KR101836125B1 - Method for generating shape feature information of model and method for analyzing shape similarity using theory - Google Patents

Abstract

The present invention relates to a method for generating shape feature information for a model. The method comprises a step (a) of dividing the model into a plurality of meshes; and a step (b) of generating shape feature information based on the plurality of meshes. The shape feature information may relate to a distance between arbitrary sample points existing in each of the two meshes when the plurality of meshes is sampled as a plurality of samples constituting a pair of two meshes. It is possible to solve a problem such as the waste of time and the decrease of productivity.

Description

METHOD FOR GENERATING SHAPE FEATURE INFORMATION AND METHOD FOR ANALYZING SHAPE SIMILARITY USING THEORY [0002]

The present invention relates to a shape feature information generation method and a shape similarity analysis method of a model.

The shape information obtained from 3D measurement equipment is actively utilized in the reverse engineering and quality inspection fields required in various industries.

The three-dimensional measuring device can be classified into a contact type and a non-contact type depending on whether or not the object is in contact with the object. Although the contact speed is somewhat lower than the non-contact type in which the three-dimensional spatial coordinates of the measurement object are measured by using an optical technique such as a camera or a laser, the probe is directly contacted to the measurement object, It is less affected by the environment and has the advantage that the measurement accuracy is superior to the non-contact type.

The contact type measurement is a method of using a coordinate measuring machine (CMM), which requires a separate space for measurement, and a method of changing a tool to a probe on a machine tool, (OMM, On-Machine Measurement) method. CMM has a disadvantage in that the measurement accuracy is higher than that of meteorological measurement but the size of the measurement object is limited and the initial investment cost is high. In addition, after machining such as roughing, medium cutting and finishing, CNC (computerized numerical control) It is necessary to move the measuring instrument to a three-dimensional measuring machine. On the other hand, the meteorological measurement can be performed directly in the work space by replacing the tool with a probe in the course of machining in the CNC machining center or after machining is completed.

The meteorological measurement has the advantage of measuring the machined state of the workpiece and solving various problems occurring during machining so that the machining can be performed immediately. However, it has the following limitations. First, although the existing shape model in which the measurement point is established and the feature of the new object (new model) are similar to each other, the measurement point must be re-designated to the new model every time, And the productivity is low. Second, since the position and the measurement path of the measurement point for the new model are set according to the empirical judgment of the operator, there is a problem that the measurement quality is affected by the skill level of the operator.

The technique of the present invention is disclosed in Japanese Laid-Open Patent Publication No. 2015-093346 (published on May 5, 2015).

It is an object of the present invention to solve the above-mentioned problems of the conventional art, and it is an object of the present invention to provide a method and apparatus for generating a path by newly specifying a measurement point in a new model, And to provide a shape feature information generation method and a shape similarity analysis method of a model capable of solving a problem of waste of time and productivity caused by collision inspection.

The present invention has been made to solve the above problems of the prior art and it is an object of the present invention to solve the problem that a difference in measurement quality is caused according to skill of an operator by setting a measurement point and a measurement point position for a new model according to an empirical judgment of an operator And to provide a shape feature information generation method and a shape similarity analysis method of a model that can be used.

It is to be understood, however, that the technical scope of the embodiments of the present invention is not limited to the above-described technical problems, and other technical problems may exist.

According to an aspect of the present invention, there is provided a method for generating shape feature information for a model, the method comprising: (a) dividing the model into a plurality of meshes; and (b) Wherein the shape feature information is generated by sampling the plurality of meshes with a plurality of samples constituting a pair of two meshes, wherein the shape feature information is present in each of the two meshes Lt; RTI ID = 0.0 > a < / RTI >

In addition, the mesh may be a triangular mesh.

The step (b) includes the steps of: (b1) arbitrarily selecting two sample meshes among the plurality of meshes, sampling the distance between arbitrary sample points located in each of the selected two sample meshes as one sample (B2) normalizing the distance values of each of the plurality of samples included in the sampling list to generate a normalized sampling list, and (b3) ) A step of allocating a normalized distance value of each of the plurality of samples by dividing the index into a plurality of indexes, and calculating the number of samples allocated to each index, thereby generating a feature vector corresponding to the primary shape feature information .

The plurality of indices may be set such that the normalized distance values of each of the plurality of samples are allocated by dividing the distance interval.

The step (b) may further include the step of (b4) generating a probability distribution-based histogram corresponding to the secondary shape feature information using the feature vector.

이며, 수학식 1에서 FV _i 는 상기 특징 벡터의 배열에서 i번째의 요소 값, n은 상기 특징 벡터의 배열의 크기, H _i 는 i번째의 히스토그램 값을 나타낸다.Further, the histogram may be generated by applying the feature vector to Equation (1). Here, Equation (1)

In Equation (1), FV _i represents the i- th element value in the array of feature vectors, n represents the size of the array of feature vectors, and H _i represents the i- th histogram value.

이며, 수학식 2에서, P는 샘플 메쉬 내부에 존재하는 임의의 샘플 포인트, A, B 및 C는 샘플 메쉬의 꼭지점의 좌표, r ₁ 은 꼭지점 A로부터 선분 BC까지의 비율, r ₂ 는 선분 BC와 평행한 샘플 메쉬 내부 선분의 비율을 나타낸다. Also, in the step (b), the arbitrary sample point may be generated based on Equation (2). Here, Equation (2)

And, in equation 2, P is the coordinate, r ₁ of the vertex of the sample mesh may be any sample point, A, B and C existing inside the sample mesh proportion to a line segment BC from vertices A, r ₂ is a line segment BC Of the sample mesh.

Meanwhile, the shape similarity analysis method according to an embodiment of the present invention includes the steps of (a) calculating the similarity between shape feature information of a new model and shape feature information of an existing model stored in a database, And the shape feature information of the existing model stored in the database can be generated by applying the shape feature information generation method described above.

