CN108876711B - Sketch generation method, server and system based on image feature points - Google Patents

Abstract

The invention discloses a sketch generation method, a server and a system based on image feature points. The sketch generation method based on the image feature points comprises the following steps: carrying out preprocessing operation on the obtained entity image; extracting feature points of the edges of the preprocessed entity images based on a wavelet self-adaptive Harris detection operator; and (3) fitting the contour of the extracted characteristic points of the edge of the entity image by adopting a segmented spline interpolation method to obtain a smooth and closed curve contour and generate a corresponding sketch. The method can improve the design efficiency and meet the use requirements of common designers.

Description

A sketch generation method, server and system based on image feature points

technical field

The invention belongs to the field of three-dimensional model modeling and design, and in particular relates to a sketch generation method, server and system based on image feature points.

Background technique

With the continuous advancement of science and technology, people's interest in 3D models has gradually increased, and 3D modeling has become a hot topic in the field of computer research. Whether it is in 3D game design, machinery manufacturing, smart home, or medical treatment, 3D models will be involved in different degrees. However, with the continuous improvement of customer demand, the efficiency of model design is also accelerating. In the process of 3D model design, the drawing of sketches is an important link. Sketches are drawn more efficiently.

There have been many researches in this area. For example, Peng et al. analyzed the hand-drawn sketch model through information entropy analysis and constructed a sketch descriptor. Liu Kai, Sun Zhengxing and others proposed an interactive method for generating three-dimensional complex surfaces, which supports users The shape of the 3D model is constructed by modifying the hand-drawn sketch, and the relevant personnel complete the drawing of the sketch through the constraint mechanism of the outline geometric feature recognition method based on the Hough transform method and the random circle detection method (RCD).

But there are still the following problems:

1) The sketch descriptor is drawn through information entropy analysis, and the structure is relatively simple. In terms of drawing, the sketching method is less efficient;

2) Although the geometric feature outline can be identified by the Hough transform method and the random circle detection method, the basic features of straight lines and arcs are all recognized. It is not suitable for curved wireframe models. For example: for graphics with curved inflection points and uneven graphics, the Hough transform method and random circle detection method can only detect straight lines and random lines in the graphics, and the detection results are not very ideal. However, the present invention directly processes on the image, uses wavelet-based adaptive Harris detection operator to directly extract image edge points to further extract inflection points and feature points, and the detection result is more definite.

3) The mainstream sketching software involves many tools, and it is difficult for ordinary users to draw related models in a short time.

To sum up, in the actual design process, as the needs of customers gradually increase, how to speed up the drawing efficiency of sketches and improve the design work of models is the key problem that designers need to solve.

Contents of the invention

In order to solve the deficiencies of the prior art, the first object of the present invention is to provide a sketch generation method based on image feature points, which reduces time complexity and improves drawing efficiency.

A method for generating sketches based on image feature points of the present invention, comprising:

Perform preprocessing operations on the acquired entity images;

Based on the wavelet adaptive Harris detection operator to extract the feature points of the edge of the preprocessed solid image;

The feature points on the edge of the extracted solid image are fitted with the segmented spline interpolation method to obtain a smooth and closed curve profile, and the corresponding sketch is generated.

Further, the preprocessing operation on the acquired entity image includes noise reduction processing and boundary tracking of the entity image.

Since the image may be degraded by noise pollution during the transmission process, the first task is to denoise the image.

Further, the acquired entity image is denoised by sequentially adopting the median filter denoising method and the dilation operation method.

In this way, the clarity of the image can be guaranteed while the noise of the image is removed.

Further, the specific process of extracting the feature points of the edge of the preprocessed entity image based on the wavelet adaptive Harris detection operator includes:

Convert the preprocessed entity image into a binary image;

Calculate the horizontal and vertical components of the binary image on the preset scale, and construct the modulus and argument of the wavelet coefficient;

Calculate the gradient of the binary image in the horizontal and vertical directions according to the modulus and argument of the constructed wavelet coefficients;

According to the gradient of the binary image in the horizontal and vertical directions, use the Gaussian function to Gaussian weight the feature points, and add the Euclidean distance at the same time, calculate the difference between the points, and generate a matrix element value sequence;

In the matrix element value sequence, the corner point is further determined by calculating the Harris response value of each image pixel, and the feature point of the edge of the image is extracted by a wavelet-based adaptive Harris detection operator pair.

