Skin detection method and device based on multispectrum

Abstract

The invention discloses a multispectral-based skin detection method and device, and relates to the technical field of skin detection.The multispectral-based skin detection method comprises the steps that a plurality of spectral light sources are used for shooting a sample skin state to obtain a sample spectrogram, and the sample spectrogram is stored to serve as a data basis; establishing a display influence model of humidity on the spectral color, and establishing a display influence model of temperature on the spectral color; uniformly dividing the surface of the skin to be detected into skin identification blocks; calculating a first color difference value of the actual multispectral image to obtain a first adjusted image; calculating a second color difference value of the actual multispectral image to obtain a second adjusted image; and obtaining the existence condition of the actual skin state in the second adjusted image. Through a model building module, a color compensation module and a detection analysis module, an actual skin state is obtained and compared with a database skin graph, a skin care scheme with similar qualification is found out, and targeted health care and nursing are carried out on the skin.

Description

Technical Field

[0001] The present invention relates to the technical field of skin detection, and in particular to a multi-spectrum-based skin detection method and device. Background Art

[0002] In recent years, with the improvement of people's living standards, people's health awareness and pursuit of beauty have been continuously improved, and skin care and health care have become a growing industry. At the same time, consumers have demands for various skin characteristic parameters, among which, they are particularly concerned about the skin condition. The use of multispectral images can complete the detection of skin.

[0003] However, the detection of skin condition is relatively complicated. When multiple abnormal conditions occur at the same time and the abnormal conditions are irregularly distributed, it is easy to misjudge the abnormal conditions. At the same time, different skin conditions can easily lead to differences in multispectral images of the same abnormal condition, making it impossible to carry out targeted health care. Summary of the invention

[0004] In order to solve the above technical problems, a multi-spectral based skin detection method and device are provided. This technical solution solves the problems raised in the above background technology.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] A multi-spectral based skin detection method, comprising:

[0007] Obtain at least one sample skin condition and a repair plan for the sample skin condition, photograph the sample skin condition with multiple sets of spectral light sources, obtain a sample spectrum graph, and store the sample spectrum graph to form a database as a data basis, wherein the acquisition condition of the sample spectrum graph is that the skin surface temperature is equal to the reference temperature and the skin surface humidity is equal to the reference humidity;

[0008] Establish a model for the effect of humidity on the display of spectral colors, and establish a model for the effect of temperature on the display of spectral colors;

[0009] Evenly divide the skin surface to be detected into at least one skin recognition block, obtain the actual temperature of the skin recognition block, obtain the actual humidity of the skin recognition block, and obtain the actual multispectral image of the skin recognition block;

[0010] Based on the display influence model of humidity on spectral color, a first color difference value of the actual multi-spectral image is calculated, and the color of the actual multi-spectral image is adjusted according to the first color depth difference value to obtain a first adjusted image;

[0011] Based on the display influence model of temperature on spectral color, a second color difference value of the actual multi-spectral image is calculated, and the color of the first adjusted image is adjusted according to the second color depth difference value to obtain a second adjusted image;

[0012] Based on the database, the second adjusted image is analyzed to obtain the existence of the actual skin condition in the second adjusted image;

[0013] A repairing solution for the actual skin condition is determined according to the existence of the actual skin condition of at least one second adjusted image.

[0014] Preferably, the method of photographing the sample skin condition using multiple groups of spectral light sources to obtain the sample spectrum graph and storing it to form a database comprises the following steps:

[0015] Get 8 spectra, 8 spectra for skin dryness, wrinkles, dark spots, acne scars, plumpness, oil, pores and aging;

[0016] The eight spectra were randomly combined, and each combination was regarded as a spectral light source;

[0017] The high-definition camera takes pictures of the skin under the irradiation of spectral light source and stores them in a computer, mobile phone or cloud to form a database. The spectral light source is scanned and the data storage and transmission are carried out wirelessly.

