KR101364686B1 - Method and apparatus for image reconstruction by digital hand scanner - Google Patents

Abstract

The present invention relates to a method and apparatus for reconstructing a split image using a hand scanner, and more particularly, to a hand scanner capable of reconstructing a plurality of split images scanned by a hand scanner through image estimation and reconstruction to an original single image. A split image reconstruction method and apparatus are used.
In the method of reconstructing a split image using a hand scanner of the present invention, in a method of reconstructing a split image scanned through a hand scanner into a single image, the scanning speed is determined and the single image is sequentially divided according to the determined scanning speed. Scanning the data to obtain divided images such that overlapping portions are formed between adjacent divided images (S210); Setting a virtual reference compartment obtained by dividing the scanned divided image by the same number in the horizontal and vertical directions (S220); Calculating a duplicate image area by comparing a reference section of the selected divided image with a reference section of another adjacent divided image (S230); Calculating and storing an image damage value between adjacent divided images through the calculated overlapping image area (S240); Calculating a correction value based on an average value of stored image damage values (S250); Generating an integrated single image by applying the correction values between the divided images (S260); .

Description

Method and device for reconstructing split image using hand scanner {METHOD AND APPARATUS FOR IMAGE RECONSTRUCTION BY DIGITAL HAND SCANNER}

The present invention relates to a method and apparatus for reconstructing a split image using a hand scanner, and more particularly, to a hand scanner capable of reconstructing a plurality of split images scanned by a hand scanner through image estimation and reconstruction to an original single image. A split image reconstruction method and apparatus are used.

Recently, due to the development of a handheld device, an image reconstruction method through a handheld device has been used and developed in many fields. In particular, a handheld device, which is a representative handheld device, is a device that has undergone many technological advancements due to its convenience in use, and in particular, medical devices, smart phones, and other devices have been continuously developed.

In general, a process of reconstructing a single image by partially scanning an area larger or longer than the area that can be scanned using a hand scanner includes acquiring a partial image and finding an optimal reconstruction point by overlapping the partial images. And an algorithm according to the step of reconstructing the image based on the point.

To this end, Korean Patent Publication No. 10-2011-0037143, published on April 13, 2011, has been optimized through measurement means for configuring reference data for image merging by measuring at least one measurement item among image pixel size and URL. A method of determining a merge target image combination is presented.

However, the image restoration method proposed by the above-described patent does not provide any compensation method for the horizontal and rotational directions of the image sensor, so there is a problem that accurate restoration cannot be achieved.

The present invention has been made to solve the above problems, and an object of the present invention is to divide the target image into a plurality of areas to calculate the optimal connection point of each divided area and the horizontal of the distorted target image based on the connection point. And a split image reconstruction method and apparatus using a hand scanner capable of reconstructing an accurate target image by compensating vertical shift values and rotation angles.

In the split image reconstruction method using the hand scanner of the present invention for achieving the above object, in the method of reconstructing the divided image scanned through the hand scanner into a single image, the scanning speed is determined, Sequentially dividing and scanning a single image, wherein the divided images are obtained such that overlapping portions are formed between adjacent divided images; Setting a virtual reference section in which the scanned divided image is divided into equal numbers in the horizontal and vertical directions; Calculating a duplicate image area by comparing a reference section of the selected divided image with a reference section of another adjacent divided image; Calculating and storing an image damage value between adjacent divided images through the calculated overlapping image area; Calculating a correction value based on an average value of stored image damage values; Generating the combined single image by applying the correction values between the divided images; Characterized in that consists of.

The image damage value may be calculated by comparing the overlapping image areas between adjacent divided images, by using a horizontal shift value, a vertical shift value, and a rotation angle between the divided images.

,

,

를 갖도록 하여 갖도록 하여 분할이미지 간 30% 이상이 중복이 이루어지도록 구성될 수 있다.Duplicate parts of each split image have coordinate parameters

,

,

It can be configured to have more than 30% overlap between the divided images to have.

