CN110148125A - Adaptive skin oil and fat detection method based on color detection - Google Patents

Abstract

The invention relates to the field of skin detection, in particular to an adaptive skin oil detection method based on color detection. The present invention firstly detects the pore area, and further detects the oil in the pore area, eliminating the possible interference in the non-pore area, and at the same time, by detecting the oil in the pores of the skin, the size and area of the oil in the pores is used as an index to measure the oiliness of the skin , which can well reflect the exuberant degree of human sebaceous gland secretion, which is more accurate and meaningful than the single measurement method of skin surface oil; the oil appears red under ultraviolet light, and the image formed under ultraviolet light source is processed It is more accurate to detect red, supplemented by detection on RGB images, than to directly detect red on RGB images.

Description

Adaptive skin oil detection method based on color detection

technical field

The invention relates to the field of skin detection, in particular to an adaptive skin oil detection method based on color detection.

Background technique

With the continuous expansion and application of computer image processing technology to the medical field, people pay more attention to dermatological cosmetology and the incidence of skin diseases. The research on the method of skin detection using image processing technology is increasingly accepted and valued by people. At present, both home and abroad have begun to study the skin analysis system to detect the quality of the skin. As an important indicator to measure skin quality, oil is one of the references for professionals to maintain facial skin. However, there are limited methods for skin oil detection at home and abroad, and most of them detect the oil on the skin surface, not from the source of oil. , the detection accuracy is low, so it is of great significance to develop an efficient and convenient skin oil detection method.

Contents of the invention

The technical problem to be solved by the present invention is to provide an adaptive skin oil detection method based on color detection that makes the measurement more accurate, efficient and convenient by detecting oil in the pore area.

In order to solve the above technical problems, the technical solution adopted by the present invention to solve the technical problems is:

A color detection-based adaptive skin oil detection method, the specific steps comprising:

S1, respectively irradiating the detection site with a white light source and an ultraviolet light source to obtain image A under the white light source and image B under the ultraviolet light source; S2, analyze and process the image A, and obtain the pore area information of the detection site; S3, Extract and process the corresponding area information of the pore area information in image A in image B, extract the pixels of the pore area and extract the pixels at the corresponding positions of a single-channel image C of the same size; S4, search image C Connected domain in , and count the number and area of grease.

Preferably, step S2 includes analyzing and processing image A including the following steps:

S21, converting image A into a grayscale image; S22, performing subregional dynamic threshold segmentation processing on the grayscale image; S23, performing grayscale inversion on the image after subregional dynamic threshold segmentation processing, and then corroding and expanding, and the remaining The white pixels are the pore area of image A.

Preferably, in step S21, after the image A is converted into a grayscale image, median filtering is performed on the grayscale image to eliminate isolated noise points.

Preferably, in step S22, the sub-area dynamic threshold segmentation processing of the grayscale image is specifically: performing dynamic threshold segmentation on the largest inscribed circle area and the remaining area in the grayscale image respectively, and merging the darker areas of the two areas , and then perform dynamic threshold segmentation on the merged area to obtain the optimal threshold and segment the entire grayscale image.

Preferably, in step S23, the image is corroded and expanded to remove small isolated points in the non-pore area of the image and fill the holes.

Preferably, in step S3, the specific steps of processing the corresponding area information in image B include:

S31. Assign (0, 0, 0) to the pixel value of the non-pore area in image B to obtain image B1; S32. Convert image B1 from RGB space to HSV space.

Preferably, in step S3, the specific steps of extracting the pixels in the pore area include:

S33, in the image B1 converted to HSV space, extract the pixels whose H component is in the range of (0, 25) or (156, 180), the S component is in the range of (70, 255), and the V component is in the range of (70, 255). point; S34, assign the corresponding position of the pixel obtained in step S33 in the image C to 255, and assign 0 to other positions.

Preferably, in step S4, image C is firstly subjected to erosion and dilation processing, and connected domains are searched on image C after erosion and dilation processing.