The second object of the present invention is to provide a sketch generation server based on image feature points, which reduces time complexity and improves drawing efficiency.

A kind of sketch generation server based on image feature points of the present invention comprises:

A preprocessing module, which is configured to: perform a preprocessing operation on the acquired entity image;

The edge feature point extraction module is configured to: extract the feature points of the edge of the preprocessed entity image based on the wavelet adaptive Harris detection operator;

The edge feature point fitting module is configured to: use a segmented spline interpolation method to fit the outline of the extracted feature points on the edge of the solid image to obtain a smooth and closed curve outline, and generate a corresponding sketch.

Further, in the preprocessing module, performing preprocessing operations on the acquired entity image includes noise reduction processing and boundary tracking of the entity image.

Further, in the preprocessing module, the acquired entity image is denoised by sequentially adopting a median filter denoising method and an expansion operation method.

Further, the edge feature point extraction module includes:

The binary image conversion submodule is configured to: convert the preprocessed entity image into a binary image;

The modulus and argument construction submodule of the wavelet coefficient is configured to: calculate the horizontal and vertical components of the binary image on a preset scale, and construct the modulus and argument of the wavelet coefficient;

The horizontal and vertical direction gradient calculation sub-module is configured to: calculate the binary image gradient in the horizontal and vertical directions according to the modulus and argument of the constructed wavelet coefficients;

The matrix element value sequence generation sub-module is configured to: use the Gaussian function to Gaussian weight the feature points according to the gradient of the binary image in the horizontal and vertical directions, and add the Euclidean distance at the same time, calculate the difference between the points, and generate a sequence of matrix element values;

The feature point extraction submodule is configured to: in the matrix element value sequence, further determine the corner point by calculating the Harris response value of each image pixel, and extract the image edge by using a Harris detection operator based on wavelet adaptation feature points.

The third object of the present invention is to provide a sketch generation system based on image feature points, which reduces time complexity and improves drawing efficiency.

A sketch generation system based on image feature points according to the present invention includes the aforementioned sketch generation server based on image feature points.

Compared with prior art, the beneficial effect of the present invention is:

(1) The present invention extracts feature points according to the solid image, controls the Harris corner detection operator by using the Euclidean distance difference, and outlines the contour edge in combination with the segmental spline interpolation algorithm, and then completes the design of the sketch, reducing the time complexity, Improved drawing efficiency.

(2) From the perspective of the image, the present invention detects and processes the key points of the image according to the corner point detection algorithm, and completes the sketch generation work through the segmented spline interpolation fitting, which has a good sketch drawing effect, can improve the design efficiency, and satisfies the Common designer use.

Description of drawings

The accompanying drawings constituting a part of the present application are used to provide further understanding of the present application, and the schematic embodiments and descriptions of the present application are used to explain the present application, and do not constitute improper limitations to the present application.

FIG. 1 is a flow chart of a sketch generation method based on image feature points in the present invention.

Fig. 2(a) is structural element A.

Figure 2(b) is the processed object X.

Figure 2(c) is the result of the processed object X after the dilation operation.

Figure 3(a) is the original graph of the dilation operation.

Figure 3(b) is a diagram of the results of one expansion operation.

Figure 3(c) is a graph of the results of the second expansion operation.

Figure 3(d) is a diagram of the results of three expansion operations.

Figure 4(a) is the initial outline of the flower.

Figure 4(b) is the initial outline of the monkey.

Figure 5(a) is the initial contour feature points of the flower.

Figure 5(b) is the initial contour feature points of the monkey.

Figure 6(a) is a schematic diagram of B-spline fitting.

Figure 6(b) is a schematic diagram of least squares fitting.

Fig. 6(c) is a schematic diagram of cubic sub-spline interpolation fitting.

Figure 7(a) is a sketch of the flower obtained by cubic sub-spline interpolation fitting.

Figure 7(b) is the sketch of the monkey obtained by cubic sub-spline interpolation fitting.

FIG. 8 is a schematic structural diagram of a sketch generation server based on image feature points in the present invention.

Detailed ways

It should be pointed out that the following detailed description is exemplary and intended to provide further explanation to the present application. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

It should be noted that the terminology used here is only for describing specific implementations, and is not intended to limit the exemplary implementations according to the present application. As used herein, unless the context clearly dictates otherwise, the singular is intended to include the plural, and it should also be understood that when the terms "comprising" and/or "comprising" are used in this specification, they mean There are features, steps, operations, means, components and/or combinations thereof.

