Robot face sketching system

Abstract

The invention discloses a robot face sketch system, relates to the technical field of robots, solves the problems that the confirmation precision of an original contour line is not high and a contour region is not optimized, and determines sketch parameters according to different feature mean values and numerical intervals to which the feature mean values belong. And the uniqueness of each facial feature in the original image can be fully reflected. Due to the fact that the numerical value intervals and the sketch parameters are drawn up by operators according to experience, the numerical value intervals and the sketch parameters can well reflect due expression modes of different facial features during sketch, personalized sketch processing of the facial features of different individuals is achieved, sketch results are more vivid, and user experience is improved. According to the method based on the pixel value and the gradient value, the contour in the face surface high-definition image can be accurately recognized, and errors possibly caused by manual judgment are avoided. Through detailed mathematical operation, such as transverse gradient and vertical gradient calculation, the change of the facial contour can be accurately captured, so that an accurate contour line is obtained.

Description

Technical Field

[0001] The invention relates to the technical field of robots, in particular to a robot face sketching system. Background Art

[0002] The sketching robot obtains the image information to be drawn through a camera or other image input device. Then, it uses a deep learning algorithm to recognize and analyze the image, identify key elements such as the outline of the object, lines, light and shadow in the image, and convert it into a data form that the computer can understand45; based on the results of the image analysis, the robot's control system will plan the path and order of drawing; through complex algorithms, the movement trajectory and movements of the robotic arm or drawing tool during the drawing process are calculated to ensure that each part of the image can be accurately depicted5; finally, the robot's robotic arm accurately controls the drawing tool (such as pencil, charcoal, etc.) to draw the sketch according to the generated path planning. During the drawing process, the robotic arm will adjust the angle, pressure and other parameters of the drawing tool as needed to achieve different line effects and shadow performances;

[0003] The application with publication number CN108230238A discloses a robot face sketching system, comprising: a webcam, an image processing unit connected to the webcam, a coordinate conversion unit connected to the image processing unit, a computer connected to the coordinate conversion unit, a SCARA robot controlled by the computer, and a brush fixing mechanism controlled by the computer, which is located on the brush fixing mechanism of the SCARA robot and is used to fix the brush. The present invention can process a portrait into a simple outline pattern before drawing, and uses a low-cost webcam and SCARA robot to draw facial features more flexibly and efficiently, which is more practical.

[0004] When the robot is sketching, it usually determines the corresponding face contour based on the corresponding face image, and then performs sketching based on the face contour. Originally, the sketch is only based on the more obvious contour, and the contour line is not specifically confirmed based on the pixel features between the pixels. As a result, the accuracy of the confirmed contour line is not high. After the contour sketch is completed, the corresponding contour area is not optimized, and the sketch image cannot be made more representative of the human figure. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a robot face sketching system, which solves the problems of low accuracy in original contour line confirmation and non-optimization of contour areas.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: a robot face sketching system, comprising:

[0007] The face image acquisition end uses a high-definition probe to acquire a high-definition surface image of the face, and transmits the acquired high-definition surface image to the image contour analysis end;

[0008] The image contour analysis end confirms the different pixel values ​​associated with different pixel points in the acquired surface high-definition image, and then confirms the contour line from the surface high-definition image based on the comprehensive performance of the pixel values ​​between different pixel points. The specific method is as follows:

[0009] Based on the different groups of pixels in the surface HD image, the different pixel values ​​associated with different pixels are calibrated as X i , where i represents different pixels;

[0010] Confirm the gradient value TD associated with the corresponding pixel point i :Mark the current pixel as the main pixel, and the pixel value of the main pixel is X i , the pixel values ​​associated with other pixels adjacent to the main pixel are calibrated as H1-H8, and the pixel values ​​of the main pixel and the surrounding adjacent pixels are arranged as follows:

