing by humans.

[0022] In addition, through improved accuracy, high accuracy can be achieved in tasks such as pattern recognition and object detection using machine learning algorithms and deep learning models.

[0023] In addition, real-time analysis is possible, allowing for rapid response in situations requiring immediate action by analyzing video in real time.

[0024] In addition, through data-driven decision support, the AI ​​system can support better decision-making by providing meaningful insights based on analyzed video data.

[0025] In addition, operating costs can be reduced by decreasing the time and manpower required for large-scale data analysis.

[0026] The effects of the present invention are not limited to those mentioned above, and various effects may be included within the scope obvious to a person skilled in the art from the contents described below. Brief explanation of the drawing

[0028] FIG. 1 is a diagram showing the schematic configuration of an image data analysis system using AI according to one embodiment of the present invention. Figure 2 is a drawing showing the image capturing device of Figure 1. Figure 3 is a drawing showing the mounting table of Figure 2. Figures 4 and 5 are drawings showing the object cleaner of Figure 3. Figure 6 is a drawing showing the support frame of Figure 4. Figure 7 is a drawing showing the air injection part of Figure 6. FIGS. 8 to 10 are drawings showing the variable discharge nozzle of FIG. 8. Specific details for implementing the invention

[0029] The following detailed description of the invention refers to the accompanying drawings, which illustrate specific embodiments in which the invention may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention. It should be understood that various embodiments of the invention are different but need not be mutually exclusive. For example, specific shapes, structures, and characteristics described herein may be implemented in other embodiments without departing from the spirit and scope of the invention in relation to one embodiment. It should also be understood that the location or arrangement of individual components within each disclosed embodiment may be changed without departing from the spirit and scope of the invention. Accordingly, the following detailed description is not intended to be limiting, and the scope of the invention is limited only by the appended claims, including all equivalents to those claimed therein, provided appropriately described. Similar reference numerals in the drawings refer to the same or similar functions across various aspects.

[0030] Hereinafter, preferred embodiments of the present invention will be described in more detail with reference to the drawings.

[0031] FIG. 1 is a diagram showing the schematic configuration of an image data analysis system using AI according to one embodiment of the present invention.

[0032] Referring to FIG. 1, an image data analysis system (10) using AI according to one embodiment of the present invention includes an image capturing device (100), an image capturing device (100), a media distribution unit (200), a database (300), and an image processing unit (400).

[0033] The video recording device (100) captures a video analysis target (P) from multiple directions, acquires at least one video, and then transmits it to a media distribution unit (200) via a network.

[0034] The media distribution unit (200) transmits the video received from the video recording device (100) to the video processing unit (400) via a network.

[0035] The image processing unit (400) estimates metadata from the image transmitted through the media distribution unit (200) using a deep learning model with a GPU embedded in the processor, stores the estimated metadata in the database (300), and decodes the image received from the media distribution unit (200) and processes it into an outputtable form.

[0036] An AI-based image data analysis system (10) according to one embodiment of the present invention having the configuration described above is a system that automatically analyzes and processes image data by utilizing artificial intelligence (AI) technology, and can extract important information from images or videos captured using an image capturing device (100) by applying AI technologies such as machine learning, deep learning, and computer vision, and can analyze the information to utilize it for a specific purpose.

[0037] For example, such a system can perform the following tasks.

[0038] Examples of this include object detection, which identifies and locates specific objects within a video; action recognition, which analyzes and recognizes the behavior of people or objects in a video; face recognition, which recognizes, compares, or identifies faces; and anomaly detection, which automatically detects abnormal behavior or situations.

[0039] Accordingly, the AI-based image data analysis system (10) according to one embodiment of the present invention having the configuration described above can be utilized in various fields such as surveillance, security, autonomous driving, healthcare, and media analysis, and can significantly reduce the time and effort required for direct human analysis by increasing the processing speed and accuracy of image data through AI.

[0041] Figure 2 is a drawing showing the image capturing device of Figure 1.

