WO2021238843A1

WO2021238843A1 - Multi-spectrum image detection technique-based ambient color temperature testing method and system

Info

Publication number: WO2021238843A1
Application number: PCT/CN2021/095461
Authority: WO
Inventors: 任玉; 王朔; 聂刚; 刘晓慧; 周浩; 刘禹辰; 蔡红星; 姚治海
Original assignee: 吉林求是光谱数据科技有限公司
Priority date: 2020-05-25
Filing date: 2021-05-24
Publication date: 2021-12-02
Also published as: CN111551266B; CN111551266A

Abstract

A multi-spectrum image detection technique-based ambient color temperature testing method and a system. The method comprises: acquiring, by means of a spectrum chip (2), an image having a uniform brightness and multi-spectrum data of 1st, 2nd, 3rd,..., m color blocks in a reference color board, and obtaining N multi-spectrum data matrices for each of the color blocks; sequentially performing temporal and spatial noise reduction processing on the N multi-spectrum data matrices to obtain n pieces of multi-spectrum data, and sequentially determining, in this manner, reference multi-spectrum data x_st(n, I_CCT) for the m-th color block at 1st, 2nd, 3rd,..., I color temperature values; acquiring, by using a mobile phone or a camera comprising the color temperature testing system, multi-spectrum data x_et(n) of multiple specific color regions in a scene image, calculating an average of the multi-spectrum data in the same frequency band; and performing, by means of a color temperature matching module (5), standard deviation calculation with respect to ambient lighting multi-spectrum data and the reference multi-spectrum data x_st(n, I_CTT), wherein a color temperature value having the minimum deviation is an ambient color temperature value. The invention achieves fast, accurate, and efficient color temperature testing.

Description

Method and system for testing environmental color temperature based on multispectral image detection technology

Technical field

The invention relates to the technical field of spectrum identification and the technical fields of video monitoring and mobile terminals, and in particular to an environmental color temperature testing method and system based on multi-spectral image detection technology.

Background technique

Different light sources have different spectral components and distributions, which are called color temperature in colorimetry. Affected by the color temperature of the environment, the pictures taken by the image sensor often show color cast. For example, the pictures taken in a higher color temperature environment will be bluish, and the pictures taken in a lower color temperature environment will be reddish. In addition to the color cast in the picture, the light also causes the picture to be over-saturated or under-saturated, and the color of the picture is seriously distorted. At present, people use white balance adjustment technology to deduct or supplement the influence of ambient light to eliminate this color shift, but the existing adjustment technology is focused on the white balance algorithm, for example: [A CMOS image sensor signal processing automatic white balance algorithm , Fang Jianrong et al., Computer Engineering, Vol.41, No.9, 2015], [Scene Fusion System Based on Ambient Light Detection, Dong Yue et al., Acta Photonica Sinica, Vol.49 No.1, 2020], [Automatic based on color temperature estimation Research on white balance algorithm, Wang Min et al., Vol.22No.12,2011], its hardware uses RGB color temperature sensor, and this white balance adjustment technology only uses the three values of red, green and blue for calculation, which is complicated by the algorithm. Too high, lack of information, resulting in insufficient color temperature correction and other shortcomings.

In response to the above problems, some mobile phones currently use eight-channel multi-spectral color temperature sensors, use semantic color reproduction algorithms, and are equipped with AI deep learning algorithm libraries to achieve the measurement of ambient light color temperature (CCT) under complex light sources and backgrounds. However, this technology uses a single-point sensor to sum and average all light signals. When the color temperature of the incident light source is different in different directions or positions, it cannot be effectively distinguished. Therefore, the environmental applicability is insufficient and the accuracy is low.

Summary of the invention

In view of the above problems, the purpose of the present invention is to provide an environmental color temperature testing method based on multispectral image detection technology, which uses a spectrum chip to achieve rapid, accurate and efficient color temperature detection.

In order to achieve the above objectives, the present invention is achieved by adopting the following technical solutions:

An environmental color temperature test method based on multi-spectral image detection technology, which specifically includes the following steps:

Step S1. Place a standard color plate containing m standard colors under a color temperature standard lamp, and adjust the distance between the color temperature standard lamp and the standard color plate, and the distance between the standard color plate and the spectrum chip; wherein, the color temperature output by the color temperature standard lamp The value is 1; the spectroscopic chip includes a spectroscopic element and an image sensor, and the spectroscopic element is used to divide incident light into n wavebands, and transmit radiation information on different spectral wavebands to the light sensing surface of the underlying image sensor, Make the captured image include n photos of different spectral bands, so that for a certain pixel area in the image, corresponding to n spectral data, that is, the multi-spectral data collection of a certain area of the target is realized;