[0018] Preferably, the step of establishing a model of the effect of humidity on the display of spectral colors comprises the following steps:

[0019] Obtaining a value range of skin humidity, dividing the value range of skin humidity at equal intervals, and obtaining at least one humidity point;

[0020] Superimpose the value at the humidity point with the reference humidity to obtain the characteristic humidity;

[0021] Under the condition that the skin humidity is characteristic humidity, a humidity spectrum image of the sample skin is obtained;

[0022] Under the condition that the skin humidity is the reference humidity, a reference spectrum image of the sample skin is obtained, and the image range of the humidity spectrum image is consistent with that of the reference spectrum image;

[0023] The humidity influence color formula is used to calculate the first color deviation of the humidity spectrum image;

[0024] The value at the humidity point is paired with the first color deviation degree and fitted to obtain a humidity fitting function, wherein the value at the humidity point is an independent variable and the first color deviation degree is a dependent variable;

[0025] The formula for humidity affecting color is as follows:

[0026]

[0027] Wherein, A is the first color deviation, B is the sum of the pixel values ​​of all pixels in the humidity spectrum image, C is the sum of the pixel values ​​of all pixels in the reference spectrum image, and n is the total number of pixels in the reference spectrum image.

[0028] Preferably, the step of establishing a temperature-spectral-color display influence model comprises the following steps:

[0029] Acquire a value range of the skin temperature, divide the value range of the skin temperature into equal intervals, and obtain at least one temperature point;

[0030] Superimpose the value at the temperature point with the reference temperature to obtain the characteristic temperature;

[0031] Under the condition that the skin temperature is a characteristic temperature, a temperature spectrum image of the sample skin is obtained;

[0032] Under the condition that the skin temperature is the reference temperature, a reference spectrum image of the sample skin is obtained, and the image range of the temperature spectrum image is consistent with that of the reference spectrum image;

[0033] Using the temperature-affected color formula, the second color shift degree of the temperature spectrum image is calculated;

[0034] Pairing and fitting the value at the temperature point with the second color shift degree to obtain a temperature fitting function, wherein the value at the temperature point is an independent variable and the second color shift degree is a dependent variable;

[0035] The temperature affects the color formula as follows:

[0036]

[0037] Wherein, D is the second color deviation, E is the sum of the pixel values ​​of all pixels in the temperature spectrum image, F is the sum of the pixel values ​​of all pixels in the reference spectrum image, and m is the total number of pixels in the reference spectrum image.

[0038] Preferably, the step of uniformly dividing the skin surface to be detected into at least one skin identification block comprises the following steps:

[0039] Obtain the allowable temperature error and allowable humidity error during skin detection;

[0040] Evenly dividing the maximum size of the skin surface to obtain at least one size point;

[0041] According to the size point, the skin is divided into at least one area block, and the side length of the area block is equal to the value at the size point;

[0042] Perform temperature measurement and humidity measurement in the area block. When the temperature fluctuation amplitude in the area block is less than the allowable temperature error and the humidity fluctuation amplitude in the area block is less than the allowable humidity error, the area block is taken as the target area block.

[0043] When the area blocks divided according to the size point are all target area blocks, the size point is taken as the target size point;

[0044] The maximum value of the target size point is used as the side length of the skin recognition block, and the skin surface to be detected is evenly divided into at least one skin recognition block according to the side length of the skin recognition block, wherein the skin recognition block is a square.

[0045] Preferably, the step of obtaining the actual multispectral image of the skin recognition block comprises the following steps:

[0046] The skin recognition block is illuminated by a spectral light source and photographed to obtain an actual multispectral image of the skin recognition block.

[0047] Preferably, the calculating and obtaining the first color difference value of the actual multispectral image, and adjusting the color of the actual multispectral image according to the first color depth difference value comprises the following steps:

[0048] Substituting the actual humidity of the skin recognition block into the humidity fitting function to obtain a first color difference value;

[0049] Superimposing the first color difference value with the pixel value of the pixel point in the actual multispectral image to obtain a first corrected pixel value;

[0050] The pixel values ​​of the pixels in the actual multispectral image are corrected to the first corrected pixel values, and the first adjusted image is redisplayed.

[0051] Preferably, the calculating and obtaining the second color difference value of the actual multispectral image, and adjusting the color of the first adjusted image according to the second color depth difference value comprises the following steps:

[0052] Substituting the actual temperature of the skin recognition block into the temperature fitting function to obtain a second color difference value;

[0053] Superimposing the second color difference value with the pixel value of the pixel point in the first adjusted image to obtain a second corrected pixel value;

[0054] The pixel values ​​of the pixels in the first adjusted image are corrected to second corrected pixel values, and the second adjusted image is redisplayed.

