TWI723541B - Sheet metal identification method - Google Patents

Abstract

A sheet metal identification method is executed by a processing device and includes the following steps: A. Receive a sheet metal image; B. Normalize the image to obtain a converted image file; C. Obtain a converted image file based on the converted image file Image feature point data; D. Compare the image feature point data with a plurality of sheet metal feature models in a database to obtain at least one target sheet metal feature model; E. Output the identification mark of the target sheet metal feature model. A sheet metal identification system includes: a platform, an image module, and the aforementioned processing device to capture sheet metal images and output the sheet metal images to the processing device.

Description

Sheet metal identification method

The present invention is related to sheet metal identification; in particular, it refers to an automated sheet metal identification method and identification system.

Sheet metal work is a type of mechanical processing. The work involves cutting and cutting the metal into semi-finished products of various shapes, and then processing or assembling the semi-finished products for use. The application range of finished sheet metal is very wide. Life is also very closely related. With the advancement of science and industry, sheet metal processing operations tend to be sophisticated and automated manufacturing and development to improve product quality and productivity.

The conventional sheet metal identification is manual identification. After the sheet metal is made into semi-finished products of various shapes through the previous process, the operator will identify the various sheet metal according to the sheet metal appearance data manual, and determine the subsequent process of the sheet metal. However, this method is not only time-consuming, but also inefficient. In addition, in order to meet various needs, the shape of the sheet metal has changed a lot. The use of manual identification and identification obviously requires a lot of labor and time costs. So, how Providing a system for automatically identifying sheet metal is one of the directions of the inventor's painstaking research.

In view of this, the purpose of the present invention is to provide a sheet metal identification method and identification system, which can automatically identify sheet metal and improve the problem of inefficiency of manual sheet metal identification.

To achieve the above objective, the present invention provides a sheet metal identification method, which is executed by a processing device, which includes the following steps: A. Receive a sheet metal image; B. Normalize the image to obtain a converted image file C. Obtain an image feature point data according to the converted image file; D. Compare the image feature point data with a plurality of sheet metal feature models in a database to obtain at least one target sheet metal feature model, wherein the The sheet metal feature model corresponds to a plurality of sheet metals, and each of the sheet metal feature models includes a feature point data and an identification mark, the feature point data is for comparison with the image feature point data; the target sheet metal feature model is the sheet metal At least one sheet metal feature model whose feature point data of the feature model matches or is similar to the image feature point data; E. Output the identification mark of the target sheet metal feature model.

The present invention further provides a sheet metal identification system, including a platform and an image module, wherein the platform has a top surface for placing at least one sheet metal; the image module is arranged on the platform to capture the at least one sheet metal Sheet metal image, and output the sheet metal image; and the above-mentioned processing device, which is electrically connected to the image module.

The effect of the present invention is that by comparing the image feature point data of the sheet metal with a plurality of sheet metal feature models in the database, the target sheet metal feature model is obtained. In this way, the user can obtain the target sheet metal feature model according to the target sheet metal feature model. Identify the identity of the sheet metal to facilitate subsequent manufacturing processes.

〔this invention〕

1: Sheet metal identification system

10: Platform

12: Top surface

20: light source

30: Image module

40: processing device

41: printing device

50: sheet metal

60: Shading parts

70: Convert model graphics

80: Convert image graphics

90: screen

92: Print button

S01~S04, S1~S5: steps

A1, A2, A3: area

L1: short side

L2: Long side

FIG. 1 is a schematic diagram of a sheet metal identification system according to a preferred embodiment of the present invention.

FIG. 2 is a flowchart of a sheet metal identification method according to a preferred embodiment of the present invention.

FIG. 3 is a flowchart of the sheet metal identification method of the preferred embodiment of the above-mentioned preferred embodiment.

Fig. 4 is a schematic diagram of a sheet metal of a preferred embodiment of the present invention.

