A false eyelash glue line detection method based on corner point sorting

Abstract

The application belongs to the intelligent identification technology field and relates to a false eyelash glue coating line detection method based on corner point sequencing, which comprises the following steps: step one, performing gray scale processing on a three-channel eyelash image captured by a camera, and then performing inversion binarization and binarization operations; step two, using open operation to extract the edge line of the false eyelash image for the binarized image, and then using Hough transformation to perform straight line detection and select the upper edge line of the eyelash; step three, using a Canny edge detection algorithm to detect the contour in the inverted binarized image, screening the candidate contour according to the contour area and shape, using the rightmost point of the candidate contour as a corner point for sequencing, and eliminating the interference data after sequencing to determine the glue coating point; and step four, calculating the position of the glue coating point in the original image according to the rotation matrix obtained through edge line detection, and then fitting the glue coating line according to the glue coating point. The application can enable the machine to automatically identify the eyelash glue coating line and simultaneously obtain the accurate position of the glue coating line.

Description

Technical Field

[0001] This invention relates to the fields of image processing and intelligent recognition and detection technology, specifically to a method for detecting adhesive lines on false eyelashes based on corner point sorting. Background Technology

[0002] The beauty industry is developing rapidly, and the false eyelash market has huge potential. However, due to production capacity limitations, there is a significant shortage of false eyelashes. A major reason for the low production volume is the slow speed of applying glue to the eyelashes. Currently, false eyelash glue application is mainly done manually. To improve efficiency and reduce labor costs, some manufacturers have started using automated equipment for glue application, but many problems still limit the speed and quality of glue application. These problems mainly lie in two aspects: 1. Current glue application machines require manual placement of the eyelashes on the designated positions on the glue application plate; 2. The glue application line position needs to be manually adjusted and set, and cannot provide real-time glue application lines for each individual eyelash. Summary of the Invention

[0003] To address the aforementioned technical problems in existing technologies, this invention proposes a method for detecting adhesive lines on false eyelashes based on corner point sorting. By performing contour filtering and corner point sorting on each eyelash image, the machine can automatically identify the adhesive lines on the eyelashes and simultaneously obtain the precise location of the adhesive lines. The specific technical solution is as follows: A method for detecting adhesive lines on false eyelashes based on corner point sorting includes the following steps: Step 1: Convert the three-channel eyelash image captured by the camera to grayscale, and then perform inverse binarization and binarization operations to obtain the corresponding inverse binarized image and binarized image. Step 2: For the binarized image, use the opening operation to extract the edge lines of the false eyelash image, and then use the Hough transform to perform line detection and select the upper edge line of the eyelash. Step 3: Use the Canny edge detection algorithm to detect contours in the inverted binarized image. Select candidate contours based on their area and shape. Sort the corner points by using the rightmost point of the candidate contour as the corner point and remove the sorted interference data to determine the glue application point. Step 4: Calculate the position of the glue application point in the original image based on the rotation matrix obtained from the edge line detection, and then fit the glue application line based on the glue application point.

[0004] Furthermore, the expressions for the inverse binarization and binarization operations are as follows: , , in, G(x,y) It is in the result image ( x , yThe grayscale value at ) g(x,y) In grayscale images ( x , y The grayscale value at ) T It is a specified threshold that is adjusted according to lighting conditions. x, y These represent the horizontal and vertical coordinates of the coordinate axes in the image, respectively.

[0005] Furthermore, step two specifically involves: first eroding the binarized image, then dilating it; subtracting the dilated result from the eroded result to obtain the edge lines; then performing Hough transform on the edge lines for line detection; and finally, determining the line type based on its left endpoint. y Sort the coordinate values ​​and select y The line with the smallest coordinate value is taken as the upper edge line; the upper edge line of the eyelash is rotated to a horizontal position to calculate the rotation matrix. R Then based on the rotation matrix R Rotate the entire image.

[0006] Furthermore, step three specifically includes the following sub-steps: Step 3.1: Use the Canny edge detection algorithm to find contours in the rotated grayscale image. Select candidate contours by setting minimum and maximum thresholds for contour area, height difference, and width difference. Step 3.2, for each candidate contour's point set, according to... x Sort the coordinate values, take the rightmost point of each contour as the corner point, and then... y The coordinate values ​​are sorted by corner points, and the noisy data at the beginning of the sorted result is removed, followed by the continuous data. N These data points are considered as adhesive application points. N This is an integer value that is dynamically adjusted based on eyelash type.