을 이용하여 조정될 수 있다.([△x]_opt: 수평이동치 한계값, [△y]_opt: 수직이동치 한계값, [△α]_opt: 회전각 한계값, △x: 수평이동치, △y: 수직이동치: △α: 회전각 변동치, M: 분할이미지의 세로길이, N: 분할이미지의 가로길이)The scan speed is a speed conversion formula to which a parameter limit value is applied.

([Δx] _opt : horizontal shift limit, [△ y] _opt : vertical shift limit, [Δα] _opt : rotation angle limit, Δx: horizontal shift, Δy: vertical shift value: Δα: rotation angle variation value, M: vertical length of split image, N: horizontal length of split image)

The split image reconstruction apparatus using the hand scanner of the present invention for achieving the above object is a device for reconstructing a single image by recombining the divided image scanned through the hand scanner, the single image is scanned by dividing at a specified speed An image input unit which acquires a divided image so that overlapping portions are formed between adjacent divided images; A reference compartment setting unit for setting a virtual reference compartment obtained by dividing the obtained divided image by the same number in the horizontal and vertical directions; A duplicate image area calculation unit for comparing the reference section of the selected divided image with a reference section of another adjacent image to calculate a duplicate image area; An image damage value calculation unit for calculating an image damage value between adjacent divided images through the calculated overlapping image area; A memory unit for temporarily storing the calculated image damage value; A correction value calculation unit that reads an image damage value from the memory unit and calculates a correction value corresponding to an average value; An image restoring unit generating the integrated single image by applying the correction value between the divided images; And a control unit.

The image corruption value calculating unit compares overlapping image areas between the divided images and calculates a horizontal shift value between the divided images, a vertical shift value measuring unit calculating a vertical shift value between the divided images, and between the divided images. It may include a rotation angle measuring unit for calculating the rotation angle of.

,

,

를 갖도록 하여 분할이미지 간 30% 이상이 중복이 이루어지도록 스캔하는 것이 바람직하다.The image input unit has a limit value of a coordinate parameter of the overlapped portion of each divided image.

,

,

It is desirable to scan so that 30% or more overlap between divided images.

을 이용하여 조정되는 것이 바람직하다.([△x]_opt: 수평이동치 한계값, [△y]_opt: 수직이동치 한계값, [△α]_opt: 회전각 한계값, △x: 수평이동치, △y: 수직이동치: △α: 회전각 변동치, M: 분할이미지의 세로길이, N: 분할이미지의 가로길이)The image input unit is a speed conversion type to which the scanning speed is applied a threshold value of the parameter

([Δx] _opt : horizontal shift limit, [Δy] _opt : vertical shift limit, [Δα] _opt : rotation angle limit, Δx: horizontal) Equivalent value, Δy: vertical shift value: Δα: rotation angle variation value, M: vertical length of split image, N: horizontal length of split image)

The present invention estimates and corrects a segmented scanned image using a method and apparatus for reconstructing a split image using a hand scanner, and accurately compensates for the original single image by performing horizontal, vertical shift and rotation in overlapping image regions between the divided images. It can be restored and applied to various imaging apparatuses (image diagnosis apparatuses, medical apparatuses, biometric apparatuses, etc.) to improve the accuracy of image processing very high.

1 is a flowchart showing a method of reconstructing a split image using the hand scanner of the present invention;
2 is a block diagram showing the configuration of a split image reconstruction apparatus using the hand scanner of the present invention.

Hereinafter, a method of reconstructing a split image using the present invention will be described in detail with reference to the accompanying drawings.

1 is a flowchart illustrating a method of reconstructing a split image using a hand scanner according to the present invention, and shows a method of performing a split scan of a single large image through a hand scanner and collecting the scanned split image into a single image.

First, in the first step (S210), the scan speed is determined, and the single image is sequentially divided and scanned according to the determined scan speed, but the divided images are acquired to form overlapping portions between adjacent divided images.

The single image refers to a large image of a size that cannot be scanned at a time through the hand scanner, and partially scans the image using the hand scanner to obtain a segmented image. At this time, the hand scanner moves at the same speed in a constant direction on the scan target image so that the scanned divided image has continuity.