[0042] Referring to FIG. 2, the image capturing device (100) includes a mounting table (110), a moving rail (120), and at least one capturing module (130).

[0043] The mounting table (110) rotates the image analysis target (P) when the image analysis target (P) is placed on it.

[0044] The moving rail (120) is formed to extend in a curved shape along the periphery of the image analysis object (P) that is seated on the mounting table (110).

[0045] In one embodiment, the moving rail (120) may be composed of two vertical supports (121) installed upright and a curved rail (122) that is rotatably connected between the two vertical supports (121) so as to allow at least one shooting module (130) to slide.

[0046] At least one shooting module (130) is connected and installed so as to be slidably moved along the moving rail (120), and moves along the moving rail (120) to capture an image of an image analysis target (P) placed on a mounting table (110) and then transmits it to a media distribution unit (200).

[0047] An image capturing device (100) having the configuration described above can enable more precise and rapid image analysis by acquiring images from various angles of view of an image analysis target (P) that is the subject of image data analysis using AI according to the present invention.

[0049] Figure 3 is a drawing showing the mounting table of Figure 2.

[0050] Referring to FIG. 3, the mounting table (110) includes a table body (111), a rotating shelf (112), and an object cleaning unit (500).

[0051] A rotary drive motor (M) for rotating a rotary shelf (112) may be installed inside the table body (111).

[0052] The rotating shelf (112) is formed in the shape of a disc and is rotatably connected and installed on the top of the table body (111), and rotates when an image analysis target (P) is placed on it.

[0053] At this time, the rotating shelf (112) may be made of various sizes corresponding to the size or shape of the image analysis target (P) to be mounted, and a fastening device such as a jig for fastening the image analysis target (P) (not shown in the drawing for convenience of explanation) may be installed on the upper side to prevent movement during the rotation process of the image analysis target (P).

[0054] The object cleaning unit (500) is installed along the circumference of the rotating shelf (112) and is driven when the image analysis object (P) is placed on the rotating shelf (112) to remove foreign matter attached to the image analysis object (P).

[0055] A mounting table (110) having the configuration described above can not only stably mount and rotate an image analysis target (P) of various shapes, but also remove foreign substances such as dust attached to the image analysis target (P) using compressed air before full-scale image shooting, thereby enabling clearer image shooting.

[0057] Figures 4 and 5 are drawings showing the object cleaner of Figure 3.

[0058] Referring to FIGS. 4 and 5, the object cleaning unit (500) includes a plurality of support frames (510), a plurality of lifting drive gears (520), a cylindrical cover (530), and a plurality of air injection units (540).

[0059] A plurality of support frames (510) are connected in an interlocking manner so that they are erected around the perimeter of the rotating shelf (112) and can move up and down in the vertical direction, thereby supporting the cylindrical cover (530) and simultaneously moving the cylindrical cover (530) up and down in the vertical direction.

[0060] A plurality of lifting drive gears (520) are rotatably installed inside the rotating shelf (112) as shown in FIG. 6 and are driven to rotate in the forward or reverse direction by being geared with a gear tooth (G1) formed at the rear end of the support frame (510) to move the support frame (510) up and down.

[0061] The cylindrical cover (530) is formed in the shape of a cylindrical pipe and is supported by a plurality of support frames (510) installed on the inward surface, and moves up and down together with the support frames (510) as they move up and down.

[0062] A plurality of air injection units (540) are installed in a row along the inner surface of the cylindrical cover (530), and when an image analysis target (P) placed on the rotating shelf (112) is positioned inside the cylindrical cover (530), compressed air is injected in the direction of the image analysis target (P) to remove foreign substances such as dust.

[0063] The object cleaning unit (500) having the configuration described above can achieve clearer image capture by removing foreign substances, such as dust attached to the image analysis object (P), using compressed air before full-scale image capture.

[0065] Figure 6 is a drawing showing the support frame of Figure 4.

[0066] Referring to FIG. 6, the support frame (510) includes a frame body (511), a support slider (512), and a rotary drive gear (513).

