CN104976951A - Device and method for identifying image - Google Patents

Abstract

The invention provides a device and a method for identifying an image. The method has the following steps. First, a first image carrying an object is captured by an image recognition device. Then, the length of the first diagonal line and the length of the second diagonal line of the object are calculated according to the first image. And outputting the identification result when the ratio of the first diagonal length to the second diagonal length is within a first range. The method can help a user to know the shape of the object and the size of the object by capturing the image of the object.

Description

Image recognition device and method

technical field

The present invention relates to an image recognition device and method thereof, in particular to an image recognition device and method for recognizing the shape and size of objects in the image.

Background technique

In the assembly line, when assembling two objects, measuring the size of the object and identifying the shape of the object is an indispensable part of the assembly process. For example, when assembling the glass into the aluminum frame, if the joint surface is sharply protruded due to excessive extrusion, or the sharp opening is formed due to incomplete sealing, it may cause injury to the packaging personnel or the personnel assembled by the end customer . In another example, if the shape of the module changes, when it is combined with the back rail at the end customer, in a slight case, it is barely assembled, resulting in residual stress at the locking point, which may cause more serious deformation or functional failure after a period of time ; Severe deformation and even chain locks cannot be carried out, causing losses to customers and manufacturers themselves.

However, there are still many problems in the current manual measurement of the size of the object and visual recognition of the shape of the object. For example, the measurer may misread the tape scale or the shape of the object itself is not easy to measure (for example, the corners of the object are arc-shaped) when measuring the object, which may cause the identification of the object to be unable to assemble due to the above errors . Moreover, if each object on the assembly line needs to be measured manually, it will be difficult to improve the efficiency of the assembly line.

Contents of the invention

In view of the above problems, the purpose of the present invention is to provide a device and method for identifying images, by comparing the number of pixels of the object in the captured image to determine the shape of the object, and by measuring the actual size of the object and the image The number of pixels in the object is obtained to obtain the size of the object, so as to help the user know the shape and size of the object only by capturing the image of the object.

The image recognition method disclosed in the present invention includes the following steps: firstly, the image recognition device is used to capture the first image of the object. Then, according to the first image, the length of the first diagonal line and the length of the second diagonal line of the object are calculated. And, when the ratio of the length of the first diagonal line to the length of the second diagonal line is within the first range, the identification result is output.

The image recognition device disclosed in the present invention has an image capture module and an image processing module. The image capturing module is used for capturing images with objects. The image processing module is coupled to the image capture module, the image processing module executes an edge detection program to obtain the positions of multiple edges of the object from the image, and the image processing module obtains the position of the object in the image according to the positions of the multiple edges of the object The positions of the first diagonal and the second diagonal among the plurality of diagonals, calculating the number of pixels on the positions of the first diagonal and the second diagonal, and judging the first diagonal and the second Whether the ratio of the number of diagonal pixels is within the first range. Wherein when the ratio of the number of pixels of the first diagonal line and the second diagonal line is within the first range, an identification result is output.

In summary, the device and method for identifying objects in an image of the present invention can obtain multiple edges of an object in an image by executing an edge detection program, and obtain multiple diagonals of an object through the multiple edges of an object, and finally Then, the shape of the object is judged by comparing the number of pixels on multiple diagonal positions of the object in the image, so that the user can know the shape of the object by capturing the image of the object.

The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments, but not as a limitation of the present invention.

Description of drawings

FIG. 1 is a functional block diagram of an image recognition device according to an embodiment of the present invention;

FIG. 2 is a side view of capturing an image of an object according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of an image with an object according to an embodiment of the present invention;

FIG. 4 is a schematic diagram of a partial image of an object according to an embodiment of the present invention;

FIG. 5 is a functional block diagram of image recognition by an apparatus for correcting and recognizing images according to an embodiment of the present invention;

FIG. 6 is a flowchart of a method for recognizing an image according to an embodiment of the present invention;

FIG. 7 is a flowchart of a method for calibrating an image recognition device according to an embodiment of the present invention.

