Method of processing image obtained from imaging device interlocked with computing apparatus and system using the same

1 . A method of processing an image, performed by a computing apparatus including a processor, the method comprising:

obtaining an image;

obtaining, from the image, analysis information corresponding to an object included in the image by using an object analysis model; and

obtaining, from the analysis information corresponding to the object, a character contained in the object by using an OCR model,

wherein the OCR model performs:

determining whether the character is displayed on a surface of the object;

extracting at least one image sample from the object on which the character is displayed;

determining a boundary line of an area in which the character is displayed from the at least one image sample;

obtaining, from the at least one image sample, image patterns located on at least one of: (i) the boundary line; or (ii) a boundary point belonging to the boundary line, as a boundary marker;

generating a character-displayed image including the boundary marker as a partial image included in the image of the object on which the character is displayed, the image of the object being included in the image obtained in the obtaining an image step; and

performing OCR on the character-displayed image.

2 . The method of claim 1 , wherein the image sample is an image pattern present at a position of at least one of the boundary line or a boundary point belonging to the boundary line; and

the image pattern includes at least one of a partial character, a border of the character, a portion of the character, or a background.

3 . The method of claim 1 , wherein the boundary marker includes at least one of:

a start marker that is an image pattern corresponding to a start character of the character; or

an end marker that is an image pattern corresponding to an end character of the character.

4 . The method of claim 3 , wherein the obtaining of, from the at least one image sample, the image patterns located on at least one of the boundary line or a boundary point belonging to the boundary line, as the boundary marker includes:

obtaining a character image including the start marker and having a character recognition rate of at least a threshold value by using a tracking controller; and

determining whether the character image includes the end marker.

5 . The method of claim 4 , further comprising:

determining the character image as the character-displayed image when the character image includes the end marker.

6 . The method of claim 4 , further comprising:

when the character image does not include the end marker, obtaining an additional character image including a next marker of a last marker among the boundary marker included in the character image and having the character recognition rate of at least the threshold value by using the tracking controller;

merging the additional character image into the character image to generate a merged character image; and

determining the merged character image as the character-displayed image when the merged character image includes the end marker.

7 . The method of claim 1 , comprising:

after the determining whether the character is displayed on the surface of the object,

obtaining an image of the object including the start point of the character and having a character recognition rate of at least a threshold value by using a tracking controller;

performing the OCR on an initial sentence area of the image of the object; and

calculating a first value of meaningfulness that is a numerical value of a meaningfulness by performing Natural Language Understanding (NLU) on a first text that is a primary result of the OCR.

8 . The method of claim 7 , further comprising:

if the first value of meaningfulness is equal to or greater than the threshold value, determining the first text to be a result text that is a result of the OCR.

9 . The method of claim 7 , further comprising:

when the first value of meaningfulness is less than the threshold value, performing the OCR on a next sentence area of the initial sentence area; performing natural language understanding on a second text that is a primary result of the OCR, and calculating a second value of meaningfulness that is a numerical value of the meaningfulness; and

when the second value of meaningfulness is equal to or greater than the threshold value, determining the second text to be a result text that is a result of the OCR.

10 . The method of claim 1 , wherein the computing apparatus is interlocked with an imaging device and a gimbal, and

wherein obtaining the image comprises:

controlling a direction of the imaging device by actuating one or more rotational axis of the gimbal by using a tracking controller; and

obtaining the image by zooming in or zooming out of the imaging device by using the tracking controller.

11 . A non-transitory computer-readable medium including a computer program, the computer program causing a computing apparatus to perform a method of processing an image, the method comprising:

obtaining an image;

obtaining, from the image, analysis information corresponding to an object included in the image by using an object analysis model; and

obtaining, from the analysis information corresponding to the object, a character contained in the object by using an OCR model,

wherein the OCR model performs:

determining whether the character is displayed on a surface of the object;

extracting at least one image sample from the object on which the character is displayed;

determining a boundary line of an area in which the character is displayed from the at least one image sample;

obtaining, from the at least one image sample, image patterns located on at least one of: (i) the boundary line; or (ii) a boundary point belonging to the boundary line, as a boundary marker;

generating a character-displayed image including the boundary marker as a partial image included in the image of the object on which the character is displayed, the image of the object being included in the image obtained in the obtaining an image step; and

performing OCR on the character-displayed image.

12 . A computing apparatus, comprising:

a processor; and

a communication unit,

wherein the processor obtains an image,

obtains, from the image, analysis information corresponding to an object included in the image by using an object analysis model, and

obtains, from the analysis information corresponding to the object, a character contained in the object by using an OCR model,

wherein the OCR model determines whether the character is displayed on a surface of the object,

extracts at least one image sample from the object on which the character is displayed,

determines a boundary line of an area in which the character is displayed from the at least one image sample,

obtains, from the at least one image sample, image patterns located on at least one of: (i) the boundary line; or (ii) a boundary point belonging to the boundary line, as a boundary marker,

generates a character-displayed image including the boundary marker as a partial image included in the image of the object on which the character is displayed, the image of the object being included in the image obtained in the obtaining an image step, and

performs OCR on the character-displayed image.