The shape feature information of the new model or the shape feature information of the existing model may be obtained by sampling a plurality of samples of two meshes in a model divided into a plurality of meshes, A histogram based on a probability distribution based on a feature vector related to a distance between arbitrary sample points present, and the step (a) includes a step of calculating a probability distribution based histogram corresponding to the new model and a probability The similarity between the distribution-based histograms can be calculated.

In addition, the similarity between the histograms can be calculated based on the Minkowski distance metric.

Also, the shape similarity analysis method according to an embodiment of the present invention may further include: (b) extracting at least one existing model from the database based on the calculated similarity.

In addition, the shape similarity analyzing method according to an embodiment of the present invention may include: (c) applying scale points of the extracted existing model and the new model to the new model, As shown in FIG.

Also, the shape similarity analysis method according to an embodiment of the present invention may include: (d) an unused measurement point generated due to a shape difference between the extracted existing model and the new model among the measured points of the extracted existing model applied to the new model And judging whether or not it exists.

In addition, the shape similarity analysis method according to an embodiment of the present invention may further include the step of (e) correcting the unmeasured measurement point when it is determined that the unused measurement point exists.

Meanwhile, an apparatus for generating shape feature information for a model according to an exemplary embodiment of the present invention includes a division unit for dividing the model into a plurality of meshes, and a shape feature information generation unit for generating shape feature information based on the plurality of meshes Wherein the shape feature information is related to a distance between arbitrary sample points existing in each of the two meshes when the plurality of meshes are sampled into a plurality of samples constituting a pair of two meshes have.

Also, the shape similarity analyzing apparatus according to an embodiment of the present invention includes a similarity analyzing unit that calculates a similarity between shape feature information of a new model and shape feature information of an existing model stored in a database, The shape feature information of the existing model stored in the shape feature information generating device may be generated by applying the shape feature information generating device.

The above-described task solution is merely exemplary and should not be construed as limiting the present disclosure. In addition to the exemplary embodiments described above, there may be additional embodiments in the drawings and the detailed description of the invention.

According to the above-mentioned problem solving means of the present invention, the histogram is generated based on the shape feature information of the new model, and the measurement point of the existing model most similar to the new model is calculated by calculating the similarity between the histogram of the new model and the histogram of the existing model, It is possible to more effectively reduce the time required for specifying a measurement point or generating a measurement path for a new model.

According to the above-mentioned problem solving means, there is an effect that the reusability of the measurement point of the existing model can be enhanced by providing the shape feature information generation method and the shape similarity analysis method of the model.

According to the above-described problem solving means of the present invention, it is possible to reduce the time required for measurement in on-machine measurement (OMM) using a touch probe without being influenced by the skill of the operator, So that the quality can be maintained uniform.

However, the effects obtainable here are not limited to the effects as described above, and other effects may exist.

1 is a block diagram schematically showing a configuration of an apparatus for generating shape feature information according to an embodiment of the present invention.
FIG. 2 is a diagram schematically illustrating a process of generating shape feature information through a shape feature information generating apparatus according to an embodiment of the present invention.
3 is a diagram for explaining an example of sample point generation and sample generation in the shape feature information generation apparatus according to an embodiment of the present invention.
4 is a diagram illustrating a process of generating a feature vector from a sampling list in an apparatus for generating shape feature information according to an embodiment of the present invention.
5 is a diagram showing a schematic configuration of a shape similarity analyzer according to an embodiment of the present invention.
6A to 6C are views showing an example of similarity analysis in the shape similarity analyzer according to an embodiment of the present invention.
7 is a diagram illustrating a process of generating a measurement path of a new model based on a measurement point of an existing model applied to a new model in the shape similarity analysis apparatus according to an embodiment of the present invention.
8 is a diagrammatic representation of an overall flow of a method for generating a measurement path for a new model in accordance with one embodiment of the present application.
9 is a flowchart illustrating a method for generating shape feature information according to an exemplary embodiment of the present invention.
10 is a flowchart illustrating an operation of the shape similarity analysis method according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. It should be understood, however, that the present invention may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In the drawings, the same reference numbers are used throughout the specification to refer to the same or like parts.

Throughout this specification, when an element is referred to as being "connected" to another element, it is intended to be understood that it is not only "directly connected" but also "electrically connected" or "indirectly connected" "Is included.

It will be appreciated that throughout the specification it will be understood that when a member is located on another member "top", "top", "under", "bottom" But also the case where there is another member between the two members as well as the case where they are in contact with each other.

Throughout this specification, when an element is referred to as "including " an element, it is understood that the element may include other elements as well, without departing from the other elements unless specifically stated otherwise.

Through the analysis of the similarity between the histogram of the new model based on the shape feature information and the histogram of the existing model, we apply the feature points of the existing model, which is most similar to the new model, to the new model to improve the reusability of the existing model. To a technique that can more effectively shorten the time taken to generate a measurement path for a new model.

FIG. 1 is a block diagram schematically showing the configuration of an apparatus for generating shape characteristic information 100 according to an embodiment of the present invention.

Referring to FIG. 1, the shape feature information generating apparatus 100 according to an embodiment of the present invention can generate shape feature information for a model, and includes a division unit 110 and a shape feature information generation unit 120 can do.