FIG. 1 is a flow chart of a sketch generation method based on image feature points in the present invention.

As shown in Figure 1, a kind of sketch generation method based on image feature point of the present invention comprises:

S1: Perform a preprocessing operation on the acquired entity image.

The preprocessing operation on the acquired entity image includes noise reduction processing and boundary tracking of the entity image.

Since the image may be degraded by noise pollution during the transmission process, the first task is to denoise the image.

The obtained solid image is denoised by using the median filter denoising method and the dilation operation method in turn.

In this way, the clarity of the image can be guaranteed while the noise of the image is removed.

As shown in Figure 2(a)-Figure 2(c), the expansion operation method includes:

Translate the structural element A shown in Figure 2(a) by a to get Aa, if Aa hits X;

As shown in Figure 2(b), record point a, which is the shaded point in the figure.

The set of points a that satisfy the above conditions is called the result of X being expanded by A. Expressed as: D(X)={a|Aa↑X}=X A, as shown in Figure 2(c), where X is the object to be processed, A is the structural element, for any one in the shaded part At point a, Aa hits X, and the shaded part is the result of X being expanded by A.

The results of the expansion operation are compared, as shown in Figure 3(a)-Figure 3(d), the boundary points tracked above are subjected to median filtering processing and the boundary expansion is performed through the expansion operation to eliminate the pixels contained in the image target area fine noise.

S2: Based on the wavelet adaptive Harris detection operator, the feature points of the edge of the preprocessed solid image are extracted.

The specific process of extracting the feature points of the edge of the preprocessed solid image based on the wavelet adaptive Harris detection operator includes:

Step S21: It is necessary to construct a wavelet function. Let Φ(x,y) be a two-dimensional smooth function, then the first-order derivatives along the x and y directions are defined as two basic wavelets

Step S22: take the convolution of the basic wavelet and the image in two directions as the components in the horizontal and vertical directions of the wavelet, then any two-dimensional image function f(x, y) has two wavelet transform components. Servings are:

The component along the y-vertical direction is:

Where wt ₁ and wt ₂ are the gradient values of the image along the x and y directions, where b is a constant.

Step S23: Calculate the image modulus (formula 1) and argument (formula 2) according to the wavelet direction scale j, and perform corner detection along the direction of the local maximum value of the gradient modulus.

和

分别指的是在不同方向梯度上的模值。in,

and

Respectively refer to the modulus value on the gradient in different directions.

Specifically, the specific process of extracting the feature points of the edge of the preprocessed solid image based on the wavelet adaptive Harris detection operator includes:

Convert the preprocessed entity image into a binary image;

Calculate the horizontal and vertical components of the binary image at scale i, and construct the modulus and argument of the wavelet coefficients;

Calculate the gradient of the binary image in the horizontal and vertical directions according to the modulus and argument of the constructed wavelet coefficients;

Among them, in formula (3), V _m,n is defined as the coefficient of the Gaussian window at (m,n), and A and B are the horizontal and vertical components along the m and n directions respectively; (m,n) means Gaussian Coordinate coefficients for the window's horizontal and vertical components.

生成矩阵X元素值序列；Based on the determination of the horizontal and vertical components, use the Gaussian function to Gaussian weight the feature points, and add the Euclidean distance at the same time, and perform the difference between points (denoted by d, where

Generates a sequence of matrix X element values;

In the matrix X, the corner point is further determined by calculating the Harris response value K of each image pixel, and the image is detected by the Harris detection operator based on wavelet adaptation.

K(m,n)＝det[X(m,n)]-Y·tr ² [X(m,n)] (4)

为矩阵X的特征值；Y取推荐值0.05；In the formula (4), det[X(m,n)] represents the value of the determinant of the X matrix, and tr ² [X(m,n)] represents the trace of the matrix X;

is the eigenvalue of the matrix X; the recommended value of Y is 0.05;

和

都比较突出时，即值比较大时确定为角点；而当

和

值小于预设阈值时，不认为是角点；当其中

或者是

时，确定为边缘。Among them, the image points are divided by the size of the X feature value, when

and

are relatively prominent, that is, when the value is relatively large, it is determined as a corner point; and when

and

When the value is less than the preset threshold, it is not considered as a corner; when

or

, it is determined as an edge.