Among them, reference signs

10 Device for identifying images

102 Image capture module

104 Image processing module

12 objects

14 video

1412 pixels

1414 pixels

1416 pixels

1418 pixels

1420 pixels

1422 pixels

1424 pixels

1426 pixels

1428 pixels

16 platform

18 standard test pieces

L1, L2, L3, L4 Sidelines

DI1, DI2 Diagonal

Detailed ways

The detailed features and advantages of the present invention are described in detail below in the embodiments, the content of which is sufficient to enable any person familiar with the relevant art to understand the technical content of the present invention and implement it accordingly, and according to the content disclosed in this specification, the scope of claims and the accompanying drawings , anyone skilled in the relevant art can easily understand the related objects and advantages of the present invention. The following examples are to further describe the viewpoints of the present invention in detail, but not to limit the scope of the present invention in any way.

Please refer to FIG. 1 . FIG. 1 is a functional block diagram of an image recognition device according to an embodiment of the present invention. As shown in FIG. 1 , the device 10 for identifying objects in an image includes an image capture module 102 and an image processing module 104 , wherein the image processing module 104 is coupled to the image capture module 102 .

Please refer to FIG. 1 , FIG. 2 and FIG. 3 together, wherein FIG. 2 is a side view of capturing an image of an object according to an embodiment of the present invention, and FIG. 3 is a schematic diagram of an image with an object according to an embodiment of the present invention. As shown in the figure, the image capture module 102 is used to capture an image 14 of an object 12 . Furthermore, when the user wants to identify the object 12, the user can first place the object 12 in parallel on a platform 16, and the platform 16 is placed under the object device 10 in the identification image, and then the user controls the image capture. The capture module 102 is used to capture the image 14 having the object 12 therein. In practice, the image capture module 102 can be, but not limited to, a camera lens, a video camera lens, a network camera lens, or other suitable photographic equipment capable of capturing the image 14 .

Next, the aforementioned image processing module 104 executes the edge detection program to obtain the positions of multiple edges of the object 12 from the image 14, and the image processing module 104 obtains the positions of the multiple edges L1, L2, L3, and L4 of the object 12 in the image. The positions of the diagonal DI1 and the diagonal DI2 among the plurality of diagonals of the object 12 in 14 . For example, after the image capturing module 102 captures the image 14 with the object 12 (such as a rectangle), it sends the image 14 to the image processing module 104 to perform an edge detection procedure. The operation of the edge detection procedure will be described in detail later. After the image processing module 104 finds the positions of the edges L1, L2, L3, and L4 of the object 12 in the image 14 through the above-mentioned edge detection program, it passes through the intersection positions of the edge L1, the edge L2, and the edge L4, and the edge L3 and the edge Find the position of the intersection of L2 and sideline L4, and find out the positions of diagonal line DI1 and diagonal line DI2 in image 14 . The image processing module 104 may be, but not limited to, a micro processor (micro processor), an image processing unit (graphics processing unit), a central processing unit (central process unit) or other elements suitable for computing processing, but the present invention does not refer to This is the limit.

The image processing module 104 calculates the number of pixels on the diagonal line DI1 and the diagonal line DI2 , and determines whether the ratio of the number of pixels on the diagonal line DI1 and the diagonal line DI2 is within a first range. When the image processing module 104 determines that the ratio of the number of pixels on the diagonal line DI1 to the number of pixels on the diagonal line DI2 is within the first range, it outputs an identification result. Furthermore, after the image processing module 104 obtains the positions of the diagonal line DI1 and the diagonal line DI2, it calculates the number of pixels of the diagonal line DI1 and the diagonal line along the diagonal line DI1 and the diagonal line DI2 respectively. The number of pixels of DI2.

Next, after calculating the ratio of the number of pixels on the diagonal line DI1 to the number of pixels on the diagonal line DI2, divide the number of pixels on the diagonal line DI1 by the number of pixels on the diagonal line DI2 to obtain a ratio . If the obtained ratio value is within the first range (for example, the range of 0.9 to 1.1), it means that the number of pixels of the diagonal line DI1 and the diagonal line DI2 are equal or the difference is within an acceptable range, which also means that the diagonal line The actual lengths of DI1 and diagonal DI2 are equal or the difference is within an acceptable range. At this time, a recognition result is output. This recognition result can be to judge that the recognized object 12 is a rectangle, or judge that the recognized object 12 is a five-dimensional Polygon, hexagon or any polygonal shape, the present invention is not limited thereto.