5 is a schematic diagram of the target sheet metal feature models arranged in order according to the comparison value according to a preferred embodiment of the present invention

In order to explain the present invention more clearly, a preferred embodiment is given below in conjunction with the drawings in detail as follows. Please refer to FIG. 1, which is a sheet metal identification system 1 according to a preferred embodiment of the present invention, which includes a platform 10, an image module 30 and a processing device 40, in which:

The platform 10 has a top surface 12 for placing at least one sheet metal 50. The image module 30 is disposed above the platform 10 and includes a camera for capturing the image of the at least one sheet metal 50 and outputting the image of the sheet metal 50. The processing device 40 is electrically connected to the image module 30 to receive the image of the sheet metal 50 output by the image module 30; preferably, the sheet metal identification system includes a light source 20 which is arranged under the platform 10, The light source 20 can be an LED light-emitting panel, which includes a plurality of LEDs, and the above-mentioned platform can be a platform with a light-transmitting function. In this way, the image module arranged above the platform can also achieve capture and placement The effect of the image of the sheet metal on the platform is not limited to the above-mentioned embodiment.

In this embodiment, the sheet metal identification system 1 includes a shading member 60, which is disposed on the platform 10 and located around the platform 10 to prevent the interference of external light from the platform 10. In practice, The sheet metal identification system 1 may not be provided with the shading member 60. Thereby, when the user places a sheet metal 50 on the platform 10, the light of the LED light-emitting board can pass through the platform, and is partially blocked by the sheet metal 50, and the light not blocked by the sheet metal 50 enters the camera, thereby, In the camera of the image module 30, a clear sheet metal image can be obtained, And the sheet metal image has the outer contour of the sheet metal 50. If the sheet metal 50 has a hollow area (such as perforations or openings), the sheet metal image has an inner contour of the sheet metal 50.

Please refer to FIG. 2, which is the sheet metal identification method provided by the present invention, which is executed by the above-mentioned processing device 40. The sheet metal identification method includes a database creation step, and the database creation step is as follows: S01. A two-dimensional model drawing file; the user inputs the two-dimensional model drawing file of the sheet metal to the processing device 40, and the processing device 40 receives the two-dimensional model drawing file of the sheet metal, wherein the two-dimensional model drawing file uses an engineering drawing The two-dimensional vector graphics file generated by the software means that the user processes the sheet metal 50 according to the two-dimensional model file. The file format can be DWG file format, DXF file format or other file formats.

The database creation step further includes S02. Performing the normalized conversion of the length unit of the two-dimensional model drawing file to obtain a converted model drawing file; the processing device 40 performs the two-dimensional model drawing file as the length unit of the side length For normalization conversion, for example, the length units in different two-dimensional model files may be marked inconsistently. For example, some use cm to represent length, and some use mm to represent length. Therefore, the processing device 40 transfers the two-dimensional The length unit of the graphics of the model drawing file is converted into a consistent unit (for example, mm is used as the consistent length unit), and then a converted model drawing file with the same length unit is obtained. The file format of the converted model drawing file can be BMP , JPG, or PNG file format or other image file format.

Then perform step S03. Obtain a feature point data according to the conversion model drawing file; preferably, the feature point data includes at least one feature point item, and the at least one feature point item includes the area, perimeter, center of mass, and At least one of rectangle degree, hole ratio, minimum external rectangle aspect ratio, Hu moment invariant of sheet metal, and distance from center to outer frame, please refer to Figure 3, where the rectangle degree is enclosed by the periphery of the sheet metal The ratio of the area A1 to the area A2 of the smallest rectangle that can enclose the contour of the sheet metal. The hole ratio is the ratio of the area enclosed by the periphery of the sheet metal A1 minus the area A3 of the holes on the sheet metal to the area enclosed by the periphery of the sheet metal A1, the minimum external moment Type length and width The ratio is the ratio of the short side L1 to the long side L2 of the smallest rectangle that can enclose the outer contour of the sheet metal, and the Hu moment invariant is Hu. MK proposed the concept of moment invariant in 1962, using the nonlinear combination of geometric moments. A set of moment invariants with desired scale invariance, translation invariance, and rotation invariance are generally called Hu moment invariants. Hu moment invariants have the characteristic of not changing with the position and direction of the image. That is, even after the image is rotated, zoomed, or translated, the Hu moment remains unchanged.

Then perform step S04. Connect an identification mark corresponding to the sheet metal of step S01 to the feature point data in step S03 to form a sheet metal feature model and store it in the database, where the identification mark may include the product number of the sheet metal 50, etc. The product data for the user to identify the sheet metal 50; preferably, in step S04, it further includes adding the conversion model image file in step S02 to the sheet metal feature model. Finally, perform step S05. Repeat steps S01 to S04 several times to store a plurality of sheet metal feature models of different sheet metals in the database; in this way, the establishment of the database is completed.