[0067] The frame body (511) is installed to be upright around the perimeter of the rotating shelf (112) and to be able to move up and down in the vertical direction, and is connected by being engaged with the lifting drive gear (520) through gear coupling.

[0068] A plurality of support sliders (512) are installed at regular intervals along the front edge of a frame body (511) facing the inward surface of a cylindrical cover (530), and are capable of horizontal movement by engaging with a horizontal movement groove (531) formed extending horizontally along the inward surface of the cylindrical cover (530), and support the cylindrical cover (530).

[0069] The rotary drive gear (513) is rotatably installed on the support slider (512) and engages with the gear teeth formed on the inner side of the horizontal movement groove (531) through gear coupling, and is driven to rotate in the forward or reverse direction to rotate the cylindrical cover (530).

[0070] The support frame (510) having the configuration described above can not only stably support the cylindrical cover (530) and move it up and down in the vertical direction, but also rotate the cylindrical cover (530) during or after the movement is completed so that compressed air is sprayed from all directions by the air injection unit (540).

[0072] Figure 7 is a drawing showing the air injection part of Figure 6.

[0073] Referring to FIG. 7, the air injection unit (540) includes an injection port (541), an injection nozzle (542), and a variable discharge nozzle (543).

[0074] The nozzle (541) is formed by being recessed into the inner surface of the cylindrical cover (530) and is shaped like a cone with an inner diameter that gradually decreases towards the inside.

[0075] The injection nozzle (542) is installed inside the injection port (541) and injects compressed air through the injection port (541).

[0076] A variable discharge nozzle (543) is installed around the inlet of the nozzle (541) to induce the compressed air discharged through the nozzle (541) to spread or concentrate.

[0078] FIGS. 8 to 10 are drawings showing the variable discharge nozzle of FIG. 8.

[0079] Referring to FIGS. 8 to 10, the variable discharge nozzle (543) includes a wing mounting groove (5431), a first guide wing (5432), a first gap cover (5433), a second guide wing (5434), a second gap cover (5435), a first drive thread (5436), a second drive thread (5437), a wing drive gear (5438), and a wing support jaw (5439).

[0080] The wing seating groove (5431) is formed as a larger cone shape along the circumference of the inlet of the nozzle (541) and on an inclined surface corresponding to the inclined surface of the nozzle (541).

[0081] A plurality of first guide wings (5432) are spaced apart at regular intervals along the inner side of the wing mounting groove (5431).

[0082] The first gap cover (5433) is made of a thin film of an elastic material capable of expanding and contracting and is installed between a plurality of first guide wings (5432).

[0083] The second guide wing (5434) is connected and installed so as to be rotatable in the direction of the nozzle (541) at the front end of the first guide wing (5432), and a plurality of them are spaced apart at regular intervals along the wing mounting groove (5431).

[0084] At this time, the first guide wing (5432) and the second guide wing (5434) may be made of a material such as elastic plastic that can be partially unfolded or bent in correspondence with the shape of the wing seating groove (5431) as they are stored or exposed along the wing seating groove (5431).

[0085] Alternatively, the wing mounting groove (5431) may preferably provide sufficient space for the first guide wing (5432) and the second guide wing (5434) to move without resistance while being exposed or inserted after being mounted.

[0086] The second gap cover (5435) is made of a thin film of an elastic material capable of expanding and contracting and is installed between a plurality of second guide wings (5434).

[0087] The first driving screw thread (5436) is installed along the front-rear length direction along each inward surface of a plurality of first guide wings (5432).

[0088] A plurality of second drive threads (5437) are connected to the first drive threads (5436) and are installed along the front-rear length direction along each inward surface of a plurality of second guide wings (5434).