Step S2. Start the color temperature standard lamp and the spectrum chip. The light waves emitted by the color temperature standard lamp irradiate the standard color plate. The color temperature standard lamp presents a uniform spot on the standard color plate. The light reflected by the standard color plate enters the spectrum chip, and the spectrum chip Obtain images and multispectral data with uniform brightness for the 1.2.3....m color squares in the standard color palette, and continuously collect N times, that is, N sets of multispectral data matrices are obtained for each color square;

Step S3: Perform temporal and spatial noise reduction processing on the N sets of multispectral data matrices of each color square. The temporal noise reduction processing method is to average the corresponding positions of the N sets of multispectral data matrices, and finally obtain an averaged multispectral data Matrix; spatial noise reduction processing is to average the multispectral data of the same band in the same color block area, and finally obtain n multispectral data, as the standard multispectral data of the m-th color block under the color temperature value, and store it in In the standard data module; in this way, determine the standard multispectral data xst(n, ICCT) of the m-th color square at the 1.2.3....I color temperature value in turn;

Step S4. Use a mobile phone or camera with a color temperature test system to obtain multi-spectral data x _et (n) in a variety of specific color regions in the scene picture, average the multi-spectral data of the same wavelength band, and then use the color temperature matching module to reduce the amount of ambient light Spectral data and standard multi-spectral data x _st (n, I _CCT ) are calculated for standard deviation. If the calculated standard deviation value is less than the standard deviation threshold St, it means that the color temperature of the ambient light in the area is equal to the standard multi-spectral data Corresponding color temperature value; if it is greater than St, then calculate the next column of data;

The standard deviation calculation formula is:

…

.

As a preference of the present invention, in step S4, when a specific color area in the scene picture is selected, white, blue, green, and red are the main ones.

Another object provided by the present invention is to provide an environmental color temperature testing system based on multispectral image detection technology. The system includes a control module, a spectrum chip, a data processing module, a standard data module, and a color temperature matching module; wherein the control module Connected with the spectrum chip and data processing module, used for the spectrum data acquisition of the spectrum chip and the start of the data processing module for command control;

The spectrum chip is connected to the data processing module, the images and multi-spectral data collected by the spectrum chip are sent to the data processing module, and the data processing module performs noise reduction processing; the spectrum chip includes a spectroscopic element and an image sensor, and the spectroscopic element is used to The incident light is divided into n bands to realize the function of spectral light splitting; the light splitting element transmits the radiation information of the same shooting target in different spectral bands to the light sensing surface of the underlying image sensor, so that the captured image includes n different spectra A photo of the band, so that for a certain pixel area in the image, it corresponds to n spectral data, that is, the multi-spectral data collection of a certain area of the target is realized; for obtaining the multi-spectral data of a certain area of the target image, all the pixels in the area are used The corresponding band is obtained by averaging, that is, the number of multispectral data corresponding to a certain area of the target image is n;

The data processing module is connected with the standard data module and the color temperature matching module;

The standard data module is connected to the color temperature matching module; the standard data module is used to pre-store the multi-spectral data corresponding to the standard color temperature value. The multi-spectral data includes m categories, corresponding to m standard colors, that is, one category represents one type Standard color, each category contains the standard multi-spectral number sequence corresponding to the I set of standard color temperature values, the number of standard multi-spectral data in each column is the total number of bands of the spectrum chip n, and the standard multi-spectral data of each category is x _st (n , I _CCT );

The color temperature matching module is connected to the control module, and is used to calculate the color temperature value of the ambient light in the captured image and transmit it to the control module. The calculation method is the standard variance method, that is, the color temperature value standard deviation threshold S _t is preset, Calculate the standard deviation of the ambient light multispectral data x _et (n) of each area in the captured image according to the category and the standard multispectral data x _st (n, I _CCT ) in the category according to the following formula, if the calculated standard If the variance value is less than St, it means that the color temperature value of the ambient light in the area is equal to the color temperature value corresponding to the standard multispectral data; if it is greater than St, proceed to the next column of data calculation;

The standard deviation calculation formula is:

…

.

As a preference of the present invention, the control module is also used to perform white balance calibration of the image according to the color temperature value calculated by the color temperature matching module.

As a preference of the present invention, the spectroscopic element is a filter type spectroscopic element, a dispersion type spectroscopic element, an interference type spectroscopic element or a diffraction type spectroscopic element.

As a preference of the present invention, the image sensor is a silicon-based image sensor, specifically a CMOS image sensor or a CCD image sensor, which is used to convert the split light signal into an electrical signal and output it as a digital signal or code. The exposure time is On the order of milliseconds to seconds.