[0011] Confirm the horizontal gradient TX associated with this main pixel point in the horizontal direction i , the operator template used is Among them TX i =H1×(-1)+H2×0+H3×1+H4×(-2)+X i ×0+H5×2+H6×(-1)+H7×0+H8×1;

[0012] Confirm the vertical gradient SX associated with this main pixel in the vertical direction i , the computing power template used is Among them SX i =H1×(-1)+H2×(-2)+H3×(-1)+H4×0+X i ×0+H5×0+H6×1+H7×2+H8×1;

[0013] use: Confirm the gradient value TD of the main pixel i ;

[0014] The gradient values ​​associated with different pixels in the surface high-definition image are confirmed in turn, and the gradient values ​​TD of the confirmed different pixels are converted into i Compare with the preset value Y1. If TD i ≥Y1, the corresponding pixel point is marked as a contour point. If TD i <Y1, no processing is performed;

[0015] Connect and confirm the contour points that appear continuously in the surface high-definition image to determine the contour line belonging to the surface high-definition image;

[0016] The point pressure confirmation terminal confirms the point drawing pressure of different contour points based on the different contour lines confirmed in the surface high-definition image and the several groups of different contour points associated with the corresponding contour lines, and simultaneously generates an execution data package. The specific method is as follows:

[0017] Based on the different contour lines confirmed in the corresponding surface HD image, the gradient values ​​TD associated with the different contour points in the corresponding contour lines are locked i ;

[0018] Adopt YL i =TD i ×C1 confirms the drawing pressure YL of the corresponding contour point i , where C1 is a preset fixed coefficient factor, and the determined contour lines and the drawing pressure YL associated with different contour points in different contour lines are i Integrate into execution data packets, and transmit the execution data packets to the execution end;

[0019] The contour feature confirmation end extracts preset feature standards from the database based on the different contour lines confirmed in the surface high-definition image, checks the feature standards with the pixel point features associated with the contour lines, confirms the selected features associated with the corresponding contour lines, and transmits the confirmed selected features to the execution end, specifically in the following manner:

[0020] Based on different contour lines confirmed in the surface high-definition image, the total length of the line segments of the corresponding contour lines is confirmed, and based on the total length of the line segments, the middle point inside the contour line is confirmed, and the middle point is used as the feature point of the contour line;

[0021] With this characteristic point as the center and R1 as the radius, a group of radiating circles is confirmed, wherein R1 is a preset value, and different pixel points associated with different diameters in the radiating circle are determined, and starting from one end point of the diameter to the other end point, the pixel points associated with the corresponding diameter are sorted to confirm the pending point sequence, each group of diameters corresponds to a group of pending point sequence, and the pixel values ​​associated with different pixel points in the pending point sequence are processed with variance to confirm the characteristic variance, and the characteristic variances of the pending point sequences associated with different diameters are confirmed in turn, and the maximum characteristic variance is selected from multiple groups of characteristic variances, and the pending point sequence corresponding to the maximum characteristic variance is used as the characteristic sequence;

[0022] Confirm the correlation difference of the corresponding pixel values ​​of adjacent pixels in the feature sequence: the pixel value associated with a group of pixels in the proposed feature sequence is X1, and the pixel value associated with the adjacent pixels of this pixel is calibrated as X2, and the correlation difference = |X1-X2| is used to confirm the different correlation differences related to the adjacent pixels in the feature sequence in turn;

[0023] The confirmed groups of correlation differences are processed by averaging, and the processed mean is used as the characteristic mean;

[0024] Confirm the numerical ranges associated with different feature standards from the database, all of which are preset ranges, confirm the numerical range to which the feature mean value belongs, mark the feature standard associated with the numerical range as the selected feature, and transmit the sketch parameters associated with the selected feature to the execution end;

[0025] The execution end preferentially controls the robot based on the determined execution data packet, so that the robot performs a preliminary sketch according to the determined contour line and the drawing pressure associated with different contour points.