[0055] Preferably, analyzing the second adjusted image to obtain the actual skin condition in the second adjusted image comprises the following steps:

[0056] Obtaining the area of ​​the second adjusted image and dividing it by the area of ​​the sample spectrum graph to obtain a target zoom ratio;

[0057] Use 0 and the target scaling ratio as the endpoints of the interval to form a scaling interval, and evenly select at least one scaling point in the scaling interval;

[0058] Scaling the sample spectrum graph according to the value at the scaling point to obtain a first sample spectrum graph;

[0059] Use 0 and 360 degrees as the endpoints of the interval to form a rotation interval, and evenly select at least one rotation point in the rotation interval;

[0060] Rotate the image in the first sample spectrum graph according to the value at the rotation point to obtain the second sample spectrum graph, and match the value at the zoom point corresponding to the first sample spectrum graph to the second sample spectrum graph;

[0061] Acquire a second sample spectrum graph appearing in the second adjusted image as a target sample spectrum graph;

[0062] The sample skin condition corresponding to the target sample spectrum is taken as the actual skin condition and summarized.

[0063] A multi-spectral skin detection device, used to implement the multi-spectral skin detection method, comprising:

[0064] A sample acquisition module, which acquires at least one sample skin condition and a repair plan for the sample skin condition, photographs the sample skin condition with multiple sets of spectral light sources, obtains a sample spectrum graph, and stores the sample spectrum graph to form a database as a data basis. The acquisition conditions of the sample spectrum graph are that the skin surface temperature is equal to the reference temperature and the skin surface humidity is equal to the reference humidity;

[0065] A model building module, wherein the model building module builds a model of the influence of humidity on the display of spectral colors and a model of the influence of temperature on the display of spectral colors;

[0066] A data acquisition module, wherein the data acquisition module evenly divides the surface of the skin to be detected into at least one skin recognition block, obtains the actual temperature of the skin recognition block, obtains the actual humidity of the skin recognition block, and obtains the actual multispectral image of the skin recognition block;

[0067] a color compensation module, wherein the color compensation module calculates a first color difference value of the actual multispectral image based on a display influence model of humidity on spectral color, adjusts the color of the actual multispectral image according to the first color depth difference value to obtain a first adjusted image, and calculates a second color difference value of the actual multispectral image based on a display influence model of temperature on spectral color, adjusts the color of the first adjusted image according to the second color depth difference value to obtain a second adjusted image;

[0068] A detection and analysis module analyzes the second adjustment image based on a database to obtain the existence of the actual skin condition in the second adjustment image, and determines a repair plan for the actual skin condition based on the existence of the actual skin condition of at least one second adjustment image.

[0069] Compared with the prior art, the present invention has the following beneficial effects:

[0070] Through the model building module, the color compensation module and the detection and analysis module, a model of the display influence of humidity on spectral color and a model of the display influence of temperature on spectral color can be established, thereby eliminating the influence of humidity and temperature on the image, and determining the existence of the actual skin condition in the second adjusted image by rotation and scaling, thereby avoiding the situation where abnormal conditions are ignored due to irregularities, and thus obtaining the actual skin condition in a more comprehensive manner, thereby providing targeted health care and treatment for the skin. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 Schematic diagram of the process of the multi-spectral based skin detection method of the present invention;

[0072] Figure 2 A schematic diagram of the process of photographing the skin condition of a sample using multiple sets of spectral light sources, obtaining a sample spectrum graph, and storing the spectrum graph to form a database according to the present invention;

[0073] Figure 3 A schematic diagram of a flow chart of establishing a model of the effect of humidity on the display of spectral colors according to the present invention;

[0074] Figure 4 A schematic diagram of a flow chart of establishing a model of the effect of temperature on the display of spectral color according to the present invention;

[0075] Figure 5 It is a schematic diagram of a process of uniformly dividing the surface of the skin to be detected into at least one skin recognition block according to the present invention;

[0076] Figure 6 A schematic diagram of a process of calculating a first color difference value of an actual multispectral image and adjusting the color of the actual multispectral image according to the first color depth difference value according to the present invention;

[0077] Figure 7 A schematic diagram of a process of calculating a second color difference value of an actual multi-spectral image and adjusting the color of a first adjustment image according to the second color depth difference value according to the present invention;

[0078] Figure 8The present invention is a flow chart of analyzing the second adjusted image to obtain the actual skin condition in the second adjusted image. DETAILED DESCRIPTION

[0079] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0080] Reference Figure 1 As shown, a multi-spectral based skin detection method comprises:

[0081] Obtain at least one sample skin condition and a repair plan for the sample skin condition, photograph the sample skin condition with multiple sets of spectral light sources, obtain a sample spectrum graph, and store the sample spectrum graph to form a database as a data basis, wherein the acquisition condition of the sample spectrum graph is that the skin surface temperature is equal to the reference temperature and the skin surface humidity is equal to the reference humidity;