S3: Using the segmental spline interpolation method to fit the outline of the feature points on the edge of the extracted solid image, obtain a smooth and closed curve outline, and generate a corresponding sketch.

For example: using cubic spline interpolation processing:

Step S31: After detecting the contour feature points of the solid image, fitting the contour feature points by cubic spline function.

Step S32: Import the above detected and extracted data points into origin for segment processing, sort x according to the coordinate positions of the data points, put the results into the array c, and iteratively process until all the key points have sequence.

Step S33: Generate an n*1 matrix according to the length n of the array c.

插值拟合，要使h₀＝0则j＝1；因此，将y和x转变为与h有关的函数插值点；其中，h_j表示样条插值上的特征点；h_j-1表示h_j前的插值点；x_j-1,y_j-1分别表示在h_j-1插值点的横纵坐标；x_j,y_j分别表示h_j插值点的横纵坐标。According to the cubic spline interpolation formula,

Interpolation fitting, to make h ₀ =0 then j=1; therefore, transform y and x into function interpolation points related to h; where, h _j represents the feature point on the spline interpolation; h _j-1 represents h The interpolation point before _j ; x _j-1 and y _j-1 represent the horizontal and vertical coordinates of the interpolation point at h _j-1 respectively; x _j and y _j represent the horizontal and vertical coordinates of the interpolation point of h _j respectively.

Step S34: According to the key points and smoothness, sequentially call the data points in the array c, perform the fitting operation through the cubic sub-spline interpolation method, fit a smooth spline curve, and finally obtain the complete outline of the sketch, and intersect in Designer review process.

A practical application example is given below:

Utilize the photograph of flower and monkey as example to illustrate the process of drafting and the superiority of the present invention.

Step 1: Process on the two images given by the customer. In order to facilitate the processing of image denoising and enhancement, the two example images are converted into grayscale images. Using the boundary tracking algorithm, find the starting point of the tracking in the gray image, and start from this point to track the boundary in the eight neighborhoods, and record its position coordinates after expansion operation and coordinates.

Step 2: For comparison and analysis, carry out sampling preprocessing at equidistant intervals, and record the position coordinates of the points and the number of sampling points, as shown in Figure 4(a), where the number of initial contour points of flowers is 3804 , the response time is 2.632207s; while the number of initial contour points of the monkey shown in Figure 4(b) is 2295, and the response time is 1.229416s.

Step 3: By adding wavelet function and Euclidean distance difference to the traditional Harris algorithm, the present invention can ensure that the local features of the image contour are complete, and obtain image feature points in a short time as shown in Figure 5(a)-Figure 5(b) shown.

Step 4: Cubic segmental spline interpolation processing.

The results of B-spline fitting, least squares fitting and cubic piecewise spline interpolation fitting are shown in Figure 6(a)-Figure 6(c).

Fitting with cubic sub-spline interpolation, the final sketch of the flower is shown in Figure 7(a), and the sketch of the monkey is shown in Figure 7(b).

In this example, taking the pictures of flowers and monkeys as an example, several corner detection algorithms are selected and compared. The detailed comparison results are shown in Table 1 and Table 2. The experimental results show that the present invention has certain advantages. superiority.

Table 1 Flower detection results

The original image Moravec algorithm Fast Algorithm Algorithm initial sampling point 3804 17 524 270 Response time 2.632207s 7.430317s 2.915905s 1.028065s

Table 2 monkey detection results

The original image Moravec algorithm Fast Algorithm Algorithm initial sampling point 2295 40 762 351 Response time 1.229416s 2.225066s 1.291240s 1.404667s

As can be seen from the table, the Harris corner point detection example figure key point number based on wavelet that the present invention proposes is 270, and the response time is 1.028065s; The Harris corner point detection example monkey figure key point number based on wavelet There are 351, and the response time is 1.404667s. In terms of time, accurate feature points can be obtained in less time, which saves time and improves efficiency for designers to draw sketches manually.

The invention also provides a sketch generation server based on image feature points, which reduces time complexity and improves drawing efficiency.

As shown in Figure 8, a kind of sketch generation server based on image feature points of the present invention includes:

(1) A preprocessing module, which is configured to: perform a preprocessing operation on the acquired entity image.