[0007] Furthermore, the method for removing interfering data from the beginning of the sorted data specifically involves: removing data from adjacent corner points. y If the difference in values ​​is greater than the set threshold and the number of consecutive corner points is less than 20, then the preceding corner point data is considered to be interference data and is removed.

[0008] Furthermore, step four specifically involves rotating the adhesive application point, i.e., based on the calculated rotation matrix. R Calculate the position of the glue application point in the original image; then use the least squares method to fit the position of the glue application line. Beneficial effects

[0009] 1. This invention uses image processing algorithms to obtain the position of the glue application line for false eyelashes in real time, reducing the positional requirements for attaching false eyelashes to the glue application plate and increasing the speed of manual eyelash application.

[0010] 2: The contour screening scheme proposed in this invention using area and shape is universal and can be applied to contour detection tasks with similar contour position distribution patterns, areas, and shapes.

[0011] 3: This invention proposes a corner selection scheme based on corner sorting, which can accurately detect corners at specific locations on objects such as false eyelashes with regular corner position arrangements.

[0012] 4. The glue coating line detection scheme of the present invention can be applied to various types of false eyelashes, providing a universal solution for the detection of glue coating lines of different types of false eyelashes. Attached Figure Description

[0013] Figure 1 This is the original image of the false eyelashes according to an embodiment of the present invention; Figure 2 This is an image of false eyelashes after inversion binarization according to an embodiment of the present invention; Figure 3 To reverse Figure 2 The resulting image after the image; Figure 4 This is a schematic diagram of the edge line detection results according to an embodiment of the present invention; Figure 5 This is a schematic diagram of all contour detection results in an image according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the detection results after corner point sorting in an embodiment of the present invention; Figure 7 This is a schematic diagram of the final glue coating line detection results in an embodiment of the present invention; Figure 8 This is a flowchart illustrating a method for detecting the glue lines on false eyelashes based on corner point sorting according to the present invention. Detailed Implementation

[0014] To make the objectives, technical solutions, and technical effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments.

[0015] like Figures 1-7 The image shown is an example of the original sampling image, processing procedure, and final result image of an embodiment of detecting adhesive lines using the method of the present invention.

[0016] like Figure 8 As shown in the figure, an embodiment of the present invention provides a method for detecting the glue line of false eyelashes based on corner point sorting, which specifically includes the following steps: Step 1: Convert the three-channel eyelash image captured by the camera to grayscale, and then perform inverse binarization and binarization operations to obtain the corresponding inverse binarized image and binarized image.

[0017] Specifically, this invention detects a single eyelash image each time, such as Figure 1 The image shown is of a single eyelash. Because the eyelashes are spaced apart at the glue application plate, each eyelash can be imaged by cropping the image.

[0018] In addition, a light source can be placed under the coating plate to provide good lighting conditions.

[0019] like Figure 2 and Figure 3 The image shown is the result of inverse binarization and binarization of a grayscale image of eyelashes. Inverse binarization is used to obtain the skeleton map of the contour, facilitating subsequent edge detection; binarization is used for subsequent opening operations to extract the edge lines of the eyelashes. The expressions for inverse binarization and binarization are shown below: , , in, G(x,y) It is in the result image ( x , y The grayscale value at ) g(x,y) In grayscale images ( x , y The grayscale value at ) T This is a specified threshold that can be adjusted according to lighting conditions. x, y These represent the x and y coordinates of the coordinate axes in the image, respectively. 。

[0020] Step 2: For the binarized image, use the opening operation to extract the edge lines of the false eyelash image, and then use the Hough transform to perform line detection and select the upper edge line of the eyelash.

[0021] Specifically, by first eroding and then dilating the binarized image, the edge lines can be obtained by subtracting the latter from the former result. Figure 4 As shown, Hough transform is then applied to the edge lines for line detection. The detected lines are then processed according to the left endpoint of the line. y Sort the coordinate values ​​and select y The line with the smallest coordinate value is taken as the upper edge line; the upper edge line of the eyelash is rotated to a horizontal position to obtain the rotation matrix. R Then based on the rotation matrix R Rotate the entire image.

[0022] Step 3: Use the Canny edge detection algorithm to detect contours in the inverted binarized image. Select candidate contours based on their area and shape. Sort the corner points by using the rightmost point of the candidate contour as the corner point and remove the sorted interference data to determine the glue application point. Figures 5-6 These are the contour detection results after rotating the eyelash image to a horizontal position and the contour detection results after sorting and filtering by corner points.

[0023] Step 3.1, in this embodiment of the invention, the Canny edge detection algorithm is used to find the contours in the rotated grayscale image. Figure 5 As there are many contours, it is necessary to filter them. This invention filters contours by setting minimum and maximum thresholds for contour area, height difference, and width difference, and finds the required contours as feature contours, i.e., candidate contours.