[0026] Preferably, it also includes:

[0027] The database stores different feature standards and different numerical ranges, wherein different feature standards are associated with different numerical ranges, and different feature standards correspond to sketch parameters of different facial features, wherein the numerical ranges and the associated sketch parameters are all preset parameters.

[0028] Preferably, the execution end performs sketch processing on the peripheral area of ​​the edge contour line according to the determined sketch parameters based on different selected features associated with different edge contour lines and based on different sketch parameters associated with the different selected features.

[0029] The present invention provides a robot face sketching system. Compared with the prior art, it has the following beneficial effects:

[0030] The present invention determines the contour line by accurately calculating the pixel points and their gradient values ​​through the image contour analysis end. This method based on pixel values ​​and gradient values ​​can accurately identify the contour in the high-definition image of the face surface, avoiding the errors that may occur in manual judgment. Through detailed mathematical operations, such as calculating the horizontal gradient and the vertical gradient, the changes in the facial contour can be accurately captured, thereby obtaining an accurate contour line;

[0031] Determining sketch parameters based on different feature means and their numerical ranges can fully reflect the uniqueness of each facial feature in the original image. Since these numerical ranges and sketch parameters are formulated by operators based on experience, they can well reflect the way different facial features should be expressed when sketching, thereby realizing personalized sketch processing of different individual facial features, making the sketch results more realistic. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a schematic diagram of the principle framework of the present invention. DETAILED DESCRIPTION

[0033] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0034] First embodiment

[0035] See also Figure 1 , the present application provides a robot face sketching system, including a face image acquisition end, an image contour analysis end, a point pressure confirmation end, a contour feature confirmation end, a database, and an execution end;

[0036] The face image acquisition end is electrically connected to the image contour analysis end input node, the image contour analysis end is electrically connected to the point pressure confirmation end input node, and the point pressure confirmation end is electrically connected to the contour feature confirmation end and the execution end input node respectively, wherein the database, the contour feature confirmation end and the execution end are electrically connected from the output node to the input node respectively;

[0037] Among them, the face image acquisition end uses a high-definition probe to acquire a high-definition surface image of the face, and transmits the acquired high-definition surface image to the image contour analysis end. Specifically, when acquiring the image, the corresponding person needs to be placed at a designated position, and then the relevant personnel acquire the high-definition image of such person;

[0038] The image contour analysis end confirms different pixel values ​​associated with different pixel points in the acquired surface high-definition image, and then confirms the contour line from the surface high-definition image based on the comprehensive performance of the pixel values ​​between different pixel points. The specific method of confirming the contour line is:

[0039] Based on the different groups of pixels in the surface HD image, the different pixel values ​​associated with different pixels are calibrated as X i , where i represents different pixels;

[0040] Confirm the gradient value TD associated with the corresponding pixel point i :Mark the current pixel as the main pixel, and the pixel value of the main pixel is X i , the pixel values ​​associated with other adjacent pixels around the main pixel are calibrated as H1-H8. In the actual image, the pixels are arranged in an array form, so the pixel values ​​associated with the surrounding of the current main pixel are eight groups, that is, the main pixel is located in the middle. The pixel values ​​of the main pixel and the surrounding adjacent pixels are arranged as follows:

[0041] Confirm the horizontal gradient TX associated with this main pixel point in the horizontal direction i , the operator template used is Among them TX i =H1×(-1)+H2×0+H3×1+H4×(-2)+X i ×0+H5×2+H6×(-1)+H7×0+H8×1;

[0042] Confirm the vertical gradient SX associated with this main pixel in the vertical direction i , the computing power template used is Among them SX i =H1×(-1)+H2×(-2)+H3×(-1)+H4×0+X i ×0+H5×0+H6×1+H7×2+H8×1;

[0043] use: Confirm the gradient value TD of the main pixel i ;