Please cooperate with FIG. 4, after completing the above database creation steps, perform the following steps: Step S1. Receive a sheet metal image; in this embodiment, the user places the sheet metal 50 to be identified on the platform 10. After the image module 30 captures the appearance image of the sheet metal 50 to be identified, the processing device 40 receives a sheet metal image output from the image module 30.

Step S2. Convert the image (also referred to as normalized conversion) to obtain a converted image file with the length unit. In this embodiment, after the processing device 40 receives the sheet metal image output from the image module 30, the image is normalized and converted. The aforementioned normalized conversion may include gray-scale binarization processing, that is, the processing device The gray scale value of the pixel whose image is larger than a critical gray scale value is set to the maximum gray scale value, and the gray scale value of the pixel smaller than the critical gray scale value is set to the minimum gray scale value to obtain the converted image image with the length unit The unit of the length sheet is the same as the length unit of the converted model drawing file (for example, both are mm). In step S2 It may include performing a pre-processing on the image, such as removing image noise, enhancing image integrity, and so on.

Step S3. Obtain an image feature point data according to the converted image file; preferably, the image feature point data includes at least one feature point item, and the at least one feature point item includes the area, perimeter, center of mass, and At least one of rectangle degree, hole ratio, minimum external rectangle aspect ratio, Hu moment invariant of sheet metal and distance from center to outer frame, among which the calculation method of rectangle degree, hole ratio and minimum external rectangle aspect ratio As mentioned before, I will not repeat it here.

Step S4. Compare the image feature point data with a plurality of sheet metal feature models in the database to obtain at least one target sheet metal feature model, wherein the sheet metal feature models correspond to a plurality of sheet metals, and each of the sheet metal features The model includes a feature point data and an identification mark, the feature point data is for comparison with the image feature point data; the target sheet metal feature model is that the feature point data of the sheet metal feature models match or match the image feature point data Approximately at least one feature model of the sheet metal; for example, the processing device 40 includes image feature point data of the sheet metal 50 such as area, perimeter, centroid, rectangle degree, hole ratio, minimum circumscribed rectangle aspect ratio, and The feature point item of Hu moment invariant of sheet metal and the feature point data of multiple sheet metal feature models in the database, such as area, circumference, centroid, rectangle degree, hole ratio, minimum external rectangle aspect ratio, and sheet metal After comparing the feature point items of Hu moment invariants, a target sheet metal feature model is obtained according to the comparison result, where the target sheet metal feature model is the most similar sheet metal feature model in the database to the sheet metal image, so First, the user can identify the product data of the sheet metal 50 by using the identification mark of the target sheet metal feature model.

Finally, step S5 is performed. The identification mark of the target sheet metal feature model is output. The identification mark may include at least one of characters, numbers, and symbols, or a combination of two or more. In one embodiment, the processing device 40 uses a printing device 41 to print Identification of the target sheet metal feature model The mark is printed as an identification label, the identification label can be attached to the sheet metal, the identification label is used for a reading tool to read, the identification label can be a one-dimensional bar code or a two-dimensional bar code, the reading tool can It is a barcode scanner; the processing device can also output the identification mark of the target sheet metal feature model as an image identification label through a projection device, and project it on the sheet metal for identification; the processing device can also output through a display device Identification mark.

In one embodiment, each of the sheet metal feature models in step S4 further includes a corresponding conversion model image file; and steps S4 and S5 include the conversion model image file of the at least one target sheet metal feature model to convert the model The graphic 70 is displayed on a screen 90. Preferably, the converted image file of the sheet metal to be recognized is also displayed on the screen as the converted image graphic 80, so that the user can confirm the target on the screen at the same time The conversion model drawing file of the sheet metal feature model and the conversion image drawing file of the sheet metal to be identified.

In one embodiment, the number of the at least one target sheet metal feature model in step S4 is plural; between steps S4 and S5, the conversion model image files of the target sheet metal feature models and the conversion image file are converted The image graphic 80 is displayed on the screen 90; Step S5 includes selecting one of the target sheet metal feature models by the person, and printing the selected identification mark of the target sheet metal feature model, that is, the target sheet metal Feature models are several sheet metal feature models in the database that are similar to the sheet metal image. The user can select a target sheet metal feature model from the target sheet metal feature models, and print the selected target sheet metal feature The identification mark of the model.