That is, the image generated when scanning the image using the hand scanner is affected by the hand scanner movement characteristics such as the moving direction and the speed of the hand scanner moving on the target image. In the first step (S210), the image pieces are sequentially obtained. The images are divided and scanned so that they can be combined, and coordinate variables (x, y, moving coordinates) for each image are generated as a set.

In the present invention, the set of coordinate variables provides the exact position of the image fragment when each segmented image is combined into a single whole image. In addition, the split image reconstruction method using the hand scanner of the present invention obtains additional information on the scanned image combination such as the scan start and end positions, the loss state of synchronization, and the adaptive scan rate by using the amount of movement of each image fragment. Calculate

For this purpose, (x, y, α) is used for the local coordinate system between two neighboring split images in the scanned divided image, and (X, Y, A) is used for the global coordinate system of the entire image object. The coordinate difference of the coordinate system is represented by (Δx, Δy, Δα). (Δx: horizontal shift value, Δy: vertical shift value, Δα: rotation angle)

In addition, overlapped portions should be formed between adjacent divided images in the scanned divided image. If no overlapping portion is formed between adjacent divided images, and if any portion has not been scanned, it cannot be reconstructed into a single image in the final step.

In addition to scanning a planar image in the design field, the task of acquiring a split image through the hand scanner is to obtain an image by scanning a body part in a medical field, or to implement a panoramic picture through a partial image. It is not limited and can be applied broadly.

In the second step S220, a virtual reference compartment in which the scanned divided image is divided into the same number horizontally and vertically is set. In other words, the reference section is a virtual partition that helps to compare the adjacent divided images, and the more detailed the comparison operation can be done as it is divided into many parts, but it may cause the processing speed to be reduced in consideration of the performance of the system. Set to set the number.

In a third step S230, the overlapping image area is calculated by comparing the reference section of the selected divided image with the reference section of another adjacent divided image. Since the overlapped regions exist between the adjacent divided images obtained in the first step S210, the overlapped image areas are calculated by comparing the reference sections set in the second step S220.

In the fourth step S240, an image damage value between adjacent divided images is calculated and stored through the calculated overlapping image area.

The image corruption value represents a degree of damage between adjacent divided images and means a shift value of relative coordinates between the divided images. In detail, the image damage value may be calculated by comparing the overlapping image areas between adjacent divided images, and using the horizontal shift value, the vertical shift value, and the rotation angle between the divided images.

Subsequently, in a fifth step S250, the correction value is calculated based on the average value of the stored image damage values. Even if the scanning operation is performed at the same speed as the predetermined step in the first step (S210), the image damage value between each adjacent divided images has a slight difference, so that the average value of the image damage value is applied as an integrated correction value. .

In the final sixth step (S260), the correction value is applied between the divided images to combine all adjacent divided images and generate an integrated single image. That is, the horizontal shift value, the vertical shift value, and the rotation angle between adjacent split images generated during the split scan operation are corrected to combine the adjacent split images into a single image.

To reiterate, first, in order to compare a plurality of fragmented images, overlapping image areas overlapping each other must exist between adjacent divided images, so that the overlapped images with the speed change of the hand scanner so that the divided images overlap each other. Adjust the area. The detailed formula is as follows.

As an example, the fragment image fragments IMG _j and IMG _{j +1} are scanned at the time t _j , t _{j +1} at the speed of V _j .

Reference segment of neighboring segmented image segment (IMG _{j +1} ) based on the reference segment of the original segmented image segment (IMG _j ) to calculate the image damage between the scanned segmented image segment IMG _j , IMG _{j +1} The optimal overlapping image area in the area is detected.