Preferably, in step S33, the pixel points whose R component in the corresponding position in the image B1 is smaller than 200 are eliminated from the extracted pixel points.

Beneficial effects of the present invention:

The present invention firstly detects the pore area, and further detects the oil in the pore area, eliminating the possible interference in the non-pore area, and at the same time, by detecting the oil in the pores of the skin, the size and area of the oil in the pores is used as an index to measure the oiliness of the skin , which can well reflect the exuberant degree of human sebaceous gland secretion, which is more accurate and meaningful than the single measurement method of skin surface oil; oil appears red under ultraviolet light, and the image formed under ultraviolet light source is processed It is more accurate to detect red, supplemented by detection on RGB images, than to directly detect red on RGB images.

Description of drawings

Fig. 1 is a flowchart of an adaptive skin oil detection method based on color detection in the present invention.

Fig. 2 is a schematic diagram of the sub-area dynamic threshold segmentation of the present invention.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, so that those skilled in the art can better understand the present invention and implement it, but the examples given are not intended to limit the present invention.

Referring to Figure 1, a color detection-based adaptive skin oil detection method, the specific steps include:

S1, respectively irradiating the detection site with a white light source and an ultraviolet light source to obtain image A under the white light source and image B under the ultraviolet light source; S2, analyze and process the image A, and obtain the pore area information of the detection site; S3, Extract and process the corresponding area information of the pore area information in image A in image B, extract the pixels of the pore area and extract the pixels at the corresponding positions of a single-channel image C of the same size; S4, search image C Connected domain in , and count the number and area of grease.

The present invention firstly detects the pore area, and further detects the oil in the pore area, eliminating the possible interference in the non-pore area, and at the same time, by detecting the oil in the pores of the skin, the size and area of the oil in the pores is used as an index to measure the oiliness of the skin , which can well reflect the exuberant degree of human sebaceous gland secretion, which is more accurate and meaningful than the single measurement method of skin surface oil; oil appears red under ultraviolet light, and the image formed under ultraviolet light source is processed It is more accurate to detect red, supplemented by detection on RGB images, than to directly detect red on RGB images.

Step S2 includes analyzing and processing image A including the following steps:

S21, converting the image A (RGB image) into a grayscale image; S22, performing subregional dynamic threshold segmentation processing on the grayscale image; S23, performing grayscale inversion on the subregional dynamic threshold segmentation processed image, and then corroding and expanding , and the remaining white pixels are the pore area of image A.

In step S22, the dynamic threshold segmentation of the grayscale image by region can eliminate the influence of uneven illumination of the standard white light source, and avoid the phenomenon that the middle circular area is brighter and the surrounding area is darker due to uneven illumination of the light source. The dynamic threshold segmentation Methods include, but are not limited to, the maximum between-class variance method.

In step S21, after the image A is converted into a grayscale image, median filtering is performed on the grayscale image to eliminate isolated noise points.

Referring to Fig. 2, in step S22, the grayscale image is subjected to sub-area dynamic threshold segmentation processing specifically as follows: dynamic threshold segmentation is performed on the two regions of the largest inscribed circle region and the remaining region in the grayscale image respectively, and the two regions Merge the darker areas in the middle, and then perform dynamic threshold segmentation on the merged area to obtain the optimal threshold T, and segment the entire grayscale image.

In step S23, the image is corroded and expanded to remove small isolated points in the non-pore area of the image and fill the holes.

In step S3, the specific steps of processing the corresponding area information in image B include:

S31. Assign (0, 0, 0) to the pixel value of the non-pore area in image B to obtain image B1; S32. Convert image B1 from RGB space to HSV space.

Grease appears red under ultraviolet irradiation, based on this, the present invention converts the RGB image into an HSV image, and detects red based on the HSV image, supplemented by detection on the RGB image, which is more accurate than directly detecting red on the RGB image

In step S3, the specific steps of extracting the pixels in the pore area include:

S33. Grease appears red under ultraviolet irradiation, and the H component is extracted from the image B1 converted to HSV space in the range of (0, 25) or (156, 180), the S component is in the range of (70, 255), and the V component is in the range of ( 70, 255) within the range of pixels; S34, assign the corresponding position of the pixel obtained in step S33 in the image C to 255, and assign 0 to other positions.