Please refer to FIG. 4 . FIG. 4 is a schematic diagram of a partial image of an object according to an embodiment of the present invention. As shown in FIG. 4 , the above-mentioned edge detection program is to compare whether the difference between the gray scale value of the pixel 1420 of the image 14 and other pixels 1412, 1418, 1420 is greater than the threshold value. , 1416, ... 1428 when the gray scale value difference is greater than the threshold value, the position of the pixel 1420 in the image 14 is the position of the pixels of the multiple edge lines L1, L2, L3, L4 of the object 12, wherein the pixel 1420 is adjacent to pixel 142 . In other words, when the user places the object 12 in parallel under the image capture module 102, the user can place a light source under the object 12, and the pixel position placed by the object 12 will be partially covered by the object 12 itself. Light, and the brightness is lower than other pixel positions where the object 12 is not placed, and the pixel brightness at the edge position where the object 12 is placed will have a larger difference in brightness than the brightness of the pixel positions outside the adjacent edge. The difference between the grayscale values of each pixel in the image 14 and its adjacent pixels is compared.

An example is used here to illustrate that after the image capture module 102 captures the image 14 of an object 12 (such as a pentagon), the image 14 with the object 12 is sent to the image processing module 104 . The image processing module 104 calculates the grayscale value of each pixel in the image 14, and starts to calculate the two grayscale values with other adjacent pixels 1412, 1414, 1416, ... 1428 centered on the pixel 1420. Whether the difference is greater than a threshold (for example, greater than 125). For example, the grayscale value of pixel 1420 is 255, and the grayscale values of pixels 1412, 1418, and 1424 adjacent to the left of pixel 1420 are 10, 20, and 50 respectively, and the pixels adjacent to the right of pixel 1420 The grayscale values of 1416, 1422, and 1428 are 168, 175, and 192 respectively, and the grayscale values of the pixel 1414 adjacent to the pixel 1420 above and the pixel 1426 below the pixel 1420 are 245 and 252, respectively. When the grayscale value of pixel 1420 is subtracted from the grayscale values of pixels 1412, 1418, and 1424, the differences are all greater than the threshold value. However, the gray scale value of the pixel 1420 is compared with the gray scale values of the pixels 1416, 1422, 1428 and the gray scale values of the pixels 1414, 1426 respectively, and they are all smaller than the threshold value. At this time, the image processing module 104 can determine that the pixel 1420 is the left line of the object 12 .

Not only that, in order to make the grayscale value of the pixel on the sideline of the present invention more distinct from the grayscale value of adjacent pixels outside the sideline, the number of light sources in the present invention can be multiple, and the number of these light sources The position can not only be placed under the object 12, but also can be placed in multiple different positions according to the needs of the user's situation, so that the edge of the object 12 of the present invention can be obtained more accurately.

In addition, the image processing module 104 is further capable of receiving a plurality of measured values, wherein each measured value corresponds to one of the actual lengths of a plurality of sidelines L1, L2, L3, L4 or a plurality of diagonal lines DI1, DI2 , and the image processing module 104 further executes the edge detection program to obtain the positions of the multiple edges L1, L2, L3, and L4 of the object 12 from the image, according to the received measurement values and the corresponding edges L1, L2 , L3 , L4 or a relational expression of the number of pixels at the positions of these diagonal lines DI1 , DI2 to calculate the size of the pixel 142 . In one embodiment, the user can measure the actual lengths of the edges L1 , L2 , L3 , L4 and the two diagonals DI1 , DI2 of the object 12 in advance, and input the measured values into the device 10 for recognizing objects in images. The image processing module 104 can calculate the pixel size of the image based on the actual lengths of the edges L1 , L2 , L3 , L4 and the two diagonals DI1 , DI2 and the number of pixels at their positions obtained from the aforementioned calculation.