Prior to the description, the shape feature information generation apparatus 100 according to an embodiment of the present invention can use a shape distribution graph algorithm to quantitatively compare the similarities among the 3D shape models. The shape distribution graph algorithm can be expressed as probability distribution histogram by using Euclidean distances between triangles in a shape model represented by a triangular mesh, and then generating feature vectors as array types. Through the shape similarity analyzer 500 to be described later, the degree of similarity between the two models can be calculated by using a value obtained by accumulating the distance deviation between the histogram graph elements of the existing model and the new model expressed in the histogram. The Shape distribution graph algorithm used in the present invention is a simple shape function capable of rapid similarity analysis. In addition, since the shape similarity analyzer 500 according to an embodiment of the present invention uses a feature vector that can express the geometric features between the shape surfaces in comparing the similarity between two shape models, it is necessary to match the coordinate systems of the two shape models And there is a robust effect on coordinate transformation compared to other similarity comparison methods.

The division unit 110 may divide the model into a plurality of meshes. Here, the mesh may be a triangular mesh, but is not limited thereto. Further, the model may be a 3D model, but is not limited thereto.

The shape feature information generation unit 120 can generate shape feature information of the model based on the plurality of divided meshes. Here, the shape feature information may relate to a distance between arbitrary sample points existing inside each of the two meshes when a plurality of meshes are sampled by a plurality of samples constituting a pair of two meshes. A more detailed description of the shape feature information generating apparatus 100 can be more easily understood with reference to FIG.

FIG. 2 is a diagram schematically illustrating a process of generating shape feature information through the shape feature information generating apparatus 100 according to an embodiment of the present invention.

Referring to FIG. 2, the shape feature information generating apparatus 100 according to an exemplary embodiment of the present invention may prepare (or load) a model for generating shape feature information in the division unit 110 (S21). Then, the partitioning unit 110 may divide the model into a plurality of meshes (S22). In step S22, the partitioning unit 110 may divide the model into a triangle mesh. Otherwise, the partitioning unit 110 may perform model triangulation in step S22.

Next, the shape feature information generation unit 120 randomly selects two sample meshes among a plurality of meshes segmented for the model, and calculates a Euclidean distance between arbitrary sample points located inside each of the two selected sample meshes (S23). The sampling process can be repeatedly performed corresponding to each of the plurality of samples to generate a sampling list (S24).

Specifically, the shape feature information generation unit 120 selects a sample mesh (S23a) from two or more of the plurality of meshes in a sampling process (S23) for extracting a geometric feature of the model, (S23b), and calculates the Euclidean distance between the two generated sample points (S23c).

At this time, before step S23a, the shape feature information generation unit 120 calculates the sum of the area of each of the divided meshes and the cumulative value of the area for each of the plurality of meshes (i.e., cumulative area value, Value) can be calculated.

Then, in step S23a, the shape feature information generation unit 120 assigns weights corresponding to the ratios of the areas of the plurality of meshes occupying the entire area to the respective meshes using the calculated cumulative area values, Can be used to select an arbitrary sample mesh.

Next, in step S23b, the shape feature information generation unit 120 generates random feature points of random (random) sample points for each of the selected arbitrary two sample meshes, so that all points in the sample mesh can be equally distributed in the shape distribution. P can be generated. At this time, the process of step S23b and step S23c can be more easily understood with reference to FIG.

FIG. 3 is a diagram for explaining an example of sample point generation and sample generation in the shape feature information generation apparatus 100 according to an embodiment of the present invention. Here, FIG. 3A shows an example of generating a sample point, and FIG. 3B shows an example of generating a sample.

Referring to FIG. 3A, the shape feature information generating unit 120 in the shape feature information generating apparatus 100 can generate an arbitrary sample point P inside one sample mesh based on the following equation (1) have.

[Equation 1]

Here, P represents the position of the sample point generated inside the sample mesh, and A, B, and C represent the vertex coordinates (x, y, z) of the sample mesh for the case where the sample mesh is a triangular mesh. Also, in generating arbitrary sample points, r ₁ and r ₂ may be generated using random and non-periodic pseudo-random number sequences (PRNS). Specifically, r ₁ means the ratio from the vertex A to the line segment BC, and r ₂ means the ratio of the inner line segment of the sample mesh (triangle) parallel to the line segment BC. Also, r ₁ and r ₂ mean a random number ranging from 0 to 1. Thus, according to r ₁ and r ₂ , an arbitrary sample point P can be generated within one sample mesh as shown in FIG. 3 (a).

After generating the sample points P for one sample mesh among the arbitrary two sample meshes selected, the shape feature information generating unit 120 generates the shape feature information P (x, y) of the arbitrary two sample meshes selected based on Equation A sample point P can be generated for one sample mesh.

Then, when two sample points located inside the two selected sample meshes, that is, two sample points generated on the surface of the model are P ₁ and P ₂ , respectively, then in step S 23 c, Unit 120 may calculate the Euclidean distance between two sample points (P ₁ , P ₂ ).

The shape feature information generating unit 120 in step S23 can sample the Euclidean distance between two sample points for arbitrarily selected two sample meshes as one sample through steps S23a to S23c, Can be performed. For example, referring to FIG. 3 (b), the shape feature information generating unit 120 may sample the Euclidean distance between sample points P ₁ and P ₂ for arbitrary two sample meshes as one sample , And the Euclidean distance between sample points P ₃ and P ₄ for any other two sample meshes can also be sampled as another sample.

The apparatus for generating shape characteristic information 100 according to an embodiment of the present invention may perform this sampling process a plurality of times. In this case, as the number of sampling times increases, the time required to perform the sampling increases while the accuracy of the existing models stored in the database, which is similar to the new model, is increased. On the other hand, if the number of sampling times is too small, it may not be possible to define a shape feature that can distinguish differences between different models. In consideration of this point, the shape feature information generating apparatus 100 according to an embodiment of the present invention can preferably perform the sampling process 1024 times, but is not limited thereto.