In step S4, the image C is first corroded and dilated, and the connected domain is searched on the image C after the erosion and dilation process.

In step S33, the pixels whose R component in the corresponding position in the image B1 is less than 200 are eliminated from the extracted pixels

Specifically, the white light source of the present invention is a standard white light source, and the images of the detection site (skin to be detected) under the irradiation of the standard white light source and the ultraviolet light source are respectively captured by a camera or other shooting equipment, which are recorded as image A and image B. Detect the area position of pores in image B, and then detect the number and area of oil in image B to complete the detection.

The above-mentioned embodiments are only preferred embodiments for fully illustrating the present invention, and the protection scope of the present invention is not limited thereto. Equivalent substitutions or transformations made by those skilled in the art on the basis of the present invention are all within the protection scope of the present invention. The protection scope of the present invention shall be determined by the claims.

Claims (9)

1. a kind of adaptive skin oil and fat detection method based on color detection, which is characterized in that specific steps include:

S1, detection position is irradiated respectively by white light source and ultraviolet light source and is obtained image A under white light source and The image B under ultraviolet light source；S2, image A is analyzed and processed, obtains the pore area information of detection position；S3, extraction It corresponding region information of the pore area information in image B and is handled in image A, the pixel for extracting pore region simultaneously will Pixel extracts on the corresponding position of the single channel image C of a secondary same size；Connected domain in S4, lookup image C, and unite Count the number and area of grease.

2. the adaptive skin oil and fat detection method based on color detection as described in claim 1, which is characterized in that step S2 Including being analyzed and processed to image A the following steps are included:

S21, image A is switched into gray level image；S22, dynamic threshold segmentation processing in subregion is carried out to gray level image；S23, will divide Regional dynamics Threshold segmentation treated image carries out gray inversion, then corrosion expansion, and remaining white pixel point is then figure As the pore region of A.

3. the adaptive skin oil and fat detection method based on color detection as claimed in claim 2, which is characterized in that step In S21, after image A is switched to gray level image, median filter process is carried out to gray level image and eliminates isolated noise spot.

4. the adaptive skin oil and fat detection method based on color detection as claimed in claim 2, which is characterized in that step In S22, dynamic threshold segmentation processing in subregion is carried out to gray level image specifically: to maximum inscribed circle region in gray level image and Remaining area carries out dynamic threshold segmentation respectively, and region darker in two regions is merged, then to the region after merging into Mobile state Threshold segmentation obtains optimal threshold, divides whole picture gray level image.

5. the adaptive skin oil and fat detection method based on color detection as claimed in claim 2, which is characterized in that step In S23, Image erosion is expanded, remove the smaller isolated point in non-pore region in image and fills up hole.

6. the adaptive skin oil and fat detection method based on color detection as described in claim 1, which is characterized in that step S3 In, the specific steps to the corresponding region information processing in image B include:

S31, to the non-pore area pixel value assignment (0,0,0) in image B, obtain image B1；S32, by image B1 from RGB sky Between be transformed into HSV space.

7. the adaptive skin oil and fat detection method based on color detection as described in claim 1, which is characterized in that in step In S3, the specific steps for extracting the pixel in pore region include:

S33, in the image B1 for being transformed into HSV space extract H component (0,25) or (156,180) range, S component (70, 255) pixel of range, V component in (70,255) range；S34, by pixel obtained in step S33 in image C Corresponding position is assigned a value of 255, and other positions are assigned a value of 0.

8. the adaptive skin oil and fat detection method based on color detection as described in claim 1, which is characterized in that step S4 In, corrosion expansion process is carried out to image C first, searches connected domain on image C after corroding expansion process.

9. the adaptive skin oil and fat detection method based on color detection as claimed in claim 7, which is characterized in that step In S33, from the pixel of extraction reject image B1 in pixel of the corresponding position R component less than 200.