In another embodiment, when the object 12 is a pentagon, the user can pre-measure the actual lengths of the five sides and three diagonals of the pentagon before capturing the image of the pentagon, And input the above measurement value into the device 10 for identifying objects in the image. The image processing module 104 then obtains the positions of the five sides of the pentagon in the image according to the aforementioned edge detection procedure, and then obtains the positions of the three diagonals of the pentagon in the image according to the positions of the five sides. Finally, the image processing module 104 can calculate the number of pixels in the image 14 according to the input side length and the actual length of the diagonal line, and the number of pixels at the position of the side length and the diagonal line in the image. size.

The aforementioned relationship can be:

Wherein, the total ratio is the addition of each measured value to the ratio of the number of pixels corresponding to these sidelines L1, L2, L3, L4 or corresponding to one of these diagonal lines DI1, DI2, and the number of measurements is The number of multiple measurements received. For example, after measuring the actual lengths of the sidelines L1, L2, L3, L4 and the two diagonal lines DI1, DI2, the image processing module 104 divides the length of the sideline L1 by the number of pixels on the sideline L1 to obtain the sideline L1 ratio. Then the image processing module 104 divides the length of the sideline L2 by the number of pixels on the sideline L2 to obtain the ratio of the sideline L2, and so on, to obtain 4 sidelines L1, L2, L3, L4 and 2 diagonal lines The ratio of DI1 and DI2. The processing module 104 then makes an arithmetic mean of the 6 ratios (that is, adds and divides the 6 ratios by 6) to obtain the pixel size of the image.

In addition to the aforesaid relational expression, the relational expression may also use a weighted average to calculate the size of the pixel 142, and the relational expression is as follows:

Wherein, the total length is the selective addition of these measurement values in a weighted form, and the total number of pixels is the addition of the number of pixels of the side lines or the corresponding diagonal lines corresponding to the measurement values. Furthermore, in order to make the size of the obtained pixel 142 more accurate, the processing module 104 can give higher weights to some of the measured values, for example, because the lengths of the two diagonal lines DI1 and DI2 are longer than those of the other side lines L1, L2, L3, and L4 are long, and the number of pixels on the two diagonal lines DI1, DI2 is also greater than the number of pixels on the other edge lines L1, L2, L3, and L4. If the length of the two diagonal lines is given a higher weight ( For example, 1.4), the error of the calculated single pixel will be smaller than the error of the calculated single pixel without weighting, that is, the calculated size of the single pixel will be more accurate.

In addition to the above, the present invention does not limit the measurement of the entire actual lengths of the sidelines L1, L2, L3, L4 and the two diagonals DI1, DI2 to calculate the pixel size of the image 14. If only the measured The size of the pixels of the image 14 can also be obtained from the sideline L1 , the sideline L3 and the diagonal line DI1 through the above two relational expressions, and the present invention is not limited thereto.

Not only that, the device 10 for identifying objects in images may further include a storage module 106, the storage module 106 is used to store the gray scale value of each pixel in the image, store the edge lines L1, L2, L3, L4 of the object 12 in the image 14, and compare The positions of the diagonals DI1, DI2 and the number of pixels stored thereon and the received measured values of the respective edges L1, L2, L3, L4 and the diagonals DI1, DI2, and the calculated size of the pixels 142 are stored In order to calculate the size of the next object by capturing the image with the next object.

In addition to the above, the present invention further has a correction program of the device 10 for correcting and recognizing images. Please refer to FIG. 1 and FIG. 5 together. FIG. 5 is a functional block diagram of the device for correcting and recognizing images according to an embodiment of the invention. As shown in the figure, the image capturing module 102 captures a second image (not shown) carrying the standard test strip 18 . Next, the image processing module 104 recognizes the positions of a plurality of edges (not shown) of the standard test strip 18 from the second image. The image processing module 104 reads the number of pixels on the plurality of edge positions of the aforementioned first standard test piece 18, and after reading, the image processing module 104 compares the number of pixels of the multiple edge positions of the manually measured standard test piece 18 received. The actual length and the corresponding number of pixels are used to calculate the actual length corresponding to each pixel (not shown in the figure).