Specifically, in the preprocessing module, performing preprocessing operations on the acquired entity image includes noise reduction processing and boundary tracking of the entity image.

Specifically, in the preprocessing module, the acquired entity image is denoised by sequentially adopting a median filter denoising method and an expansion operation method.

(2) An edge feature point extraction module, which is configured to: extract feature points of the edge of the preprocessed entity image based on a wavelet adaptive Harris detection operator.

Specifically, the edge feature point extraction module includes:

The binary image conversion submodule is configured to: convert the preprocessed entity image into a binary image;

The modulus and argument construction submodule of the wavelet coefficient is configured to: calculate the horizontal and vertical components of the binary image on a preset scale, and construct the modulus and argument of the wavelet coefficient;

The horizontal and vertical direction gradient calculation sub-module is configured to: calculate the binary image gradient in the horizontal and vertical directions according to the modulus and argument of the constructed wavelet coefficients;

The matrix element value sequence generation sub-module is configured to: use the Gaussian function to Gaussian weight the feature points according to the gradient of the binary image in the horizontal and vertical directions, and add the Euclidean distance at the same time, calculate the difference between the points, and generate a sequence of matrix element values;

The feature point extraction submodule is configured to: in the matrix element value sequence, further determine the corner point by calculating the Harris response value of each image pixel, and extract the image edge by using a Harris detection operator based on wavelet adaptation feature points.

(3) The edge feature point fitting module is configured to: use the segmented spline interpolation method to fit the outline of the extracted feature points on the edge of the solid image, obtain a smooth and closed curve outline, and generate a corresponding sketch.

The invention extracts feature points according to the solid image, controls the Harris corner detection operator by using the Euclidean distance difference, and combines the segmental spline interpolation algorithm to outline the outline edge, and then completes the design of the sketch, reduces the time complexity, and improves the drawing efficiency.

From the perspective of image, the present invention detects and processes the key points of the image according to the corner point detection algorithm, and completes the sketch generation work through segmental spline interpolation fitting, which has a good sketch drawing effect, can improve design efficiency, and satisfies ordinary designers usage of.

The invention also provides a sketch generation system based on image feature points, which reduces time complexity and improves drawing efficiency.

A sketch generation system based on image feature points according to the present invention includes a sketch generation server based on image feature points as shown in FIG. 8 .

The invention extracts feature points according to the solid image, controls the Harris corner detection operator by using the Euclidean distance difference, and combines the segmental spline interpolation algorithm to outline the outline edge, and then completes the design of the sketch, reduces the time complexity, and improves the drawing efficiency.

Those skilled in the art should understand that the embodiments of the present invention may be provided as methods, systems, or computer program products. Accordingly, the present invention can take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It should be understood that each procedure and/or block in the flowchart and/or block diagram, and a combination of procedures and/or blocks in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions may be provided to a general purpose computer, special purpose computer, embedded processor, or processor of other programmable data processing equipment to produce a machine such that the instructions executed by the processor of the computer or other programmable data processing equipment produce a An apparatus for realizing the functions specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing apparatus to operate in a specific manner, such that the instructions stored in the computer-readable memory produce an article of manufacture comprising instruction means, the instructions The device realizes the function specified in one or more procedures of the flowchart and/or one or more blocks of the block diagram.

These computer program instructions can also be loaded onto a computer or other programmable data processing device, causing a series of operational steps to be performed on the computer or other programmable device to produce a computer-implemented process, thereby The instructions provide steps for implementing the functions specified in the flow chart or blocks of the flowchart and/or the block or blocks of the block diagrams.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented through computer programs to instruct related hardware, and the programs can be stored in a computer-readable storage medium. During execution, it may include the processes of the embodiments of the above-mentioned methods. Wherein, the storage medium may be a magnetic disk, an optical disk, a read-only memory (Read-Only Memory, ROM) or a random access memory (Random Access Memory, RAM) and the like.

Although the specific implementation of the present invention has been described above in conjunction with the accompanying drawings, it does not limit the protection scope of the present invention. Those skilled in the art should understand that on the basis of the technical solution of the present invention, those skilled in the art do not need to pay creative work Various modifications or variations that can be made are still within the protection scope of the present invention.