In addition, the image damage value between the scanned fragment image fragments IMG _j and IMG _{j +1} is calculated, specifically, due to rotation or distortion from the original fragment image fragment IMG _j to the neighboring fragment image fragment IMG _{j +1} . The split images are combined by estimating the horizontal shift value, the vertical shift value, the rotation angle, and the like and by applying a correction value corresponding to the average value of the image damage value. That is, based on the best matching each point by calculating the △ x and △ y _{ji ji, i} of the linear approximation △ x IMG _{j +1} by the (linear approximation), △ y _i, α _i is obtained by △.

Three reference parameters (horizontal shift value, vertical shift value, and rotation angle) for the shift amount of the split image are presented between the reference divisions of adjacent split images having overlapping image areas. The calculated parameter set in the local coordinate system is compared with each fragment of the fragmented image and then combined in the overall coordinate system to obtain one completed image.

That is, by using the _{△ x i, △ y i,} △ α i of the slice IMG _{j +1} (world coordinates), obtain the global coordinates of the whole image buffers buffered to the appropriate position, and store the entire image when the buffering is completed.

In this case, if the smooth operation is not performed, the scan rate needs to be adjusted according to the speed change of the handscanner so that a certain area of each of the fragmented images may be overlapped continuously. One must be spaced apart by a predefined limit.

[Equation 1]

,

,

,

,

은 상기 [수학식 1]로부터 유추하여 산출할 수 있다.The limit value of each parameter described in [Equation 1] is to allow about 30% of the image pieces to overlap each other. At this time, the scan rate of the hand scanner

Can be calculated by inferring from [Equation 1].

이고,

이면,

의 y값 변위(

)를 산출할 수 있다. 마찬가지로 x값과 a값을 동일하게 산출하면 다음과 같은 스캔율이 완성된다.That is, assuming that Δy _{i +1} converges to M / 3,

ego,

If so,

Y value displacement of

) Can be calculated. Similarly, calculating the same x and a values results in the following scan rates.

&Quot; (2) "

Equation 2 is applied to a step of scanning an image to adjust a scan rate.

Hereinafter, with reference to the accompanying drawings will be described in detail a split image reconstruction apparatus using the present invention hand scanner.

FIG. 2 is a block diagram showing the configuration of a split image reconstruction apparatus using the hand scanner of the present invention. As shown in FIG. 2, the split image reconstruction apparatus using the hand scanner of the present invention includes an image input unit 110 and a reference section setting unit 120. ), A duplicate image area calculation unit 130, an image corruption value calculation unit 140, a memory unit 150, a correction value calculation unit 160, and an image restoration unit 170.

The image input unit 110 performs a function of scanning an image as an image input sensor of a hand scanner. In particular, the present invention divides and scans a single image at a specified speed, but obtains a divided image so that a overlapped portion is formed between adjacent divided images.

The single image refers to a large image of a size that cannot be scanned at a time through the hand scanner, and partially scans the single image through the image input unit 110, but the image input unit is the same speed in a constant direction on the scan target image. The image is moved so that the scanned divided image has continuity.

That is, the image generated during the scan is affected by the movement characteristics such as the moving direction and the speed of the moving image input unit 110, so that the image input unit 110 can combine the image pieces obtained sequentially. The scan is divided and a coordinate variable (x, y moving coordinate) for each image is generated as a set.

In the present invention, the set of coordinates provides the exact position of the image fragment when each split image is combined into a single whole image.

In addition, (x, y, α) is used for the local coordinate system between two neighboring divided images in the scanned divided image, and (X, Y, A) is used for the global coordinate system of the entire image object, and the local coordinate system The coordinate difference of is represented by (Δx, Δy, Δα). (Δx: horizontal shift value, Δy: vertical shift value, Δα: rotation angle)

In addition, an overlapped portion should be formed between adjacent divided images in the scanned divided image. If no overlapping portion is formed between adjacent divided images, and if any portion has not been scanned, it cannot be reconstructed into a single image in the final step.

The reference compartment setting unit 120 performs a function of setting a virtual reference compartment obtained by dividing the obtained divided image by the same number horizontally and vertically. The reference section is a virtual section that helps to compare the adjacent divided images, and when divided into many parts, a more precise comparison can be made, but it may cause a decrease in processing speed. Set to set.