[0082] Establish a model for the effect of humidity on the display of spectral colors, and establish a model for the effect of temperature on the display of spectral colors;

[0083] Evenly divide the skin surface to be detected into at least one skin recognition block, obtain the actual temperature of the skin recognition block, obtain the actual humidity of the skin recognition block, and obtain the actual multispectral image of the skin recognition block;

[0084] Based on the display influence model of humidity on spectral color, a first color difference value of the actual multi-spectral image is calculated, and the color of the actual multi-spectral image is adjusted according to the first color depth difference value to obtain a first adjusted image;

[0085] Based on the display influence model of temperature on spectral color, a second color difference value of the actual multi-spectral image is calculated, and the color of the first adjusted image is adjusted according to the second color depth difference value to obtain a second adjusted image;

[0086] Based on the database, the second adjusted image is analyzed to obtain the existence of the actual skin condition in the second adjusted image;

[0087] A repairing solution for the actual skin condition is determined according to the existence of the actual skin condition of at least one second adjusted image.

[0088] When testing the skin, since there are various problems with the skin, such as tone, scars, size of pores, color, etc., these problems use different spectra for image acquisition, and the degree of obviousness of their characteristics is different. Therefore, when identifying, the spectrum that makes the feature most easily recognizable is usually used to acquire the image. Therefore, it is necessary to use multiple different spectra to acquire images of the skin to be tested, and compare them with the sample spectrum graph of the sample skin condition in turn, so as to obtain the problems.

[0089] When acquiring multispectral images, temperature and humidity can cause subtle changes in the displayed color of the skin. For example, when a person's body temperature rises due to exercise, the skin will turn red due to the expansion of blood vessels. These changes will interfere with the recognition of the image of the skin state. Therefore, a corresponding model needs to be established to eliminate this influence.

[0090] In addition, since the recognition process relies on the sample spectrogram, which corresponds to the sample skin condition, the actual skin condition distribution may not be as shown in the sample spectrogram. It may be that one part is more serious and another part is less serious. Moreover, it may be rotated or scaled with the corresponding sample spectrogram. Therefore, it is difficult to directly obtain the result during the recognition process.

[0091] Therefore, in this solution, corresponding algorithms are set to carry out targeted processing of these situations, thereby obtaining a more accurate skin condition and then carrying out targeted care.

[0092] Reference Figure 2 As shown, using multiple groups of spectral light sources to photograph the sample skin condition, obtaining the sample spectrum graph, and storing it to form a database includes the following steps:

[0093] Get 8 spectra, 8 spectra for skin dryness, wrinkles, dark spots, acne scars, plumpness, oil, pores and aging;

[0094] The eight spectra were randomly combined, and each combination was regarded as a spectral light source;

[0095] The high-definition camera takes pictures of the skin under the irradiation of spectral light source and stores them in a computer, mobile phone or cloud to form a database. The spectral light source is scanned and the data storage and transmission are carried out wirelessly.

[0096] The problems with the sample skin condition are known, and based on big data, for each existing problem, there is a spectrum that makes the problem easiest to identify. Therefore, the spectrum is corresponded to the sample skin condition and used as the image acquisition condition for the sample spectrum graph of the sample skin condition. Here, only 8 spectra are listed, but the spectrum can be increased as the number of detection problems increases, which is completely similar to the above method.

[0097] Reference Figure 3 As shown, establishing a model for the effect of humidity on spectral color display includes the following steps:

[0098] Obtaining a value range of skin humidity, dividing the value range of skin humidity at equal intervals, and obtaining at least one humidity point;

[0099] Superimpose the value at the humidity point with the reference humidity to obtain the characteristic humidity;

[0100] Under the condition that the skin humidity is characteristic humidity, a humidity spectrum image of the sample skin is obtained;

[0101] Under the condition that the skin humidity is the reference humidity, a reference spectrum image of the sample skin is obtained, and the image range of the humidity spectrum image is consistent with that of the reference spectrum image;

[0102] The humidity influence color formula is used to calculate the first color deviation of the humidity spectrum image;

[0103] The value at the humidity point is paired with the first color deviation degree and fitted to obtain a humidity fitting function, wherein the value at the humidity point is an independent variable and the first color deviation degree is a dependent variable;

[0104] The formula for humidity affecting color is as follows:

[0105]

[0106] Wherein, A is the first color deviation, B is the sum of the pixel values ​​of all pixels in the humidity spectrum image, C is the sum of the pixel values ​​of all pixels in the reference spectrum image, and n is the total number of pixels in the reference spectrum image.