Therefore, the shape characteristic information generation unit 120 can determine whether or not the number of sampling times is equal to 1024 (S24) after step S23c. If it is determined that the number of sampling times does not meet 1024 times (S24-N), step S23a is again performed. If it is determined that the number of sampling times is equal to 1024 times (S24-Y) A sampling list for a sample (i.e., 1024 samples) can be generated (S25). Here, the sampling list may mean a list in which the Euclidean distance values of a plurality of samples are expressed in an array form.

Since the sampling list is a result of Euclidean distance between sample points, if the two models have the same feature, if the sizes of the two models are different from each other, the histogram is generated differently depending on the distance range of the sampling list So that an error may occur in calculating the similarity value. Therefore, the shape feature information generating apparatus 100 according to an exemplary embodiment of the present invention is configured to generate the shape feature information of a plurality of samples included in the sampling list so that robust similarity analysis can be performed on a scale, Each of the Euclidean distance values may be normalized (S26) to a range of 0 to 1 based on the following equation (2), whereby a normalized sampling list can be generated.

&Quot; (2) "

Here, X _i represents the distance value of the i-th sample among the plurality of samples, and X _Max And X _Min represent the maximum value and the minimum value in the sampling list, and X _{i, 0 to 1} represent the distance value of the i- th sample normalized in the range of 0 to 1.

Then, the shape feature information generation unit 120 divides the normalized distance values of each of the plurality of samples included in the normalized sampling list into a plurality of indexes, and calculates the number of samples assigned to each of the plurality of indexes The feature vector corresponding to the car shape feature information can be generated (S27). Here, the plurality of indices may be set such that the normalized distance values of each of the plurality of samples are allocated by dividing the distance section. In addition, the plurality of indexes may be set to 64 indexes by way of example, but the present invention is not limited thereto. Accordingly, the shape feature information generation unit 120 can generate 1024 distance values included in the sampling list as a feature vector of a one-dimensional array composed of 64 indices, using Equation (2). This can be more easily understood from FIG.

FIG. 4 is a diagram illustrating a process of generating a feature vector from a sampling list in the shape feature information generating apparatus 100 according to an embodiment of the present invention.

First, referring to FIG. 4A, the shape feature information generation unit 120 may generate a sampling list based on the Euclidean distance values of each of a plurality of samples (that is, 1024 samples). Here, Bin may be an ID of each of a plurality of samples. For example, the Euclidean distance value of the first sample (Bin ₁ ) is 22.4, the Euclidean distance value of the second sample (Bin ₂ ) is 23.2, and the Euclidean distance value of the third sample (Bin ₃ ) is 25.0.

Next, referring to FIG. 4B, the shape feature information generating unit 120 may normalize the distance values of each of the plurality of samples included in the sampling list based on Equation (2) A sampling list can be generated. In other words, the shape feature information generation unit 120 generates the feature vector for the model by applying Equation 2 to the distance value (i.e., the array element value of the sampling list) of each of the plurality of samples included in the sampling list can do. For example, the normalized Euclidean distance value of the first sample (Bin ₁ ) is 0.83, the normalized Euclidean distance value of the second sample (Bin ₂ ) is 0.88, and the normalized Euclidean distance value of the third sample (Bin ₃ ) Lt; / RTI >

Next, referring to FIG. 4C, the shape feature information generation unit 120 may convert each of the normalized Euclidean distance values of the plurality of samples into an index of a feature vector. For example, when generating 64 feature vectors in the shape feature information generating apparatus 100 according to an embodiment of the present invention, the shape feature information generating unit 120 may generate each of the plurality of normalized Euclidean distance values It can be converted into any one of the index 0 to the index 63. At this time, the shape feature information generating unit 120 may convert each of the normalized Euclidean distance values into an index of the feature vector based on Equation (3).

&Quot; (3) "

Here, m is an integer representing the size of the feature vector, and n is an integer representing the number of times of sampling. SL [i] represents the value of the i-th element of the normalized sampling list as a real number (i.e., the normalized Euclidean distance value of the i-th sample in the normalized sampling list), and the shape characteristic information generating unit 120 generates the SL [ i] value to the feature vector index value SL [i] by multiplying the normalized Euclidean distance value of the i-th sample by the value m.

Next, referring to FIG. 4 (d), the shape feature information generation unit 120 can generate a feature vector for the model based on the following equation (4).

&Quot; (4) "

Here, n denotes a sampling number as an integer, SL [i] denotes a value obtained by converting the normalized Euclidean distance value of the i-th sample calculated by the equation (3) to the feature vector index value, FV [k] is characterized in a k-th element of vector array, with other words, represents the element of the k-th index in the array of the plurality of indexes. In this case, since the integer k indicating the k- th index in the feature vector array is allocated from the SL [i], which is a real number, it may mean a rounded integer value.

The shape feature information generation unit 120 adds 1 to the element value of the index included in the feature vector array and adds the element value corresponding to the feature vector index value calculated for each of the plurality of samples in FIG. A feature vector corresponding to a plurality of indexes can be generated as shown in Fig. 4 (d). In other words, the shape feature information generating unit 120 in FIG. 4D can divide the normalized distance values of each of the plurality of samples into a plurality of indexes, and then assigns the number of samples allocated to each of the plurality of indexes The feature vector can be generated based on the feature vector.

For example, the shape feature information generation unit 120 may change the Euclidean distance value 23.2 in the sampling list to 0.88 through the normalization process in the case of the second sample (Bin ₂ ) of the plurality of samples. Then, the shape feature information generation unit 120 may convert the normalized Euclidean distance value 0.88 of the second sample (Bin ₂ ) to the index value 55 of the feature vector. Thereafter, the shape feature information generation unit 120 may add 1 to the element value of the 55 th index of the feature vector array. The shape feature information generation unit 120 may repeat the above method by the number of times corresponding to a plurality of samples (that is, 1024 times), and the feature vector corresponding to the primary shape feature information for the model Can be generated.