[0044] The gradient values ​​associated with different pixels in the surface high-definition image are confirmed in turn. For the related pixels that do not meet the determined characteristics (that is, the internal pixels with 8 pixels around them), there is no need to confirm the gradient values. Generally, such pixels are edge points of the image, and the related features of such points do not need to be confirmed. The gradient values ​​TD of the confirmed different pixels are i Check with the preset value Y1, where the specific value of Y1 is determined by the operator based on experience. i <Y1, no processing is performed. If TD i ≥Y1, the corresponding pixel point is marked as a contour point;

[0045] Connect and confirm the contour points that appear continuously in the surface high-definition image to determine the contour line belonging to the surface high-definition image;

[0046] Specifically, there are several groups of different pixel points in the corresponding surface high-definition image, and different pixel points have different pixel values. Based on the relevant gradient of each pixel point, the contour points are specifically confirmed, and the internal contour line of the corresponding image can be determined based on the contour points that appear continuously and sequentially in the corresponding image. Based on the determined contour line, the contour is specifically calibrated, which facilitates the subsequent robot to sketch according to the determined contour line.

[0047] Among them, the point pressure confirmation end confirms the point drawing pressure of different contour points based on different contour lines confirmed in the surface high-definition image and a number of groups of different contour points associated with the corresponding contour lines. The specific method of confirmation is:

[0048] Based on the different contour lines confirmed in the corresponding surface HD image, the gradient values ​​TD associated with the different contour points in the corresponding contour lines are locked i (The higher the gradient value, the darker the associated contour, and the lower the gradient value, the lighter the corresponding contour).

[0049] Adopt YL i =TD i ×C1 confirms the drawing pressure YL of the corresponding contour point i , where C1 is a preset fixed coefficient factor, the specific value of which is determined by the operator based on experience, and the determined contour lines and the drawing pressure YL associated with different contour points in different contour lines are i Integrate into execution data packets, and transmit the execution data packets to the execution end;

[0050] The execution end controls the robot based on the determined execution data packet, so that the robot performs a preliminary sketch according to the determined contour line and the drawing pressure associated with different contour points;

[0051] Specifically, the corresponding robot controls the mechanical arm to perform sketching according to the confirmed contour line and different drawing pressures associated with different contour points.

[0052] Second embodiment

[0053] In the specific implementation process, compared with the above-mentioned embodiment, this embodiment mainly targets specific sketch parameters of different areas in the image, and comprehensively confirms the trend characteristics of different points in the original image, which is executed by the corresponding contour feature confirmation end;

[0054] Among them, the contour feature confirmation end extracts the preset feature standards from the database based on the different contour lines confirmed in the surface high-definition image, checks the feature standards with the pixel point features associated with the contour lines, confirms the facial features associated with the corresponding contour lines, and transmits the confirmed facial features to the execution end. Its facial features include eye features, mouth and nose features, etc. Each different type of feature has different sketch parameters. The specific method of facial feature confirmation is as follows:

[0055] Based on different contour lines confirmed in the surface high-definition image, the total length of the line segments of the corresponding contour lines is confirmed, and based on the total length of the line segments, the middle point inside the contour line is confirmed, and the middle point is used as the feature point of the contour line;

[0056] With this feature point as the center and R1 as the radius, a group of radiating circles is confirmed (the radiating circles will generally include this contour line), where R1 is a preset value, which is prepared in advance by relevant operators based on experience. Different pixel points associated with different diameters in the radiating circle are determined, and starting from one end point of the diameter to the other end point, the pixel points associated with the corresponding diameter are sorted to confirm the pending point sequence. Each group of diameters corresponds to a group of pending point sequences. The pixel values ​​associated with different pixel points in the pending point sequence are processed with variance to confirm the feature variance. The feature variances of the pending point sequences associated with different diameters are confirmed in turn. The maximum feature variance is selected from multiple groups of feature variances, and the pending point sequence corresponding to the maximum feature variance is used as the feature sequence.