[0107] Different humidity has different effects on the display of spectral colors, because humidity means water droplets, which will affect the color of light. For example, the surface of dry skin is not as bright as the surface of wet skin. Therefore, in order to eliminate the influence of humidity on color, this model is established to eliminate the influence of color under different humidity conditions. As a result, the accuracy can be improved during recognition. The same is true for temperature. The surface temperature of the body has a great relationship with the human body itself, because humans are warm-blooded animals and their own temperature will not change due to external temperature. The change of their own temperature is reflected in blood flow. Blood flow will cause the skin color to tend to turn red. Therefore, this situation needs to be eliminated during recognition. Since the blood flow in blood vessels is related to temperature, a model for the influence of temperature on the display of spectral colors can be established.

[0108] Reference Figure 4 As shown, establishing a model for the effect of temperature on spectral color display includes the following steps:

[0109] Acquire a value range of the skin temperature, divide the value range of the skin temperature into equal intervals, and obtain at least one temperature point;

[0110] Superimpose the value at the temperature point with the reference temperature to obtain the characteristic temperature;

[0111] Under the condition that the skin temperature is a characteristic temperature, a temperature spectrum image of the sample skin is obtained;

[0112] Under the condition that the skin temperature is the reference temperature, a reference spectrum image of the sample skin is obtained, and the image range of the temperature spectrum image is consistent with that of the reference spectrum image;

[0113] Using the temperature-affected color formula, the second color shift degree of the temperature spectrum image is calculated;

[0114] Pairing and fitting the value at the temperature point with the second color shift degree to obtain a temperature fitting function, wherein the value at the temperature point is an independent variable and the second color shift degree is a dependent variable;

[0115] The temperature affects the color formula as follows:

[0116]

[0117] Wherein, D is the second color deviation, E is the sum of the pixel values ​​of all pixels in the temperature spectrum image, F is the sum of the pixel values ​​of all pixels in the reference spectrum image, and m is the total number of pixels in the reference spectrum image.

[0118] Reference Figure 5 As shown, evenly dividing the skin surface to be detected into at least one skin recognition block includes the following steps:

[0119] Obtain the allowable temperature error and allowable humidity error during skin detection;

[0120] Evenly dividing the maximum size of the skin surface to obtain at least one size point;

[0121] According to the size point, the skin is divided into at least one area block, and the side length of the area block is equal to the value at the size point;

[0122] Perform temperature measurement and humidity measurement in the area block. When the temperature fluctuation amplitude in the area block is less than the allowable temperature error and the humidity fluctuation amplitude in the area block is less than the allowable humidity error, the area block is taken as the target area block.

[0123] When the area blocks divided according to the size point are all target area blocks, the size point is taken as the target size point;

[0124] The maximum value of the target size point is used as the side length of the skin recognition block, and the skin surface to be detected is evenly divided into at least one skin recognition block according to the side length of the skin recognition block, wherein the skin recognition block is a square.

[0125] The skin recognition block needs to be divided as large as possible. In this way, the fewer the blocks, the less the retrieval amount will be. However, the size of the division needs to ensure that the humidity and temperature fluctuations within the skin recognition block are small and will not exceed the allowable temperature error and the allowable humidity error. Otherwise, when obtaining the temperature and humidity, the humidity and temperature fluctuations within the skin recognition block are large, and the impact on the color is also very different, resulting in the inability to use a unified standard for processing within the skin recognition block, which will increase the difficulty of processing.

[0126] Acquiring the actual multispectral image for the skin recognition patch involves the following steps:

[0127] The skin recognition block is illuminated by a spectral light source and photographed to obtain an actual multispectral image of the skin recognition block.

[0128] Reference Figure 6 As shown, calculating and obtaining a first color difference value of the actual multispectral image, and adjusting the color of the actual multispectral image according to the first color depth difference value include the following steps:

[0129] Substituting the actual humidity of the skin recognition block into the humidity fitting function to obtain a first color difference value;

[0130] Superimposing the first color difference value with the pixel value of the pixel point in the actual multispectral image to obtain a first corrected pixel value;

[0131] The pixel values ​​of the pixels in the actual multispectral image are corrected to the first corrected pixel values, and the first adjusted image is redisplayed.