In one embodiment, step S4 is based on the feature point items related to geometric features in the image feature point data, such as area, perimeter, centroid, rectangle degree, hole ratio, and minimum circumscribed rectangle aspect ratio. The feature point items are respectively compared with the feature point items of the feature point data of the sheet metal feature models to obtain a plurality of comparison values. In this embodiment, the comparison value is the value of the feature point items of the sheet metal feature model minus Feature point item of image feature point data After the target value, the absolute value is taken and then divided by the value of the feature point item of the image feature point data; and among the comparison values, at least one sheet metal feature model corresponding to a threshold value less than a threshold value is set as at least one target Sheet metal feature model, for example, the user sets the threshold value to 0.5, and the ratio of feature point items such as area, perimeter, and centroid of one of the sheet metal feature models in the database to the feature point items of the image feature point data If the pair values are all less than the threshold value (0.5), the processing device 40 determines that the sheet metal feature model in the database is the target sheet metal feature model. When the feature point items of the plurality of sheet metal feature models in the database are such as area When the comparison values of, perimeter, and center of mass with the feature point items of the image feature point data all meet the condition of less than 0.5, the processing device 40 determines that the sheet metal feature models in the database are all target sheet metal feature models .

In one embodiment, in step S4, according to the Hu moment invariants of the image feature point data and the Hu moment invariants of the feature point data of the sheet metal feature models, respectively, a plurality of similarities are calculated by the Euclidean distance algorithm. Value; and according to the similarity values, the at least one sheet metal feature model that matches or is similar to the image feature point data is set as at least one target sheet metal feature model. In practice, in step S4, the image can be first based on The Hu moment invariants of the feature point data are respectively compared with the Hu moment invariants of the feature point data of the sheet metal feature models and arranged in sequence according to the similarity value, and several most similar sheet metal feature models are selected, and the processing device 40 Then according to the feature point data of the image and the feature point items related to the geometric features, such as area, perimeter, centroid, rectangle degree, hole ratio, minimum external rectangle aspect ratio, and feature point items in the database respectively The feature point items of the feature point data of the selected sheet metal feature models are compared to obtain a plurality of comparison values, and the processing device 40 among the comparison values is less than the threshold value corresponding to the sheet metal feature model Set as the target sheet metal feature model, so that when the number of Hu moment invariants is too large, the geometric features can be used to further filter to reduce the number. As shown in Figure 5, the target sheet metal feature models are based on the comparison value They are arranged in order and displayed on the screen 90. The comparison values of the target sheet metal feature models in Figure 5 are (0.1, 0.2, 0.3) respectively. After the user clicks the target sheet metal feature model with a similarity of 0.1, and then clicks a print button 92, the processing device 40 prints the identification mark as an identification label through the printing device 41.

In addition, in step S4, the processing device 40 may also first base on the image feature point data and geometric features related feature point items, such as area, circumference, centroid, rectangle degree, hole ratio, minimum circumscribed rectangle length The feature point items of the aspect ratio are respectively compared with the feature point items of the feature point data of the sheet metal feature models in the database to obtain a plurality of comparison values respectively, and the sheet metal features whose comparison value is less than the threshold value are selected Model, and then compare the Hu moment invariants of the image feature point data with the Hu moment invariants of the feature point data of the selected sheet metal feature models in the database and arrange them in order according to the similarity value. The processing device 40 determines that the several most similar sheet metal feature models in the database are the target sheet metal feature models. In this way, when the number of geometric features after comparison is too large, the Hu moment invariant is used to further filter to reduce the number. . In an embodiment, the processing device 40 may repeatedly perform the comparison of feature point items related to geometric features and Hu moment invariants.

In one embodiment, the processing device 40 can sequentially compare the feature point items of the image feature point data with the feature point items of the sheet metal feature models. For example, the processing device 40 can first target one of the feature points The items are compared, and the sheet metal feature model whose comparison value is less than the threshold value is selected, and then another feature point item of the image feature point data is compared with the feature point item of the selected sheet metal feature model in the database, and After repeating the above steps, at least one target sheet metal feature model is obtained.