The duplicate image area calculation unit 130 calculates a duplicate image area by comparing a reference section of the selected divided image with a reference section of another adjacent divided image. Since the overlapped regions exist between the adjacent divided images obtained by the image input unit, the overlapped reference regions set by the reference division setting unit are compared to calculate the overlapped image regions.

The image damage value calculator 140 calculates an image damage value between adjacent divided images through the calculated overlapping image area. The image corruption value represents a degree of damage between adjacent divided images and means a shift value of relative coordinates between the divided images. In detail, the image corruption value includes a horizontal shift value measuring unit 141 for comparing the overlapping image areas between adjacent divided images and calculating a horizontal shift value between the divided images, and a vertical shift value measuring unit for calculating the vertical shift values between the divided images ( 142 and a rotation angle measuring unit 143 calculating a rotation angle between the divided images.

The memory unit 150 temporarily stores the image damage value calculated by the image damage tooth calculation unit 140 as a means for temporarily storing data.

The correction value calculation unit reads an image damage value from the memory unit 150 and calculates a correction value corresponding to an average value. Even if the scanning operation is performed at the same speed as the predetermined direction through the image input unit 110, the image damage value between the adjacent divided images has a minute difference, so that the average value of the image damage value can be applied as a correction value that can be integrated. do.

The image restoring unit 170 performs a function of generating an integrated single image by applying the correction value between the divided images. That is, the horizontal shift value, the vertical shift value, and the rotation angle between adjacent split images generated during the split scan operation are corrected to combine the adjacent split images into a single image.

To reiterate, first, in order to compare a plurality of fragmented images, overlapping image areas overlapping each other must exist between adjacent divided images, so that the overlapped images with the speed change of the hand scanner so that the divided images overlap each other. Adjust the area. The detailed formula is as follows.

As an example, the fragment image fragments IMG _j and IMG _{j +1} are scanned at the time t _j , t _{j +1} at the speed of V _j .

Reference segment of neighboring segmented image segment (IMG _{j +1} ) based on the reference segment of the original segmented image segment (IMG _j ) to calculate the image damage between the scanned segmented image segment IMG _j , IMG _{j +1} The optimal overlapping image area in the area is detected.

In addition, the image damage value between the scanned fragment image fragments IMG _j and IMG _{j +1} is calculated, specifically, due to rotation or distortion from the original fragment image fragment IMG _j to the neighboring fragment image fragment IMG _{j +1} . The split images are combined by estimating the horizontal shift value, the vertical shift value, the rotation angle, and the like and by applying a correction value corresponding to the average value of the image damage value. That is, based on the best matching each point by calculating the △ x and △ y _{ji ji, i} of the linear approximation △ x IMG _{j +1} by the (linear approximation), △ y _i, α _i is obtained by △.

Between the reference divisions of adjacent divided images having overlapping image areas, three parameters (tangential, parallel shift amount, rotation angle) for the shift amount of the divided images are presented. The calculated parameter set in the local coordinate system is compared with each fragment of the fragmented image and then combined in the overall coordinate system to obtain one completed image.

That is, by using the _{△ x i, △ y i,} △ α i of the slice IMG _{j +1} (world coordinates), obtain the global coordinates of the whole image buffers buffered to the appropriate position, and store the entire image when the buffering is completed.

In this case, if the smooth operation is not performed, the scan rate needs to be adjusted according to the speed change of the handscanner so that a certain area of each of the fragmented images may be overlapped continuously. One must be spaced apart by a predefined limit.

[Equation 1]

,

,

,

,

은 상기 [수학식 1]로부터 유추하여 산출할 수 있다.The limit value of each parameter described in [Equation 1] is to allow about 30% of the image pieces to overlap each other. At this time, the scan rate of the hand scanner

Can be calculated by inferring from [Equation 1].

이고,

이면,

의 y값 변위(

)를 산출할 수 있다. 마찬가지로 x값과 a값을 동일하게 산출하면 다음과 같은 스캔율이 완성된다.That is, assuming that Δy _{i +1} converges to M / 3,

ego,

If so,

Y value displacement of

) Can be calculated. Similarly, calculating the same x and a values results in the following scan rates.