[0132] Reference Figure 7 As shown, calculating and obtaining the second color difference value of the actual multispectral image, and adjusting the color of the first adjustment image according to the second color depth difference value includes the following steps:

[0133] Substituting the actual temperature of the skin recognition block into the temperature fitting function to obtain a second color difference value;

[0134] Superimposing the second color difference value with the pixel value of the pixel point in the first adjusted image to obtain a second corrected pixel value;

[0135] The pixel values ​​of the pixels in the first adjusted image are corrected to second corrected pixel values, and the second adjusted image is redisplayed.

[0136] Reference Figure 8As shown, analyzing the second adjusted image to obtain the existence of the actual skin condition in the second adjusted image includes the following steps:

[0137] Obtaining the area of ​​the second adjusted image and dividing it by the area of ​​the sample spectrum graph to obtain a target zoom ratio;

[0138] Use 0 and the target scaling ratio as the endpoints of the interval to form a scaling interval, and evenly select at least one scaling point in the scaling interval;

[0139] Scaling the sample spectrum graph according to the value at the scaling point to obtain a first sample spectrum graph;

[0140] Use 0 and 360 degrees as the endpoints of the interval to form a rotation interval, and evenly select at least one rotation point in the rotation interval;

[0141] Rotate the image in the first sample spectrum graph according to the value at the rotation point to obtain the second sample spectrum graph, and match the value at the zoom point corresponding to the first sample spectrum graph to the second sample spectrum graph;

[0142] Acquire a second sample spectrum graph appearing in the second adjusted image as a target sample spectrum graph;

[0143] The sample skin condition corresponding to the target sample spectrum is taken as the actual skin condition and summarized.

[0144] During identification, when the image obtained by rotating or scaling the sample spectrum or a combination of the two appears in the second adjusted image, it means that the sample skin condition corresponding to the sample spectrum appears in the second adjusted image. Therefore, the sample skin condition corresponding to the sample spectrum can be used as the actual skin condition. When repairing, it can be repaired in sequence according to the repair plan corresponding to the sample spectrum.

[0145] A multi-spectral skin detection device, used to implement the multi-spectral skin detection method, comprising:

[0146] A sample acquisition module, which acquires at least one sample skin condition and a repair plan for the sample skin condition, photographs the sample skin condition with multiple sets of spectral light sources, obtains a sample spectrum graph, and stores the sample spectrum graph to form a database as a data basis. The acquisition conditions of the sample spectrum graph are that the skin surface temperature is equal to the reference temperature and the skin surface humidity is equal to the reference humidity;

[0147] A model building module, wherein the model building module builds a model of the influence of humidity on the display of spectral colors and a model of the influence of temperature on the display of spectral colors;

[0148] A data acquisition module, wherein the data acquisition module evenly divides the surface of the skin to be detected into at least one skin recognition block, obtains the actual temperature of the skin recognition block, obtains the actual humidity of the skin recognition block, and obtains the actual multispectral image of the skin recognition block;

[0149] a color compensation module, wherein the color compensation module calculates a first color difference value of the actual multispectral image based on a display influence model of humidity on spectral color, adjusts the color of the actual multispectral image according to the first color depth difference value to obtain a first adjusted image, and calculates a second color difference value of the actual multispectral image based on a display influence model of temperature on spectral color, adjusts the color of the first adjusted image according to the second color depth difference value to obtain a second adjusted image;

[0150] A detection and analysis module analyzes the second adjustment image based on a database to obtain the existence of the actual skin condition in the second adjustment image, and determines a repair plan for the actual skin condition based on the existence of the actual skin condition of at least one second adjustment image.

[0151] Furthermore, the present solution also proposes a storage medium on which a computer-readable program is stored, and when the computer-readable program is called, the above-mentioned multi-spectrum-based skin detection method is executed.

[0152] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).

[0153] In summary, the advantages of the present invention are: through the model building module, the color compensation module and the detection and analysis module, a humidity display influence model on the spectral color and a temperature display influence model on the spectral color can be established, thereby eliminating the influence of humidity and temperature on the image, and determining the existence of the actual skin condition in the second adjustment image by rotation and scaling, thereby avoiding the situation where abnormal conditions are ignored due to irregularities, and thus the actual skin condition can be obtained more comprehensively, thereby providing targeted health care and care for the skin.