Through the above process, the shape feature information generation unit 120 can generate a one-dimensional array by dividing the normalized distance value of each of a plurality of samples (1024) into 64 intervals. In this case, a value corresponding to each element of the one-dimensional array means the number of specific distance values included in the corresponding section, that is, the number of samples allocated to each of the plurality of indices has a distance value within a distance range set in each index Means the number of samples, which can represent the shape characteristics of the model.

The shape feature information generator 120 may generate a probability distribution based histogram corresponding to the second shape feature information by using the feature vector of the generated model, in order to analyze the similarity between the two models described below. Here, the histogram can be generated by applying the generated feature vector to Equation (5).

&Quot; (5) "

Here, FV _i represents the i- th element value (i.e., the element value of the i-th index among the plurality of index arrays) in the feature vector array, n represents the size of the feature vector array, H _i represents the i- . For example, in FIG. 4 (d), FV _i may represent an element value of the i- th index at index 0 to index 63.

The shape feature information generation apparatus 100 according to an embodiment of the present invention may include a database 130. The shape feature information generation apparatus 100 may generate shape feature information for all models generated through the above- And the corresponding histogram may be stored in the database 130 for use in the similarity analysis described later.

Hereinafter, a method for analyzing the similarity between the shape feature information of the new model and the shape feature information of the existing model will be described based on the above description.

FIG. 5 is a diagram showing a schematic configuration of a shape similarity analyzer 500 according to an embodiment of the present invention.

Referring to FIG. 5, the shape similarity analyzer 500 according to an embodiment of the present invention may include a similarity analyzer 510.

The shape similarity analyzer 500 is a device separate from the shape feature information generating apparatus 100 in the embodiment of the present invention. However, the present invention is not limited to this, and the shape similarity analyzer 500, (100) may be implemented as a single device. Accordingly, the contents described with respect to the shape feature information generation apparatus 100 can be similarly applied to the shape similarity analysis apparatus 500, and the following description will be briefly described.

The similarity analyzer 510 may calculate the similarity by comparing the feature vector values representing the distances between arbitrary sample points specified on the surface of the model shape, which are generated for each of the existing model and the new model, in an array form . That is, the similarity analyzer 510 may calculate the similarity between the shape feature information of the new model and the shape feature information of the existing model stored in the database 130. Here, the shape feature information of the new model and the shape feature information of the existing model stored in the database 130 may be generated using the shape feature information generation apparatus 100. [

The shape feature information of the new model or the shape feature information of the existing model is obtained by sampling a plurality of samples of two meshes in a model divided into a plurality of meshes, And the similarity analyzer 510 calculates a similarity between the histogram based on the probability distribution based on the new model and the histogram based on the probability distribution based on the existing model, The degree of similarity can be calculated.

Specifically, the shape feature information of the existing models generated through the shape feature information generation apparatus 100 and the histogram based on the feature distribution information based thereon may be stored in the database 130. In order to extract a model having the highest degree of similarity with the new model among the existing models stored in the database 130 through the extracting unit 520 of the shape similarity analyzer 500, The feature information generating apparatus 100 may generate shape feature information for the new model and generate a probability distribution based histogram for the new model based on the generated feature feature information. Then, the similarity analyzer 510 can calculate the similarity between the histogram of the probability distribution based on the new model and the histogram of the probability distribution based on the existing models stored in the database 130.

The similarity analyzer 510 may calculate the similarity by comparing the distance between the histogram of the new model and the histogram of the existing model. Specifically, the similarity analyzer 510 can calculate the similarity by comparing the probability values corresponding to the respective Bin in the histogram in the histogram of the new model and the histogram of the existing model. At this time, the similarity analyzer 510 may calculate the similarity between the histograms based on the Minkowski distance metric that generalizes the distance between two points as shown in Equation (6).

&Quot; (6) "

Here, L _p (x, y) denotes a similarity value between the two histograms, x _i represents the value of the i-th in the histogram of the new model, y _i denotes the value of the i-th in the histogram of the original model, n is Feature Indicates the size of the vector array (in other words, the size of the Bin in the histogram, the index number). In addition, for example, the similarity value between two histograms can be calculated as a percentage (%) value, but is not limited thereto.

In Equation (6), p is a Manhattan distance in case of 1, and Euclidean distance in case of p is 2. In the shape similarity analyzer 500 according to an embodiment of the present invention, The similarity can be calculated using Manhattan distance for quick calculation. At this time, the Manhattan distance can be obtained as the absolute value of the difference between the coordinates of two points. Therefore, the similarity value between two histograms calculated through Equation (6) can be obtained by multiplying the histogram value of the new model calculated through Equation (5) and the histogram value of the existing model Can be calculated by summing up the difference values. This can be more easily understood from FIG. 6 (a) to FIG. 6 (c).

6A to 6C are diagrams showing an example of similarity analysis in the shape similarity analyzer 500 according to an embodiment of the present invention.

6A shows a result of comparing the degree of similarity between the new model (Ring1, query model) and the existing models stored in the database 130 (i.e., Ring2 to Ring3, Flanges1 and Bolt2) . Referring to FIG. 6A, for example, the similarity values to the new model may be 0.26593 for Ring3, 0.27343 for Rng2, 0.29296 for Ring4, 0.30273 for Flanges1, and 0.88671 for Bolt2. Here, for example, the lower the similarity value is, the higher the degree of similarity with the new model is. Accordingly, the extracting unit 520 extracts Ring3 as a model most similar to the new model among the existing models stored in the database 130, according to the extracting unit 520 described later.

On the other hand, the histogram trends are similar for the most similar model to the new model, and the difference area between the trajectories of the histograms between the two models may be small. A more detailed description follows.