[0089] The wing drive gear (5438) is connected to the inward-facing surface of the wing mounting groove (5431) on the entrance side, facing each of the multiple second drive screw threads (5437), so as to enable rotational drive, and is connected by engaging with the second drive screw threads (5437) through gear coupling. As it is rotated in the forward direction, the second guide wing (5434) is exposed from the wing mounting groove (5431), and then the first guide wing (5432) is exposed from the wing mounting groove (5431) by engaging with the first drive screw thread (5436) through gear coupling, and as it is rotated in the reverse direction, the first guide wing (5432) and the second guide wing (5434) are sequentially stored into the wing mounting groove (5431).

[0090] The wing support jaw (5439) is installed on one side and the other side of the inward surface of the first guide wing (5432), respectively, and supports the second guide wing (5434), which is exposed from the wing seating groove (5431) and rotates due to the tension of the second gap cover (5435), so that it cannot rotate any further.

[0091] In one embodiment, the frame body (511) may have a wing movement groove (5111) formed along the contact surface with the cylindrical cover (530) where the air injection part (540) moves to move the first guide wing (5432) and the second guide wing (5434) when the cylindrical cover (530) rotates after the first guide wing (5432) and the second guide wing (5434) are exposed from the wing seating groove (5431) as shown in FIG. 5.

[0092] The variable discharge nozzle (543) having the configuration described above can be implemented to enable various applications corresponding to the distribution area of ​​foreign substances or the intensity of the discharged compressed air, etc., by varying its shape such as expanding the cone shape as it is exposed from the wing seating groove (5431) or gathering again after expansion.

[0094] The embodiments described above are for illustrative purposes only, and those skilled in the art will understand that the embodiments described above can be easily modified into other specific forms without altering the technical concept or essential features of the embodiments described above. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single unit may be implemented in a distributed manner, and components described as distributed may likewise be implemented in a combined form.

[0096] The scope of protection sought through this specification is defined by the claims set forth below rather than by the detailed description above, and should be interpreted to include all modifications or variations derived from the meaning and scope of the claims and the concept of equivalents. Explanation of the symbols

[0098] 10: Video data analysis system using AI 100: Video recording device 200: Media Distribution Department 300: Database 400: Image processing unit

Claims

Claim 1 A video capturing device that captures a video analysis target from multiple directions to acquire at least one video; a media distribution unit that transmits the video received from the video capturing device; and a video processing unit that uses a deep learning model with a GPU embedded in a processor to estimate metadata from the video transmitted through the media distribution unit, stores the estimated metadata in a database, and decodes the video received from the media distribution unit and processes it into an outputtable form; wherein the video capturing device comprises: a mounting table that rotates the video analysis target when the video analysis target is placed thereon; a movable rail that extends in a curved shape along the periphery of the video analysis target placed on the mounting table; and at least one shooting module that is connected and installed to enable sliding movement on the movable rail, moves along the movable rail to capture an image of the video analysis target placed on the mounting table, and then transmits it to the media distribution unit; wherein the mounting table comprises: a table body; a rotating shelf formed in the shape of a disc, connected and installed to the top of the table body so as to be rotatable, and rotated when the video analysis target is placed thereon. and includes an object cleaning unit installed along the circumference of the rotating shelf, which is driven when the image analysis target is placed on the rotating shelf to remove foreign substances attached to the image analysis target; wherein the object cleaning unit comprises: a plurality of support frames that are upright on the circumference of the rotating shelf and are connected in an interlocking manner to enable vertical movement; and a plurality of lifting drive gears that are rotatably installed on the inner side of the rotating shelf and are driven to rotate in a forward or reverse direction by engaging with gear teeth formed at the rear end of the support frames to move the support frames up and down.An image data analysis system using AI, comprising: a cylindrical cover formed in the shape of a cylindrical pipe and supported by a plurality of support frames installed on an inward surface, which moves up and down together as the support frames move up and down; and a plurality of air injection units installed along the inner surface of the cylindrical cover, which spray compressed air toward the image analysis target to remove foreign substances when the image analysis target seated on the rotating shelf is placed inside the cylindrical cover. Claim 2 delete Claim 3 delete Claim 4 delete Claim 5 delete Claim 6 delete Claim 7 delete

Images