[0154] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A multi-spectral skin detection method, characterized in that: include: Obtain at least one sample skin condition and a repair plan for the sample skin condition, photograph the sample skin condition with multiple sets of spectral light sources, obtain a sample spectrum graph, and store the sample spectrum graph to form a database as a data basis, wherein the acquisition condition of the sample spectrum graph is that the skin surface temperature is equal to the reference temperature and the skin surface humidity is equal to the reference humidity; Establish a model for the effect of humidity on the display of spectral colors, and establish a model for the effect of temperature on the display of spectral colors; Evenly divide the surface of the skin to be detected into at least one skin recognition block, obtain the actual temperature of the skin recognition block, obtain the actual humidity of the skin recognition block, and obtain the actual multispectral image of the skin recognition block; Based on the display influence model of humidity on spectral color, a first color difference value of the actual multispectral image is calculated, and the color of the actual multispectral image is adjusted according to the first color depth difference value to obtain a first adjusted image; Based on the display influence model of temperature on spectral color, a second color difference value of the actual multi-spectral image is calculated, and the color of the first adjusted image is adjusted according to the second color depth difference value to obtain a second adjusted image; Based on the database, the second adjusted image is analyzed to obtain the existence of the actual skin condition in the second adjusted image; A repairing solution for the actual skin condition is determined according to the existence of the actual skin condition of at least one second adjusted image.

2. The multi-spectral skin detection method according to claim 1, characterized in that: The method of photographing the sample skin state with multiple groups of spectral light sources to obtain the sample spectrum graph and storing it to form a database includes the following steps: Get 8 spectra, 8 spectra for skin dryness, wrinkles, dark spots, acne scars, plumpness, oil, pores and aging; The eight spectra were randomly combined, and each combination was regarded as a spectral light source; The high-definition camera takes pictures of the skin under the irradiation of spectral light source and stores them in a computer, mobile phone or cloud to form a database. The spectral light source is scanned and the data storage and transmission are carried out wirelessly.

3. The multi-spectral skin detection method according to claim 2, characterized in that: The establishment of the display influence model of humidity on spectral color comprises the following steps: Obtaining a value range of skin humidity, dividing the value range of skin humidity at equal intervals, and obtaining at least one humidity point; Superimpose the value at the humidity point with the reference humidity to obtain the characteristic humidity; Under the condition that the skin humidity is characteristic humidity, a humidity spectrum image of the sample skin is obtained; Under the condition that the skin humidity is the reference humidity, a reference spectrum image of the sample skin is obtained, and the image range of the humidity spectrum image is consistent with that of the reference spectrum image; The humidity influence color formula is used to calculate the first color deviation of the humidity spectrum image; The value at the humidity point is paired with the first color deviation degree and fitted to obtain a humidity fitting function, wherein the value at the humidity point is an independent variable and the first color deviation degree is a dependent variable; The formula for humidity affecting color is as follows: Wherein, A is the first color deviation, B is the sum of the pixel values ​​of all pixels in the humidity spectrum image, C is the sum of the pixel values ​​of all pixels in the reference spectrum image, and n is the total number of pixels in the reference spectrum image.

4. The multi-spectral skin detection method according to claim 3, characterized in that: The establishment of the display influence model of temperature on spectral color comprises the following steps: Acquire a value range of the skin temperature, divide the value range of the skin temperature into equal intervals, and obtain at least one temperature point; Superimpose the value at the temperature point with the reference temperature to obtain the characteristic temperature; Under the condition that the skin temperature is a characteristic temperature, a temperature spectrum image of the sample skin is obtained; Under the condition that the skin temperature is the reference temperature, a reference spectrum image of the sample skin is obtained, and the image range of the temperature spectrum image is consistent with that of the reference spectrum image; Using the temperature-affected color formula, the second color shift degree of the temperature spectrum image is calculated; Pairing and fitting the value at the temperature point with the second color shift degree to obtain a temperature fitting function, wherein the value at the temperature point is an independent variable and the second color shift degree is a dependent variable; The temperature affects the color formula as follows: Wherein, D is the second color deviation, E is the sum of the pixel values ​​of all pixels in the temperature spectrum image, F is the sum of the pixel values ​​of all pixels in the reference spectrum image, and m is the total number of pixels in the reference spectrum image.