As a further preference of the present invention, the light splitting element is a filter film, and the filter film is a single-layer structure including M cycles, and each cycle includes T ₁ , T ₂ ... T _n units , Are made of n kinds of materials with known and different light transmittance through coating and etching one by one, and each unit covers a pixel of the image sensor, so that the filter film corresponding to each pixel has the same Or different spectral transmittance; the calculation method of spectral data is shown in formula (1),

S _i =∫I(λ)T _i (λ)η(λ)dλ, (1)

Among them, S is the intensity value of the optical signal output by the image sensor, I is the incident spectrum, which is the signal to be solved, T is the spectral transmittance of the filter film, η is the quantum efficiency of the image sensor, and λ is the incident wavelength.

As a further preference of the present invention, the method for preparing the filter film is as follows: select n kinds of polyimide filter film materials with different spectral transmittances, and first coat the first filter film on the image sensor Material, and then apply an etching layer, according to the corresponding relationship with the image sensor pixels, save the necessary places, and etch away the unnecessary places; then apply the second filter film material, and then apply another The layer etching layer, according to the corresponding relationship with the image sensor pixels, keep the necessary places and etch away the unneeded places; cycle in turn until all n kinds of filter film materials are coated on the image sensor image, as mentioned above After n kinds of filter film materials are coated and etched one by one, a complete filter film with M periodicities is finally formed, and each period includes T ₁ , T ₂ ......T _n unit.

The advantages and positive effects of the present invention are:

1. The present invention calculates the color temperature value by using the ambient light multispectral data obtained by the spectrum chip, which has more information than the traditional color temperature sensor, strong environmental applicability, simple algorithm, small memory required, and accurate color temperature. , Fast and real-time detection, with the advantages of fast test speed and high accuracy.

2. The color temperature test method provided by the present invention can obtain the color temperature value of each area of the target image, and obtain a more accurate color temperature value of the ambient light after averaging, and then use the existing white balance algorithm to process the color image, and the obtained image is more reliable. Authenticity.

3. The spectrum chip used in the color temperature test system of the present invention has the advantages of wide spectrum range, small size, high spectral resolution, light weight, simple structure, convenient operation, fast detection speed, etc., combined with a simple data processing module to form a low cost The color temperature sensor is suitable for all electronic devices with display and camera functions, such as smart phones, tablet computers, notebook computers, driving recorders, etc., with a wide range of applications.

Description of the drawings

Figure 1 is a schematic block diagram of the color temperature test system of the present invention;

Figure 2 is a schematic diagram of the filter film of the present invention;

Figure 3 is a flow chart of the color temperature value test of the present invention;

Fig. 4 is the standard multi-spectral data of the white block of the standard color plate when the color temperature standard lamp emits 3 color temperature values of 2800K, 5000K and 6500K according to the present invention.

Reference signs: control module 1, spectrum chip 2, data processing module 3, standard data module 4, color temperature matching module 5.

Detailed ways

In the following description, for illustrative purposes, in order to provide a comprehensive understanding of one or more embodiments, many specific details are set forth. However, it is obvious that these embodiments can also be implemented without these specific details. In other examples, for the convenience of describing one or more embodiments, well-known structures and devices are shown in the form of block diagrams.

Embodiment 1 An environmental color temperature test system based on multi-spectral image detection technology

Referring to Figure 1, an environmental color temperature testing system based on multispectral image detection technology provided by the present invention includes: a control module 1, a spectrum chip 2, a data processing module 3, a standard data module 4, and a color temperature matching module 5; The control module 1 is connected with the spectrum chip 2 and the data processing module 3, and is used for the spectrum data collection of the spectrum chip 2 and the data processing module 3 to start to perform command control, and at the same time perform the white balance of the image according to the color temperature value calculated by the color temperature matching module 5 calibration;

The spectrum chip 2 is connected to the data processing module 3. The image and the multispectral data matrix collected by the spectrum chip 2 are sent to the data processing module, and the data processing module performs temporal and spatial noise reduction processing; the spectrum chip includes a spectroscopic element and a CMOS image sensor The light splitting element is used to divide the incident light into n wavelength bands to realize the function of spectral light splitting; the light splitting element transmits the radiation information of the same shooting target in different spectral wavelength bands to the light sensing surface of the underlying image sensor, so that all The captured image includes n photos of different spectral bands, so that for a certain pixel area in the image, corresponding to n spectral data, that is to achieve the multi-spectral data collection of a certain area of the target; for obtaining the multi-spectral data of a certain area of the target image The data is obtained by averaging the corresponding wavebands of all pixels in the area (the spectral data corresponding to the pixels with the same wavebands are averaged), that is, the number of multispectral data corresponding to a certain area of the target image is n;

The data processing module 3 is connected to the standard data module 4 and the color temperature matching module 5;