6 (b) shows a histogram of Ring3, which is the most similar model to the new model Ring1. 6B, in the case of the histogram between Ring3 and the new model, which is the model with the highest similarity to the new model Ring1, the number of values distributed in a specific distance section (i.e., The similarity of histogram trends can be confirmed.

On the other hand, Fig. 6 (c) shows a histogram of Bolt2, which is a model not similar to the new model Ring1. 6 (c), in the case of the histogram between Bolt2, which is the model with the lowest similarity to the new model Ring1, and the new model, the difference of the histogram trajectory between the two models is larger than that between the new model of FIG. It can be confirmed that the difference is larger than the trajectory difference. According to FIG. 6, it can be seen that the area between the differences of the histogram trajectories is smaller in the model with higher similarity to the new model, and the similarity value is calculated to be close to 0.

Meanwhile, the shape similarity analyzing apparatus 500 according to an embodiment of the present invention may include an extracting unit 520.

The extracting unit 520 may extract at least one existing model from the database 130 based on the similarity calculated by the similarity analyzing unit 510. Specifically, the extracting unit 520 extracts a model most similar to the new model (i.e., the model having the lowest similarity value) among the existing models stored in the database 130, based on the similarity calculated by the similarity analyzing unit 510 Can be extracted. For example, in the case of FIG. 6, the extractor 520 may extract the Ring3 model among existing models stored in the database 130 as a similar model most similar to the new model.

The shape similarity analyzing apparatus 500 according to an exemplary embodiment of the present invention may include a measurement point applying unit 530 and the measurement point applying unit 530 may include an existing model extracted by the extracting unit 520 Model) and the scale of the new model can be adjusted in the same manner, and then the measurement point of the extracted existing model can be applied to the new model.

In addition, the shape similarity analyzing apparatus 500 according to an embodiment of the present invention includes a shape similarity analyzer 500 for analyzing an unexplained measurement point generated by a shape difference between an existing model and a new model extracted from measurement points of an existing model extracted from an extraction unit 520 applied to a new model (Not shown) for determining whether or not there is a non-usable measuring point. Here, in the shape similarity analyzer 500 according to an embodiment of the present invention, when it is determined that an unused measurement point exists, an unadjusted measurement point may be corrected. At this time, the modification to the unused measurement point may be modified by the user as an example. In another example, modifications to the unmeasured measurement points may be automatically corrected by applying known algorithms or future algorithms. The process after applying the measurement point of the existing model to the new model can be more easily understood from FIG.

FIG. 7 is a diagram illustrating a process of generating a measurement path of a new model based on a measurement point of an existing model applied to a new model in the shape similarity analysis apparatus 500 according to an embodiment of the present invention.

Referring to FIG. 7, the shape similarity analyzing apparatus 500 according to an embodiment of the present invention extracts an existing model (that is, a similar model) most similar to the new model through the extracting unit 520, From the database 130 (S71), and applies the measurement point of the similar model to the new model (S72). At this time, the database 130 may store information about measurement points corresponding to the models in addition to the shape feature information (i.e., feature vector) of the models and the corresponding histogram.

At this time, in step S72, an unused measurement point may be generated on the new model due to the shape difference between the two models. Therefore, in step S73, it is possible to determine whether an unused measurement point exists on the new model through the unused measurement point determination unit (not shown).

At this time, the unused measurement point can be generated as the measurement point is applied to the unintended position due to the scale difference between the existing model and the new model. Accordingly, when applying the measurement point of the existing model to the new model in step S72, the shape similarity analyzing apparatus 500 according to an embodiment of the present invention calculates the scale of the overall shape or feature shapes of the existing model After adjusting the same as the new model, the adjusted point of the existing model can be applied to the new model. Here, the scale adjustment between the new model and the existing model can be adjusted through various known techniques.

In addition, in step S72, even though the measurement point of the existing model and the scale of the new model are adjusted to the same, and the adjusted measurement point of the existing model is applied to the new model, the shapes of the existing model and the new model do not perfectly match An unused measuring point may be found. Accordingly, the shape similarity analyzing apparatus 500 according to an embodiment of the present invention corrects the unmeasured measuring point on the new model in accordance with the size and type, based on the feature of the new model in which the unused measuring point is located, after step S73 (S74). At this time, modification or change of the unused measurement point can be corrected based on user input through an editing interface (not shown) provided through the shape similarity analyzing apparatus 500, for example.

Thereafter, in step S75, the shape similarity analyzing apparatus 500 can generate a measurement path for the new model based on the measurement point of the existing model applied to the new model and the corrected unadjusted measurement point. At this time, in generating the measurement path for the new model, the shape similarity analyzer 500 can additionally generate a safety point and an entry point so that the probe can smoothly approach the measurement point without collision, Accordingly, the shape similarity analyzer 500 can generate the measurement path for the new model by considering the measurement point and the non-recognition point and the entry point applied to the new model through step S74.

Next, the shape similarity analyzer 500 may generate a measurement path for the new model, and then perform a collision check simulation (S76) to verify the generated measurement path.

Hereinafter, the overall flow of a method of generating a measurement path for a new model will be briefly described with reference to FIG.

8 is a diagrammatic representation of an overall flow of a method for generating a measurement path for a new model in accordance with one embodiment of the present application.

8, a measurement path generation method for a new model may be performed through the shape feature information generation apparatus 100 and the shape similarity analysis apparatus 500 described above. Therefore, the contents described with respect to the shape feature information generation apparatus 100 and the shape similarity analysis apparatus 500 can be similarly applied to the description of the measurement path generation method for the new model, even if omitted below.

Referring to FIG. 8, when a new model is input (S81), a model triangulation process, a sampling process, and a sampling normalization process for a new model are performed through the partitioning unit 110 and the shape feature information generating unit 120 (I.e., shape feature information) with respect to the new model.

Next, in step S83, the shape feature information generation unit 120 can generate a histogram for the new model based on the feature vector for the new model.