For example, if the user sets a threshold value, the processing device 40 first compares the area feature point items, and the processing device 40 compares the area feature point items of the image feature point data with those of the sheet metal feature models in the database. Area feature point project comparison, get Corresponding multiple comparison values of the area feature point items, among which, the smaller the comparison value, the closer the area feature point items are. Then select the sheet metal feature model whose comparison value is less than the threshold value. If the number of the sheet metal feature models selected is greater than a set number, the processing device 40 then focuses on the feature point item of the distance from the center to the outer frame of the image feature point data Compare the center value outer frame distance feature point items of the selected sheet metal feature model in the database to obtain a plurality of comparison values corresponding to the center to outer frame distance feature point items. Among them, the smaller the comparison value, represents the center The distance from the outer frame to the feature point items is closer. Then select the sheet metal feature model whose comparison value is less than the threshold value. If the number of sheet metal feature models whose comparison value is less than the threshold value is still greater than the set number, the processing device 40 continues to repeat the foregoing steps with other Different feature point items are compared until the number of selected sheet metal feature models is less than or equal to the set number, or until all feature point items of the image feature point data are compared, and finally, the selected sheet metal feature model is set It is the target sheet metal feature model. In this way, a lot of comparison time can be saved to improve work efficiency.

In one embodiment, the user can set different threshold values for various feature point items to achieve a better comparison result. For example, the user can set a corresponding threshold value according to each feature point item. For example, the threshold value for the area feature point item is a first threshold value, and the threshold value for the distance feature point item from the center to the outer frame is a first threshold value. Two threshold values. When the processing device 40 compares the area feature point items, the processing device 40 compares the area feature point items of the image feature point data with the area feature point items of the sheet metal feature models in the database to obtain the corresponding A plurality of comparison values of the area feature point item, and select a sheet metal feature model whose comparison value is less than the first threshold value. If the number of the sheet metal feature models selected as described above is greater than a set number, the processing device 40 then targets the image Compare the center to outer frame distance feature point item of the feature point data with the center value outer frame distance feature point item of the selected sheet metal feature model in the database, and obtain the corresponding center to outer frame distance feature point item A plurality of comparison values are selected, and a sheet metal feature model with a comparison value less than the second threshold value is selected to select a more similar sheet metal feature model again. In this way, the purpose of saving comparison time and improving work efficiency can also be achieved.

In summary, the effect of the present invention is that the user can compare the image feature point data of the sheet metal with a plurality of sheet metal feature models in the database to obtain the target sheet metal feature model. In this way, the user can obtain the target sheet metal feature model according to The target sheet metal feature model can identify the identity of the sheet metal to facilitate subsequent manufacturing processes. Furthermore, the database of the present invention does not need to use a large number of sheet metal images for smart training. The database of the present invention is inputted by two-dimensional Model drawing file creation, which provides a more efficient database creation method. In practice, if there is an existing database, there is no need to perform the database creation steps of S01~S05, just perform steps S1~S5 to identify the sheet metal.

The above are only the preferred and feasible embodiments of the present invention. Any equivalent changes made by applying the specification of the present invention and the scope of the patent application should be included in the patent scope of the present invention.

1: Sheet metal identification system

10: Platform

12: Top surface

20: light source

30: Image module

40: processing device

41: printing device

50: sheet metal

60: Shading parts

90: screen

Claims (8)