[0057] Confirm the correlation difference of the corresponding pixel values ​​of adjacent pixels in the feature sequence: the pixel value associated with a group of pixels in the proposed feature sequence is X1, and the pixel value associated with the adjacent pixels of this pixel is calibrated as X2, and the correlation difference = |X1-X2| is used to confirm the different correlation differences related to the adjacent pixels in the feature sequence in turn;

[0058] The confirmed groups of correlation differences are processed by averaging, and the processed mean is used as the characteristic mean;

[0059] Confirm the numerical ranges associated with different feature standards from the database. The numerical ranges are all preset ranges, which are prepared in advance by relevant operators based on experience. Confirm the numerical range to which the feature mean belongs, calibrate the feature standard associated with this numerical range as the selected feature, and transmit the sketch parameters associated with the selected feature to the execution end.

[0060] Among them, the database stores different feature standards and different numerical ranges, and different feature standards are associated with different numerical ranges, and different feature standards correspond to sketch parameters of different facial features, and the numerical ranges and the associated sketch parameters are all preset parameters, which are all prepared in advance by relevant operators based on experience.

[0061] Among them, the execution end, based on different selected features associated with different edge contour lines and based on different sketch parameters associated with different selected features, performs sketch processing on the peripheral area of ​​this edge contour line according to the determined sketch parameters, and the peripheral area is the determined radiation circle.

[0062] Through the detailed analysis of different contour lines in the surface high-definition image, facial features such as eye features, mouth and nose features can be accurately confirmed. This pixel point analysis method based on the middle point inside the contour line and the radial circle fully considers the geometric characteristics of the image and the change of pixel values, avoiding the problem of misjudgment of facial features that may occur in traditional methods, thus making the basis of sketching - facial feature recognition more accurate;

[0063] Determining sketch parameters based on different feature means and their numerical ranges can fully reflect the uniqueness of each facial feature in the original image. Since these numerical ranges and sketch parameters are formulated by operators based on experience, they can well reflect the way different facial features should be expressed when sketching, thereby realizing personalized sketch processing of different individual facial features, making the sketch results more realistic.

[0064] Some of the data in the above formulas are dimensionless and numerically calculated. Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0065] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. Robot face sketching system, characterized by: include: The face image acquisition end uses a high-definition probe to acquire a high-definition surface image of the face, and transmits the acquired high-definition surface image to the image contour analysis end; The image contour analysis end confirms different pixel values ​​associated with different pixel points in the acquired surface high-definition image, and then confirms the contour line from the surface high-definition image based on the comprehensive performance of the pixel values ​​between different pixel points; The point pressure confirmation terminal confirms the point drawing pressure of different contour points based on the different contour lines confirmed in the surface high-definition image and the several groups of different contour points associated in the corresponding contour lines, and simultaneously generates an execution data packet; The contour feature confirmation end extracts preset feature standards from a database based on different contour lines confirmed in the surface high-definition image, checks the feature standards with pixel point features associated with the contour lines, confirms the selected features associated with the corresponding contour lines, and transmits the confirmed selected features to the execution end; The execution end preferentially controls the robot based on the determined execution data packet, so that the robot performs a preliminary sketch according to the determined contour line and the drawing pressure associated with different contour points.