Next, in step S84, the similarity analyzer 510 calculates the similarity between the existing model and the new model by comparing the histogram of the existing models based on the feature vectors of the existing models stored in the database 130 with the histogram of the new model . After step S84, the extracting unit 520 may extract at least one existing model from the database 130 as a similar model most similar to the new model, based on the similarity calculated by the similarity analyzing unit 510. [

Next, in step S85, the measurement point application unit 530 may apply the measurement point of the existing model extracted by the extraction unit 520 to the new model.

Next, in step S86, the unused measurement point determination unit (not shown) can determine whether or not an unused measurement point exists on the new model. At this time, in the step S86, when it is judged that there is an unused measurement point, the feature of the new model in which the unused measurement point is located is grasped, and the unmeasured measurement point on the new model is corrected (Change). At this time, the modification or change of the unused measurement point may be modified based on the user input, and the number or position of the measurement point with respect to the measurement point of the existing model applied on the new model may be modified as well as the position of the unused measurement point .

Next, when the modification to the unused measurement point is completed in step S86, the measurement path for the new model can be generated based on the measurement point of the existing model and the corrected unadjusted measurement point applied to the new model in step S87.

Next, in step S88, collision check simulation may be performed for verification of the measurement path for the generated new model.

Conventionally, in a gas phase measurement using a touch probe, the time required for measurement can be varied according to the number of measurement points and the skill of the operator. Particularly, in the past, when a plurality of objects (models) are measured, the measurement points need to be repeatedly generated for each model in spite of the similar shapes of the objects. In order to solve this problem, a method of generating a measurement path for a new model according to an embodiment of the present invention includes a step of calculating a feature vector of an existing 3D shape model previously built in the database 130 and a feature vector of a new model By performing the similarity analysis based on the probability distribution based histogram, the measurement points of the existing model extracted based on the similarity analysis can be applied (or reused) to the new model. Accordingly, a method of generating a measurement path for a new model according to an exemplary embodiment of the present invention can effectively shorten the time taken to generate a measurement point for a new model and provide a measurement quality at a certain level It is effective.

Hereinafter, the operation flow of the present invention will be briefly described based on the details described above.

9 is a flowchart illustrating a method for generating shape feature information according to an exemplary embodiment of the present invention.

The shape feature information generating method shown in FIG. 9 can be performed by the shape feature information generating apparatus 100 described above. Therefore, even if omitted below, the contents described with respect to the shape feature information generating apparatus 100 can be similarly applied to the shape feature information generating method.

Referring to Fig. 9, in step S91, the model can be divided into a plurality of meshes. Here, the mesh may be a triangular mesh.

Next, in step S92, the shape feature information can be generated based on the plurality of meshes segmented in step S91. Here, the shape feature information may relate to a distance between arbitrary sample points existing inside each of the two meshes when a plurality of meshes are sampled by a plurality of samples constituting a pair of two meshes.

In addition, in step S92, two sample meshes among a plurality of meshes are arbitrarily selected, and sampling of the Euclidean distances between arbitrary sample points located inside each of the selected two sample meshes as one sample is performed for each of a plurality of samples So that the sampling list can be generated. In addition, in step S92, the distance value of each of the plurality of samples included in the sampling list may be normalized to generate a normalized sampling list. In step S92, the normalized distance value of each of the plurality of samples is divided into a plurality of indexes, and the number of samples allocated to each index is calculated to generate a feature vector corresponding to the primary shape feature information. Here, the plurality of indices may be set such that the normalized distance values of each of the plurality of samples are allocated by dividing the distance section.

In step S92, a probability distribution based histogram corresponding to the secondary shape feature information can be generated using the feature vector. Here, the histogram can be generated by applying the feature vector to Equation (1).

Also, in step S92, any sample point may be generated using Pseudo-Random Number Sequences (PRNS).

In the above description, steps S91 and S92 may be further divided into further steps or combined in fewer steps, according to embodiments of the present disclosure. Also, some of the steps may be omitted as necessary, and the order between the steps may be changed.

FIG. 10 is a flowchart illustrating an operation of the shape similarity analysis method according to an embodiment of the present invention.

The shape similarity analysis method shown in FIG. 10 can be performed by the shape similarity analysis apparatus 500 described above. Therefore, even if the contents are omitted in the following description, the contents described for the shape similarity analyzer 500 can be similarly applied to the shape similarity analysis method.

Referring to FIG. 10, in step S101, it is possible to calculate the similarity between the shape feature information of the new model and the shape feature information of the existing model stored in the database. At this time, the shape feature information of the new model and the shape feature information of the existing model stored in the database can be generated by applying the shape feature information generation method.

When the model feature information of the new model or the shape feature information of the existing model is sampled by a plurality of samples constituting a pair of two meshes in a model divided into a plurality of meshes, A histogram based on a probability distribution based on a feature vector relating to a distance between arbitrary sample points. In step S101, the similarity between the histogram based on the probability distribution based on the new model and the histogram based on the probability distribution corresponding to the existing model is calculated can do. Here, the similarity between the histograms can be calculated based on the Minkowski distance metric.

Next, in step S102, at least one existing model may be extracted from the database based on the similarity calculated in step S101.

Next, in step S103, the scale of the existing model and the new model extracted in step S102 may be adjusted to the same, and then the measurement point of the extracted existing model may be applied to the new model.

After step S103, it is determined whether there is an unused measurement point generated due to the shape difference between the extracted model and the new model extracted from the extracted measurement points of the existing model applied to the new model. If it is determined that an unapplied measuring point is present, a step may be taken where a correction to the unapplied measuring point is made.

In the above description, steps S101 to S103 may be further divided into additional steps or combined into fewer steps, according to embodiments of the present disclosure. Also, some of the steps may be omitted as necessary, and the order between the steps may be changed.