&Quot; (2) "

Equation 2 is applied to a step of scanning an image to adjust a scan rate.

It is to be understood that the invention is not limited to the disclosed embodiment, but is capable of many modifications and variations within the scope of the appended claims. It is self-evident.

110: image input unit 120: reference compartment setting unit
130: duplicate image area calculation unit 140: image corruption calculation unit
141: horizontal shift measurement unit 142: vertical shift measurement unit
143: rotation angle measurement unit 150: memory unit
160: correction value calculation unit 170: image restoration unit

Claims (8)

In the method of reconstructing a single image by collecting the divided image scanned by the hand scanner,
Determining a scanning speed, and scanning a single image by sequentially dividing the scan according to the determined scanning speed, obtaining the divided images so that overlapping portions are formed between adjacent divided images, and generating coordinate variables for each divided image ( S210);
Setting a virtual reference compartment obtained by dividing the scanned divided image by the same number in the horizontal and vertical directions (S220);
Calculating a duplicate image area by comparing a reference section of the selected divided image with a reference section of another adjacent divided image (S230);
Calculating and storing an image damage value between adjacent divided images through the calculated overlapping image area (S240);
Calculating a correction value based on an average value of stored image damage values (S250);
Generating an integrated single image by applying the correction values between the divided images (S260); Lt; / RTI >
Duplicate parts of each split image have coordinate parameters

,

,

Over 30% overlap between split images,
The scan speed is a speed conversion formula to which a parameter limit value is applied.

Split image reconstruction method using a hand scanner, characterized in that the adjustment using. ([△ x] _opt : horizontal shift value limit value, [△ y] _opt : vertical shift value limit value, [△ α]] _opt : rotation angle limit value, △ x: horizontal shift value, △ y: vertical shift value: Δα: rotation angle variation value, M: vertical length of split image, N: horizontal length of split image)
The method of claim 1,
The image damage value is calculated by calculating a horizontal shift value, a vertical shift value, and a rotation angle between the divided images by comparing overlapping image regions between adjacent divided images.
delete delete In the device for reconstructing a single image by collecting the divided image scanned through the hand scanner,
An image input unit 110 which scans by dividing a single image at a predetermined speed, but obtains a divided image so that a overlapped portion is formed between adjacent divided images, and generates coordinate variables for each divided image;
A reference compartment setting unit 120 for setting a virtual reference compartment obtained by dividing the obtained divided image by the same number in the horizontal and vertical directions;
A duplicate image area calculation unit 130 for comparing the reference section of the selected divided image with the reference section of another adjacent image to calculate a duplicate image area;
An image damage value calculation unit 140 for calculating an image damage value between adjacent divided images through the calculated overlap image area;
A memory unit 150 in which the calculated image damage value is temporarily stored;
A correction value calculation unit 160 for reading an image damage value from the memory unit 150 and calculating a correction value corresponding to an average value;
An image restoring unit (170) for generating an integrated single image by applying the correction value between the divided images; , ≪ / RTI >
The image input unit 110 is a limit value of the coordinate parameter of the overlapping portion of each divided image

,

,

To have more than 30% overlap between split images,
Scan rate is speed conversion with parameter limit

Split image reconstruction apparatus using a hand scanner, characterized in that the adjustment using. ([△ x] _opt : horizontal shift value limit value, [△ y] _opt : vertical shift value limit value, [△ α]] _opt : rotation angle limit value, △ x: horizontal shift value, △ y: vertical shift value: Δα: rotation angle variation value, M: vertical length of split image, N: horizontal length of split image)
The method of claim 5,
The image corruption value calculation unit 140 compares overlapping image areas between the divided images to calculate a horizontal shift value between the divided images, and a vertical shift value to calculate the vertical shift values between the divided images. And a rotation angle measurement unit (143) for calculating a rotation angle between the divided images.
delete delete

Images