[0024] Step 3.2, for each candidate contour's point set, according to... x Sort the coordinate values, take the rightmost point of each contour as the corner point, and then... y The coordinate values ​​are sorted by corner points, and then the interfering data at the beginning of the sorted list is removed. Specifically, the method for removing interfering data at the beginning involves considering two adjacent corner points... y If the difference in values ​​is greater than a threshold of 5, and the number of consecutively collected points is less than 20, then the previously collected data is considered to be interference data; then the subsequent consecutive data... N ( N (This can be dynamically adjusted according to eyelash type) These data points are considered as application points, resulting in the filtered contour effect as follows: Figure 6 As shown in the image, the comparison of the results demonstrates that the solution of this invention is effective and accurate.

[0025] Step 4: Calculate the position of the glue application point in the original image based on the rotation matrix obtained from the edge line detection, and then fit the glue application line based on the glue application point.

[0026] Specifically, the rotating adhesive application point is determined based on the previously calculated rotation matrix. R The positions of the adhesive application points in the original image were calculated; then, the positions of the adhesive application lines were fitted using the least squares method. The detection results are as follows: Figure 7 As shown.

[0027] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention in any way. Although the implementation process of the present invention has been described in detail above, those skilled in the art can still modify the technical solutions described in the foregoing examples or make equivalent substitutions for some of the technical features. All modifications and equivalent substitutions made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for detecting adhesive lines on false eyelashes based on corner point sorting, characterized in that, Includes the following steps: Step 1: Convert the three-channel eyelash image captured by the camera to grayscale, and then perform inverse binarization and binarization operations to obtain the corresponding inverse binarized image and binarized image. Step 2: For the binarized image, use the opening operation to extract the edge lines of the false eyelash image, and then use the Hough transform to perform line detection and select the upper edge line of the eyelash. Specifically, the process involves first eroding the binarized image, then dilating it. The edge lines are obtained by subtracting the dilated result from the eroded result. A Hough transform is then applied to the edge lines for line detection. The detected lines are then analyzed based on their left endpoints. y Sort the coordinate values ​​and select y The line with the smallest coordinate value is taken as the upper edge line; The rotation matrix is ​​calculated by rotating the upper edge of the eyelashes to a horizontal position. R Then based on the rotation matrix R Rotate the entire image; Step 3: Use the Canny edge detection algorithm to detect contours in the inverted binarized image. Select candidate contours based on their area and shape. Sort the corner points by using the rightmost point of the candidate contour as the corner point and remove the sorted interference data to determine the glue application point. Step 4: Calculate the position of the glue application point in the original image based on the rotation matrix obtained from the edge line detection, and then fit the glue application line based on the glue application point.

2. The method for detecting false eyelash glue lines based on corner point sorting as described in claim 1, characterized in that, The expressions for the inverse binarization and binarization operations are as follows: ， ， in, G(x,y) It is in the result image ( x , y The grayscale value at ) g(x,y) In grayscale images ( x , y The grayscale value at ) T It is a specified threshold that is adjusted according to lighting conditions. x, y These represent the horizontal and vertical coordinates of the coordinate axes in the image, respectively.

3. The method for detecting false eyelash glue lines based on corner point sorting as described in claim 1, characterized in that, Step three specifically includes the following sub-steps: Step 3.1: Use the Canny edge detection algorithm to find contours in the rotated grayscale image. Select candidate contours by setting minimum and maximum thresholds for contour area, height difference, and width difference. Step 3.2, for each candidate contour's point set, according to... x Sort the coordinate values, take the rightmost point of each contour as the corner point, and then... y The coordinate values ​​are sorted by corner points, and the noisy data at the beginning of the sorted result is removed, followed by the continuous data. N These data points are considered as adhesive application points. N This is an integer value that is dynamically adjusted based on eyelash type.

4. The method for detecting false eyelash glue lines based on corner point sorting as described in claim 3, characterized in that, The method for removing interfering data from the head of the sorted data is specifically as follows: adjacent corner points y If the difference in values ​​is greater than the set threshold and the number of consecutive corner points is less than 20, then the preceding corner point data is considered to be interference data and is removed.

5. The method for detecting false eyelash glue lines based on corner point sorting as described in claim 3, characterized in that, Step four specifically involves rotating the adhesive application point, i.e., based on the calculated rotation matrix. R Calculate the position of the glue application point in the original image; then use the least squares method to fit the position of the glue application line.

Images