The shape feature information generation method and the shape similarity analysis method according to an embodiment of the present invention may be implemented in the form of a program command which can be executed through various computer means and recorded in a computer readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the medium may be those specially designed and constructed for the present invention or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks, and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

Further, the above-described shape feature information generation method and shape similarity analysis method may be implemented in the form of a computer program or an application executed by a computer stored in a recording medium.

It will be understood by those of ordinary skill in the art that the foregoing description of the embodiments is for illustrative purposes and that those skilled in the art can easily modify the invention without departing from the spirit or essential characteristics thereof. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. For example, each component described as a single entity may be distributed and implemented, and components described as being distributed may also be implemented in a combined form.

The scope of the present invention is defined by the appended claims rather than the detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalents should be interpreted as being included in the scope of the present invention.

100: shape feature information generation device
110: minute installment
120: shape feature information generating unit
130: Database

Claims (17)

A method for generating shape feature information for a model,
(a) dividing the model into a plurality of meshes; And
(b) generating shape feature information based on the plurality of meshes,
, ≪ / RTI &
Wherein the shape feature information relates to a distance between arbitrary sample points existing in each of the two meshes when the plurality of meshes are sampled into a plurality of samples constituting a pair of two meshes,
The step (b)
(b1) arbitrarily selecting two sample meshes among the plurality of meshes, and sampling, as one sample, the distance between arbitrary sample points located in each of the two selected sample meshes, corresponding to each of the plurality of samples Repeatedly generating a sampling list;
(b2) normalizing a distance value of each of the plurality of samples included in the sampling list to generate a normalized sampling list;
(b3) assigning a normalized distance value of each of the plurality of samples to a plurality of indexes, calculating a number of samples allocated to each index, and generating a feature vector corresponding to the primary shape feature information; And
(b4) generating a probability distribution based histogram corresponding to the secondary shape feature information using the feature vector,
Wherein the histogram is generated by applying the feature vector to Equation (1).
[Equation 1]

Here, FV _i denotes an i- th element value in the array of feature vectors, n denotes the size of the array of feature vectors, and H _i denotes an i- th histogram value. The method according to claim 1,
Wherein the mesh is a triangular mesh. delete The method according to claim 1,
Wherein the plurality of indexes are set so that the normalized distance values of each of the plurality of samples are allocated by dividing the distance interval. delete delete The method according to claim 1,
And in the step (b), the arbitrary sample point is generated based on the following equation (2).
&Quot; (2) "

Where P is an arbitrary sample point existing in the sample mesh, A, B and C are the coordinates of the vertices of the sample mesh, r ₁ is the ratio from the vertex A to the line segment BC, r ₂ is the sample mesh parallel to the line segment BC, Indicates the percentage of internal segments. In the shape similarity analysis method,
(a) calculating the similarity between the shape feature information of the new model and the shape feature information of the existing model stored in the database,
Wherein the shape feature information of the new model and the shape feature information of the existing model stored in the database are generated by applying the shape feature information generation method of claim 1. 9. The method of claim 8,
Wherein the shape feature information of the new model or the shape feature information of the existing model is sampled from a plurality of samples constituting a pair of two meshes in a model divided into a plurality of meshes, A probability distribution based histogram generated based on a feature vector relating to a distance between arbitrary sample points,
Wherein the step (a) calculates a degree of similarity between a probability distribution based histogram corresponding to the new model and a probability distribution based histogram corresponding to the existing model. 10. The method of claim 9,
Wherein the similarity between the histograms is calculated based on a Minkowski distance metric. 9. The method of claim 8,
(b) extracting at least one existing model from the database based on the calculated similarity. 12. The method of claim 11,
(c) adjusting the scale between the extracted existing model and the new model to the same, and then applying the extracted measurement point of the existing model to the new model. 13. The method of claim 12,
(d) judging whether or not an unused measurement point generated by a shape difference between the extracted existing model and the new model among the measurement points of the extracted existing model applied to the new model exists . 14. The method of claim 13,
(e) if the unmeasured measurement point is determined to be present, the correction to the unmeasured measurement point is performed. A shape feature information generating apparatus for a model,
A dividing unit dividing the model into a plurality of meshes; And
A shape feature information generating unit for generating shape feature information based on the plurality of meshes,
, ≪ / RTI &
Wherein the shape feature information relates to a distance between arbitrary sample points existing in each of the two meshes when the plurality of meshes are sampled into a plurality of samples constituting a pair of two meshes,
Wherein the shape feature information generating unit comprises:
Sampling two sample meshes out of the plurality of meshes and sampling a distance between arbitrary sample points located inside each of the two selected sample meshes as one sample is repeated corresponding to each of the plurality of samples Generate a sampling list,
Normalizing a distance value of each of the plurality of samples included in the sampling list to generate a normalized sampling list,
A normalized distance value of each of the plurality of samples is divided into a plurality of indexes, a number of samples allocated to each index is calculated, a feature vector corresponding to the primary shape feature information is generated,
Generating a histogram based on a probability distribution corresponding to the secondary shape feature information using the feature vector,
And the histogram is generated by applying the feature vector to the following equation (5).
&Quot; (5) "

Here, FV _i denotes an i- th element value in the array of feature vectors, n denotes the size of the array of feature vectors, and H _i denotes an i- th histogram value. In the shape similarity analyzer,
And a similarity analyzer for calculating the similarity between the shape feature information of the new model and the shape feature information of the existing model stored in the database,
Wherein the shape feature information of the new model and the shape feature information of the existing model stored in the database are generated by applying the shape feature information generating apparatus of claim 15. A computer-readable recording medium storing a program for executing the method of any one of claims 1, 2, 4, and 14 in a computer.

Images