After the image processing module 104 obtains the actual length corresponding to each pixel, the image processing module 104 directly sends a command signal to the image capture module 102 to instruct the image capture module 102 to capture a third image (not shown). After the image capture module 102 captures the third image, it sends the third image to the processing module 104 for image processing, so as to obtain the number of pixels at multiple edge positions of the standard test strip 18 in the third image. The processing module 104 calculates the multiple sidelines of the standard test piece 18 through the actual length corresponding to each pixel obtained from the aforementioned calculation, and the number of pixels on the multiple sideline positions of the standard test piece 18 obtained from the third image. length.

When the processing module 104 obtains the above-mentioned actual lengths of the multiple edges of the standard test strip 18 manually measured, and the lengths of the multiple edges calculated by the actual length corresponding to each pixel of the third image, the processing module 104 then judges to Whether the ratio of the aforementioned two lengths is within a second range (for example, 0.9 to 1.1), if the processing module 104 determines that the ratio is within the second range, the processing module 104 stops the calibration procedure. If the processing module 104 judges that the aforementioned ratio is not within the second range, the processing module 104 will send a command signal to the image capture module 102, instructing the image capture module 102 to recapture a fourth image (not shown in the figure) ), and re-execute the steps of the calibration procedure of the aforementioned device 10 for calibrating and recognizing images.

In order to enable those skilled in the art to better understand the device for identifying objects in an image of the present invention, further description will be given below in conjunction with the method for identifying objects in an image of the present invention. Please refer to FIG. 1 , FIG. 2 and FIG. 6 together. FIG. 6 is a flowchart of a method for recognizing an image according to an embodiment of the present invention. As shown in FIG. 6 , in step S400 , the image processing module 104 receives a plurality of measured values, wherein each measured value corresponds to one of the actual lengths of the plurality of edges or the plurality of diagonals of the corresponding object 12 . In step S402 , the image capture module 102 captures the image 14 with the object 12 . In step S404 , the image processing module 104 executes an edge detection program to obtain the positions of a plurality of edges L1 , L2 , L3 , L4 of the object 12 from the image 14 .

In step S406, the image processing module 104 obtains the positions of the diagonal line DI1 and the diagonal line DI2 among the multiple diagonal lines of the object 12 in the image 14 according to the positions of the multiple edge lines L1, L2, L3, and L4 of the object 12 , the image processing module 104 . In step S408 , the image processing module 104 calculates the number of pixels at the positions of the diagonal line DI1 and the diagonal line DI2 of the object in the image. In step S410 , the image processing module 104 calculates the number of pixels at the positions of the edges L1 , L2 , L3 , and L4 of the object 12 in the image. In step S412 , the image processing module 104 calculates the pixel size of the image 14 according to the number of pixels at the positions of the edges L1 , L2 , L3 , L4 and the diagonals DI1 , DI2 of the object 12 . In step S414 , the image processing module 104 determines whether the ratio of the number of pixels of the diagonal line DI1 and the diagonal line DI2 is within the first range. In step S416, when the image processing module 104 judges that the ratio of the number of pixels of the diagonal line DI1 and the diagonal line DI2 is within the first range, the image processing module 104 outputs a recognition result, which means that the object 12 is determined to be a rectangle . In step S418, when the image processing module 104 judges that the ratio of the number of pixels of the diagonal line DI1 and the diagonal line DI2 is not within the first range, the image processing module 104 outputs another recognition result, which is the judgment object 12 is not a rectangle.

It is worth noting that the order of the aforementioned steps S400, S402, S404, and S406 can be changed. For example, the present invention can firstly perform the steps S402, S404, and S406, and then perform the actions of the step S400. Moreover, the order of step S408 and step S410 can also be reversed, that is, the number of pixels at the positions of the edges L1, L2, L3, and L4 of the object 12 in the image 14 can be calculated first, and then the number of pixels of the object 12 in the image 14 can be calculated. The number of pixels at the positions of the diagonal lines DI1 and DI2 depends on user requirements, and the present invention is not limited thereto.