A sheet metal identification method, executed by a processing device, includes the following steps: A. receiving a sheet metal image; B. converting the image to obtain a converted image file with a length unit; C. according to the conversion Obtain an image feature point data from the image file; D. Compare the image feature point data with a plurality of sheet metal feature models in a database to obtain at least one target sheet metal feature model, wherein the sheet metal feature models correspond to each other A plurality of sheet metal, and each of the sheet metal feature models includes a feature point data and an identification mark connected to the feature point data, the feature point data is for comparison with the image feature point data; the target sheet metal feature model is the sheet metal At least one sheet metal feature model whose feature point data of the feature model matches or is similar to the image feature point data; E. Output the identification mark of the target sheet metal feature model; wherein each feature point data of the sheet metal feature model and the image feature The point data all include at least one feature point item, and the at least one feature point item includes the Hu moment invariant of the sheet metal. In step D, the Hu moment invariant of the image feature point data is respectively based on the characteristics of the sheet metal feature models. The Hu moment invariant of the point data is calculated by the Euclidean distance algorithm to obtain a plurality of similarity values, and the at least one sheet metal feature model that is consistent with or similar to the image feature point data is set as At least one feature model of the target sheet metal. The sheet metal identification method according to claim 1, wherein in step E, the identification mark of the target sheet metal feature model is printed as an identification label, and the identification label is used for reading by a reading tool. The sheet metal identification method according to claim 1, wherein the at least one feature point item includes the area, circumference, centroid, rectangular degree, hole ratio, and minimum external rectangular length and width of the sheet metal One of the ratio and the distance from the center to the outer frame, where the rectangular degree is defined as the ratio of the area enclosed by the sheet metal periphery to the area of the smallest rectangle that can enclose the sheet metal outline, and the hole ratio is defined as the sheet metal periphery The area deducts the ratio of the hole area on the sheet metal to the area enclosed by the sheet metal periphery. The minimum circumscribed rectangular aspect ratio is defined as the ratio of the short side to the long side of the smallest rectangle that can enclose the outline of the sheet metal. The sheet metal identification method of claim 1, wherein before step A, it further includes a database creation step for creating the sheet metal feature models of the database, and the database creation step includes: A1. receiving a sheet metal A two-dimensional model drawing file; A2. The two-dimensional model drawing file is normalized and converted to obtain a conversion model drawing file with the length unit; A3. A feature point data is obtained according to the conversion model drawing file; A4 Connect one of the identification marks corresponding to one of the sheet metal in step A1 to the feature point data in step A3 to form a feature model of the sheet metal and store it in the database; A5. Repeat steps A1 to A4 multiple times. The sheet metal identification method according to claim 4, wherein step A4 further includes adding the conversion model drawing file in step A2 to a sheet metal feature model; each of the sheet metal feature models in step D further includes a corresponding conversion model Drawing files; between steps D and E, the conversion model drawing file of the at least one target sheet metal feature model is graphically displayed on a screen. A method for identifying sheet metal, executed by a processing device, includes the following steps: A. receiving a sheet metal image; B. converting the image to obtain a converted image file with a length unit; C. Obtain an image feature point data according to the converted image file; D. Compare the image feature point data with a plurality of sheet metal feature models in a database to obtain at least one target sheet metal feature model, wherein the sheet metal The feature models respectively correspond to a plurality of sheet metals, and each of the sheet metal feature models includes a feature point data and an identification mark connected to the feature point data, the feature point data is for comparison with the image feature point data; the target sheet metal feature The model is at least one sheet metal feature model whose feature point data of the sheet metal feature models matches or is similar to the image feature point data; E. Outputs the identification mark of the target sheet metal feature model; wherein a database is further included before step A The establishment step is used to establish the sheet metal feature models of the database. The database establishment step includes: A1. receiving a two-dimensional model image file of the sheet metal; A2. normalizing the two-dimensional model image file , To obtain a conversion model drawing file with the length unit; A3. Obtain a feature point data according to the conversion model drawing file; A4. Link an identification mark corresponding to one of the sheet metal in step A1 to one in step A3 Feature point data to form a sheet metal feature model and store it in the database; A5. Repeat steps A1 to A4 several times; wherein step A4 further includes adding the conversion model drawing file in step A2 to a sheet metal feature model Each of the sheet metal feature models in step D further includes a corresponding conversion model drawing file; steps D and E include the conversion model drawing file of the at least one target sheet metal feature model graphically displayed on a screen; wherein In step D, the number of the at least one target sheet metal feature model is plural; between steps D and E, the conversion model drawing file of the target sheet metal feature model and the conversion image drawing file are The graph is displayed on the screen; Step E includes selecting one of the target sheet metal feature models by a person, and printing the selected identification mark of the target sheet metal feature model. The sheet metal identification method according to claim 3, wherein in step D, the feature point items of the feature point data of the image are compared with the feature point items of the feature point data of the sheet metal feature models to obtain a plurality of ratios respectively. Value; and at least one sheet metal feature model corresponding to a threshold value in the comparison values is set as at least one target sheet metal feature model. The sheet metal identification method according to claim 3, wherein the number of the feature point data and the at least one feature point item of the image feature point data is plural and the feature point items are different; wherein in step D, It is based on the comparison of one of the feature point items of the image feature point data with the feature point items of the feature point data of the sheet metal feature models to obtain a plurality of comparison values respectively, and select the comparison values to be less than A sheet metal feature model of a threshold value; at least once, another feature point item of the image feature point data is compared with the feature point item of the selected sheet metal feature model in the database to obtain at least one sheet metal feature model and set Is at least one feature model of the target sheet metal.

Images