2. The robot face sketching system according to claim 1, characterized in that: The specific method of the image contour analysis end to confirm the contour line from the surface high-definition image is: Based on the different groups of pixels in the surface HD image, the different pixel values ​​associated with different pixels are calibrated as X i , where i represents different pixels; Confirm the gradient value TD associated with the corresponding pixel point i :Mark the current pixel as the main pixel, and the pixel value of the main pixel is X i , the pixel values ​​associated with other pixels adjacent to the main pixel are calibrated as H1-H8, and the pixel values ​​of the main pixel and the surrounding adjacent pixels are arranged as follows: Confirm the horizontal gradient TX associated with this main pixel point in the horizontal direction i , the operator template used is Among them TX i =H1×(-1)+H2×0+H3×1+H4×(-2)+X i ×0+H5×2+H6×(-1)+H7×0+H8×1; Confirm the vertical gradient SX associated with this main pixel in the vertical direction i , the computing power template used is Among them SX i =H1×(-1)+H2×(-2)+H3×(-1)+H4×0+X i ×0+H5×0+H6×1+H7×2+H8×1; use: Confirm the gradient value TD of the main pixel i ; The gradient values ​​associated with different pixels in the surface high-definition image are confirmed in turn, and the gradient values ​​TD of the confirmed different pixels are converted into i Compare with the preset value Y1. If TD i ≥Y1, the corresponding pixel point is marked as a contour point; The contour points that appear continuously in the surface high-definition image are connected and confirmed to determine the contour line belonging to this surface high-definition image.

3. The robot face sketching system according to claim 2, characterized in that: If TD i <Y1, no processing is performed.

4. The robot face sketching system according to claim 2, characterized in that: The point pressure confirmation terminal confirms the point drawing pressure of different contour points in the following specific manner: Based on the different contour lines confirmed in the corresponding surface HD image, the gradient values ​​TD associated with the different contour points in the corresponding contour lines are locked i ; Adopt YL i =TD i ×C1 confirms the drawing pressure YL of the corresponding contour point i , where C1 is a preset fixed coefficient factor, and the determined contour lines and the drawing pressure YL associated with different contour points in different contour lines are i The data packets are integrated into execution data packets, and the execution data packets are transmitted to the execution end.

5. The robot face sketching system according to claim 1, characterized in that: The specific method of the contour feature confirmation end for confirming the selected feature is as follows: Based on different contour lines confirmed in the surface high-definition image, the total length of the line segments of the corresponding contour lines is confirmed, and based on the total length of the line segments, the middle point inside the contour line is confirmed, and the middle point is used as the feature point of the contour line; With this characteristic point as the center and R1 as the radius, a group of radiating circles is confirmed, wherein R1 is a preset value, and different pixel points associated with different diameters in the radiating circle are determined, and starting from one end point of the diameter to the other end point, the pixel points associated with the corresponding diameter are sorted to confirm the pending point sequence, each group of diameters corresponds to a group of pending point sequence, and the pixel values ​​associated with different pixel points in the pending point sequence are processed with variance to confirm the characteristic variance, and the characteristic variances of the pending point sequences associated with different diameters are confirmed in turn, and the maximum characteristic variance is selected from multiple groups of characteristic variances, and the pending point sequence corresponding to the maximum characteristic variance is used as the characteristic sequence; Confirm the correlation difference of the corresponding pixel values ​​of adjacent pixels in the feature sequence: the pixel value associated with a group of pixels in the proposed feature sequence is X1, and the pixel value associated with the adjacent pixels of this pixel is calibrated as X2, and the correlation difference = |X1-X2| is used to confirm the different correlation differences related to the adjacent pixels in the feature sequence in turn; The confirmed groups of correlation differences are processed by averaging, and the processed mean is used as the characteristic mean; Confirm the numerical ranges associated with different feature standards from the database, all of which are preset ranges, confirm the numerical range to which the feature mean belongs, calibrate the feature standard associated with this numerical range as the selected feature, and transmit the sketch parameters associated with the selected feature to the execution end.

6. The robot face sketching system according to claim 5, characterized in that: Also includes: The database stores different feature standards and different numerical ranges, wherein different feature standards are associated with different numerical ranges, and different feature standards correspond to sketch parameters of different facial features, wherein the numerical ranges and the associated sketch parameters are all preset parameters.

7. The robot face sketching system according to claim 6, characterized in that: The execution end performs sketch processing on the peripheral area of ​​the edge contour line according to the determined sketch parameters based on the different selected features associated with the different edge contour lines and based on the different sketch parameters associated with the different selected features.

Images