Next, please refer to FIG. 1 , FIG. 2 and FIG. 7 together. FIG. 7 is a flowchart of a method for calibrating an image recognition device according to an embodiment of the present invention. As shown in the figure, in step S500 , the image processing module 104 receives manual measurement of the actual lengths of the multiple sidelines of the standard test strip 18 . In step S502 , the image capture module 102 captures a second image (not shown) carrying the standard test strip 18 . In step S504 , the image processing module 104 reads a plurality of pixel numbers at positions of a plurality of edge lines (not shown in the figure) of the standard test strip 18 . In step S506 , the image processing module 104 compares the actual lengths of the edges of the standard test strip 18 with the corresponding pixel numbers to calculate the actual length corresponding to each pixel (not shown). In step S508 , the image capturing module 102 captures a third image (not shown in the figure) carrying the standard test strip 18 .

In step S510 , the image processing module 104 calculates the lengths of multiple sidelines of the standard test strip 18 according to the third image and the actual length corresponding to each pixel. In step S512 , the image processing module 104 determines whether the ratio of the actual lengths of the edges of the standard test strip 18 to the lengths calculated from the third image is within the second range. When the image processing module 104 judges that the ratio of the actual lengths of the multiple sidelines of the standard test strip 18 to the lengths calculated by the third image is not within the second range, then the image recognition method of the present invention will be executed again from step S500. , a method for calibrating an image recognition device. In step S514, if the image processing module 104 judges that the ratio of the actual lengths of the multiple sidelines of the standard test strip 18 to the length calculated by the third image is within the second range, the image processing module 104 then proceeds to end the calibration procedure. , and then start the image recognition method of the present invention (step S400).

To sum up, the device for identifying objects in an image of the present invention can obtain multiple edges of the object in the image by executing an edge detection program, and obtain multiple diagonals of the object through the multiple edges of the object. Finally, by comparing the number of pixels on multiple diagonal positions of the object in the image, the user can know the shape of the object by capturing the image of the object. In addition, the present invention obtains the pixel size in the image by measuring the actual length of the sideline and diagonal line of the object, and the number of pixels on the sideline and diagonal line of the object in the image, so that the user can All object sizes need to be measured, only the actual length of the first object needs to be measured, other subsequent objects can directly capture the image of the object, and obtain the actual length of the edge of the subsequent object through the pixel size of the image.

Not only that, the present invention also has a calibration procedure for the device for calibrating and identifying images, through the number of pixels on the multiple edge positions of the standard test piece in the captured second image, and manually measuring the multiple edge lines of the standard test piece to calculate the actual length corresponding to the pixel. And the calculated lengths of multiple sidelines of the standard test piece are obtained through the actual lengths corresponding to the pixels and the captured third image. Finally, the processing module judges whether the ratio between the actual lengths of the manually measured multiple edges and the calculated lengths of the multiple edges is greater than the second range, so that the image recognition device can correct itself without manual correction by the user , and increase the accuracy of measuring objects in the assembly line.

Of course, the present invention can also have other various embodiments, and those skilled in the art can make various corresponding changes and deformations according to the present invention without departing from the spirit and essence of the present invention, but these corresponding Changes and deformations should belong to the scope of protection of the appended claims of the present invention.

Claims (10)