5. The multi-spectral skin detection method according to claim 4, characterized in that: The step of uniformly dividing the skin surface to be detected into at least one skin recognition block comprises the following steps: Obtain the allowable temperature error and allowable humidity error during skin detection; Evenly dividing the maximum size of the skin surface to obtain at least one size point; According to the size point, the skin is divided into at least one area block, and the side length of the area block is equal to the value at the size point; Perform temperature measurement and humidity measurement in the area block. When the temperature fluctuation amplitude in the area block is less than the allowable temperature error and the humidity fluctuation amplitude in the area block is less than the allowable humidity error, the area block is taken as the target area block. When the area blocks divided according to the size point are all target area blocks, the size point is taken as the target size point; The maximum value of the target size point is used as the side length of the skin recognition block, and the skin surface to be detected is evenly divided into at least one skin recognition block according to the side length of the skin recognition block, wherein the skin recognition block is a square.

6. The multi-spectral skin detection method according to claim 5, characterized in that: The method of obtaining the actual multispectral image of the skin recognition block comprises the following steps: The skin recognition block is illuminated by a spectral light source and photographed to obtain an actual multispectral image of the skin recognition block.

7. The multi-spectral skin detection method according to claim 6, characterized in that: The calculating and obtaining the first color difference value of the actual multispectral image, and adjusting the color of the actual multispectral image according to the first color depth difference value comprises the following steps: Substituting the actual humidity of the skin recognition block into the humidity fitting function to obtain a first color difference value; Superimposing the first color difference value with the pixel value of the pixel point in the actual multispectral image to obtain a first corrected pixel value; The pixel values ​​of the pixels in the actual multispectral image are corrected to the first corrected pixel values, and the first adjusted image is redisplayed.

8. The multi-spectral based skin detection method according to claim 7, characterized in that: The calculating and obtaining the second color difference value of the actual multispectral image, and adjusting the color of the first adjustment image according to the second color depth difference value comprises the following steps: Substituting the actual temperature of the skin recognition block into the temperature fitting function to obtain a second color difference value; Superimposing the second color difference value with the pixel value of the pixel point in the first adjusted image to obtain a second corrected pixel value; The pixel values ​​of the pixels in the first adjusted image are corrected to second corrected pixel values, and the second adjusted image is redisplayed.

9. The multi-spectral skin detection method according to claim 8, characterized in that: The analyzing the second adjusted image to obtain the actual skin condition in the second adjusted image comprises the following steps: Obtaining the area of ​​the second adjusted image and dividing it by the area of ​​the sample spectrum graph to obtain a target zoom ratio; Use 0 and the target scaling ratio as the endpoints of the interval to form a scaling interval, and evenly select at least one scaling point in the scaling interval; Scaling the sample spectrum graph according to the value at the scaling point to obtain a first sample spectrum graph; Use 0 and 360 degrees as the endpoints of the interval to form a rotation interval, and evenly select at least one rotation point in the rotation interval; Rotate the image in the first sample spectrum graph according to the value at the rotation point to obtain the second sample spectrum graph, and match the value at the zoom point corresponding to the first sample spectrum graph to the second sample spectrum graph; Acquire a second sample spectrum graph appearing in the second adjusted image as a target sample spectrum graph; The sample skin condition corresponding to the target sample spectrum is taken as the actual skin condition and summarized.

10. A multi-spectral skin detection device, used to implement the multi-spectral skin detection method according to any one of claims 1 to 9, characterized in that: include: A sample acquisition module, which acquires at least one sample skin condition and a repair plan for the sample skin condition, photographs the sample skin condition with multiple sets of spectral light sources, obtains a sample spectrum graph, and stores the sample spectrum graph to form a database as a data basis. The acquisition conditions of the sample spectrum graph are that the skin surface temperature is equal to the reference temperature and the skin surface humidity is equal to the reference humidity; A model building module, wherein the model building module builds a model of the influence of humidity on the display of spectral colors and a model of the influence of temperature on the display of spectral colors; A data acquisition module, wherein the data acquisition module evenly divides the surface of the skin to be detected into at least one skin recognition block, obtains the actual temperature of the skin recognition block, obtains the actual humidity of the skin recognition block, and obtains the actual multispectral image of the skin recognition block; a color compensation module, wherein the color compensation module calculates a first color difference value of the actual multispectral image based on a display influence model of humidity on spectral color, adjusts the color of the actual multispectral image according to the first color depth difference value to obtain a first adjusted image, and calculates a second color difference value of the actual multispectral image based on a display influence model of temperature on spectral color, adjusts the color of the first adjusted image according to the second color depth difference value to obtain a second adjusted image; A detection and analysis module analyzes the second adjustment image based on a database to obtain the existence of the actual skin condition in the second adjustment image, and determines a repair plan for the actual skin condition based on the existence of the actual skin condition of at least one second adjustment image.

Images