Claims (8)

1. A sketch generation method based on image feature points, characterized in that, comprising: Perform preprocessing operations on the acquired entity images; Based on the wavelet adaptive Harris detection operator to extract the feature points of the edge of the preprocessed solid image; Using the segmented spline interpolation method to fit the outline of the feature points on the edge of the extracted solid image, obtain a smooth and closed curve outline, and generate a corresponding sketch; The specific process of extracting the feature points of the edge of the preprocessed solid image based on the wavelet adaptive Harris detection operator includes: Convert the preprocessed entity image into a binary image; Calculate the horizontal and vertical components of the binary image on the preset scale, and construct the modulus and argument of the wavelet coefficient; Calculate the gradient of the binary image in the horizontal and vertical directions according to the modulus and argument of the constructed wavelet coefficients; According to the gradient of the binary image in the horizontal and vertical directions, use the Gaussian function to Gaussian weight the feature points, and add the Euclidean distance at the same time, calculate the difference between the points, and generate a matrix element value sequence; In the matrix element value sequence, the corner point is further determined by calculating the Harris response value of each image pixel, and the feature point of the edge of the image is extracted by a wavelet-based adaptive Harris detection operator pair. 2. A sketch generation method based on image feature points as claimed in claim 1, wherein the preprocessing operation on the acquired entity image includes noise reduction processing and boundary tracking of the entity image. 3. A method for generating sketches based on image feature points as claimed in claim 2, characterized in that, the acquired entity image is denoised by sequentially adopting a median filter denoising method and an expansion operation method. 4. A sketch generation server based on image feature points, characterized in that, comprising: A preprocessing module, which is configured to: perform a preprocessing operation on the acquired entity image; The edge feature point extraction module is configured to: extract the feature points of the edge of the preprocessed entity image based on the wavelet adaptive Harris detection operator; The edge feature point fitting module is configured to: use the segmented spline interpolation method to perform contour fitting on the feature points of the extracted solid image edge, obtain a smooth and closed curve contour, and generate a corresponding sketch; The specific process of extracting the feature points of the edge of the preprocessed solid image based on the wavelet adaptive Harris detection operator includes: Convert the preprocessed entity image into a binary image; Calculate the horizontal and vertical components of the binary image on the preset scale, and construct the modulus and argument of the wavelet coefficient; Calculate the gradient of the binary image in the horizontal and vertical directions according to the modulus and argument of the constructed wavelet coefficients; According to the gradient of the binary image in the horizontal and vertical directions, use the Gaussian function to Gaussian weight the feature points, and add the Euclidean distance at the same time, calculate the difference between the points, and generate a matrix element value sequence; In the matrix element value sequence, the corner point is further determined by calculating the Harris response value of each image pixel, and the feature point of the edge of the image is extracted by a wavelet-based adaptive Harris detection operator pair. 5. a kind of sketch generation server based on image feature point as claimed in claim 4, is characterized in that, in described preprocessing module, carrying out preprocessing operation to the entity image that obtains comprises denoising processing and the boundary of entity image track. 6. a kind of sketch generation server based on image feature point as claimed in claim 5, it is characterized in that, in described preprocessing module, adopt median filter denoising method and dilation operation method successively to reduce the entity image that obtains noise. 7. a kind of sketch generation server based on image feature point as claimed in claim 4, is characterized in that, described edge feature point extraction module comprises: The binary image conversion submodule is configured to: convert the preprocessed entity image into a binary image; The modulus and argument construction submodule of the wavelet coefficient is configured to: calculate the horizontal and vertical components of the binary image on a preset scale, and construct the modulus and argument of the wavelet coefficient; The horizontal and vertical direction gradient calculation sub-module is configured to: calculate the binary image gradient in the horizontal and vertical directions according to the modulus and argument of the constructed wavelet coefficients; The matrix element value sequence generation sub-module is configured to: use the Gaussian function to Gaussian weight the feature points according to the gradient of the binary image in the horizontal and vertical directions, and add the Euclidean distance at the same time, calculate the difference between the points, and generate a sequence of matrix element values; The feature point extraction submodule is configured to: in the matrix element value sequence, further determine the corner point by calculating the Harris response value of each image pixel, and extract the image edge by using a Harris detection operator based on wavelet adaptation feature points. 8. A sketch generation system based on image feature points, comprising the sketch generation server based on image feature points according to any one of claims 4-7.

Images