1. A method for identifying images, comprising: capturing a first image containing an object by means of an image recognition device; calculating the length of a first diagonal and the length of a second diagonal of the object based on the first image; and When the ratio of the length of the first diagonal line to the length of the second diagonal line is within a first range, an identification result is output. 2. The image recognition method according to claim 1, wherein the step of calculating the length of the first diagonal and the length of the second diagonal of the object comprises: Executing an edge detection program to obtain positions of edges of the object from the first image; and Obtaining the positions of the first diagonal and the second diagonal of the object in the first image according to the positions of the edges of the object, and reading the first diagonal and the second pair of diagonals The number of pixels at the position of the diagonal line to calculate the length of the first diagonal line and the length of the second diagonal line of the object. 3. The image recognition method according to claim 2, wherein the edge detection program is to compare whether the difference between the gray scale values of a first pixel and a second pixel of the first image is greater than a threshold value, when the difference between the grayscale values of the first pixel and the second pixel of the first image is greater than the threshold value, the position of the first pixel in the first image is the The positions of the pixels of the edge, where the second pixel is adjacent to the first pixel. 4. The method for identifying images according to claim 3, further comprising: receiving a plurality of measurements, each of which corresponds to one of the actual lengths of the edges or the diagonals; and After the step of executing the edge detection program to obtain the positions of the edges of the object from the first image, obtain and output according to a relational expression between the measured values and the positions of the edges and the diagonals The pixel size of the first image. 5. The method for identifying images according to claim 4, wherein the relational expression is: Wherein, the total ratio is the addition of the ratios of the measured values to the corresponding pixel numbers of the side lines and the diagonal lines, and the measured number is the number of received measured values. 6. The method for identifying images according to claim 4, wherein the relational expression is: Wherein, the total length is the selective addition of the measured values in a weighted form, and the total number of pixels is the sum of the pixel numbers of the corresponding side lines or the corresponding diagonal lines. 7. The image recognition method according to claim 1, characterized in that, before the step of capturing the first image carrying the object by using the image recognition device, the method comprises: The device for correcting the image recognition, wherein the calibration procedure of the device for recognizing images includes: (a) Receive manual measurement of the actual lengths of multiple sidelines of a standard test piece; (b) using the image recognition device to capture a second image containing the standard test strip; (c) based on the second image, read a plurality of pixel numbers at the positions of the edges of the standard test strip; (d) Comparing the actual lengths of the edges of the standard test piece with the corresponding pixel numbers, to calculate the actual length corresponding to each pixel; (e) reusing the image recognition device to capture a third image containing the standard test strip; (f) calculating the lengths of the sidelines of the standard test piece according to the third image and the corresponding actual length of each pixel; and (g) judging whether the ratio of the actual length of the edges of the standard test piece to the length calculated by the third image is within a second range; Wherein if it is judged that the ratio of the actual lengths of the edges of the standard test piece to the length calculated by the third image is within the second range, the calibration procedure ends, and if the actual length of the edges of the standard test piece is judged When the ratio to the length calculated by the third image is not within the second range, repeat steps (a) to (g). 8. A device for identifying images, characterized in that it comprises: an image capturing module, used for capturing an image with an object; and An image processing module, coupled to the image capture module, the image processing module executes an edge detection program to obtain the positions of multiple edges of the object from the image, and the image processing module according to the position of the edges of the object The position obtains the position of a first diagonal and a second diagonal among the multiple diagonals of the object in the image, and calculates the pixels on the position of the first diagonal and the second diagonal number, and judging whether the ratio of the number of pixels of the first diagonal line and the number of pixels of the second diagonal line is within a first range; Wherein when the ratio of the number of pixels of the first diagonal line and the second diagonal line is within the first range, an identification result is output. 9. The image recognition device according to claim 8, wherein the edge detection program is to compare whether the difference between the grayscale values of a first pixel and a second pixel of the image is greater than a threshold value, When the difference between the grayscale values of the first pixel and the second pixel of the image is greater than the threshold value, the position of the first pixel in the image is the position of the pixels of the borders of the object, Wherein the second pixel is adjacent to the first pixel, the image processing module receives a plurality of measured values, wherein each of the measured values corresponds to one of the actual lengths of the side lines or the diagonal lines, and the image After the processing module executes the edge detection program to obtain the positions of the edges of the object from the image, it obtains and outputs the The dimensions of the pixels of the image, where the relation is: Wherein, the total ratio is the addition of the ratios of the measured values to the corresponding pixel numbers of the side lines and the diagonal lines, and the measured number is the number of the received measured values. 10. The image recognition device according to claim 8, wherein the edge detection program is to compare whether the difference between the grayscale values of a first pixel and a second pixel of the image is greater than a threshold value, When the difference between the grayscale values of the first pixel and the second pixel of the image is greater than the threshold value, the position of the first pixel in the image is the position of the pixels of the borders of the object, Wherein the second pixel is adjacent to the first pixel, the image processing module receives a plurality of measured values, wherein each of the measured values corresponds to one of the actual lengths of the side lines or the diagonal lines, and the image After the processing module executes the edge detection program to obtain the positions of the edges of the object from the image, it obtains and outputs the The dimensions of the pixels of the image, where the relation is: Wherein, the total length is the selective addition of the measured values in a weighted form, and the total number of pixels is the sum of the pixel numbers of the corresponding side lines or the corresponding diagonal lines.