The standard data module 4 is connected to the color temperature matching module 5. The standard data module 4 is used to pre-store the multi-spectral data corresponding to the standard color temperature. The multi-spectral data includes m categories, corresponding to m standard colors, that is, one category Represents a standard color, each category contains the standard multi-spectral number sequence corresponding to the I set of standard color temperature values, the number of standard multi-spectral data in each column is the total number of bands of the spectrum chip n, and the standard multi-spectral data of each category is x _st (n, I _CCT );

The color temperature matching module 5 is connected to the control module 1, and is used to calculate the color temperature value of the ambient light in the captured image and transmit it to the control module 1. The control module 1 performs the main white balance calibration of the image according to the color temperature value. ; The color temperature calculation method is the standard deviation method, that is, the color temperature value standard deviation threshold S _t is preset, and the ambient light multispectral data x _et (n) of each area in the captured image is sequentially compared with the standard multispectral data in the category according to the category x _st (n, I _CCT ) calculate the standard deviation according to the following formula. If the calculated standard deviation value is less than St, it means that the color temperature of the ambient light in the area is equal to the color temperature value corresponding to the standard multispectral data; if it is greater than St , And then calculate the next column of data;

The standard deviation calculation formula is:

…

.

Further, the light splitting element is a filter type light splitting element, a dispersion type light splitting element, an interference type light splitting element or a diffraction type light splitting element.

In this embodiment, the light splitting element adopts a self-made filter film. The filter film has a single-layer structure and includes M cycles, and each cycle includes T ₁ , T ₂ ... T ₉ units, all of which are It is made up of 9 materials with known and different light transmittances through coating and etching one by one (see Figure 2). Each unit covers a pixel of the image sensor, so that each pixel corresponds to a filter film Have the same or different spectral transmittance; the spectral data calculation method of the spectrum chip is shown in formula (1),

S _i =∫I(λ)T _i (λ)η(λ)dλ, (1)

Among them, S is the intensity value of the optical signal output by the image sensor, I is the incident spectrum, which is the signal to be solved, T is the spectral transmittance of the filter film, η is the quantum efficiency of the image sensor, and λ is the incident wavelength;

The method for preparing the filter film is as follows: select 9 kinds of polyimide filter film materials with different spectral transmittances, first coat the first filter film material on the image sensor, and then coat a layer of engraving Etching layer (epoxy resin material), according to the corresponding relationship with the image sensor pixels, keep the needed places and etch away the unneeded places; then apply the second filter film material, and then apply a layer of etching The etching layer, according to the corresponding relationship with the image sensor pixels, keep the necessary places and etch away the unneeded places; cycle in turn until all n kinds of filter film materials are coated on the image sensor image, the above 9 kinds After the filter film material is coated and etched one by one, a complete filter film with M periodicities is finally formed, and each period includes T ₁ , T ₂ ... T ₉ units.

The image sensor is a silicon-based image sensor, specifically a CMOS image sensor or a CCD image sensor, which is used to convert the split light signal into an electrical signal and output it as a digital signal or code, and its exposure time is on the order of milliseconds to seconds.

In order to enable those skilled in the art to understand the multi-spectral data stored in the standard data module of this application, specific examples are given below.

For example: if the filter film of the present invention is made of 9 kinds of polyimide filter film materials with different spectral transmittances, the number of multi-spectral data corresponding to a certain area of the target image finally obtained is 9; if the multi-spectral The data includes 24 categories (corresponding to 24 standard colors), and each category contains 3 sets of standard multi-spectral series corresponding to standard color temperature values (2800K, 5000K and 6500K). The number of standard multi-spectral data in each column is the spectrum chip The total number of bands is 9, and the standard multispectral data of each category is x _st (9,3). Refer to Table 1 for the specific data distribution.

Table 1 Multi-spectral data distribution table

The color temperature test system provided in this application can be directly applied to existing mobile phones or cameras. When the mobile phone camera is used to take pictures, the temperature measurement system automatically calculates the color temperature value of the ambient light, and then outputs it to the control module 1. The color temperature value can be calibrated for the main white balance of the image using the existing method.

Embodiment 2 An environmental color temperature test system based on multispectral image detection technology specifically includes the following steps:

Step S1. Place the standard color plate under the color temperature standard lamp, adjust the distance between the color temperature standard lamp and the standard color plate to 50cm, and the distance between the standard color plate and the spectrum chip 2 to 10cm; the system starts self-checking. After the self-checking is normal, the color temperature The standard lamp and the spectrum chip 2 are in a preheating standby state; wherein, the standard color temperature lamp is an LED standard color temperature lamp, the color temperature range is 2500～8500K (±200K), the dimming range is 0%～100%, and the maximum output power is 10W, adjust the color temperature standard lamp so that the output color temperature values are 2800K, 5000K and 6500K respectively; the spectrum chip 2 uses hyperspectral pixel-level coating chip, model specification: QS-A-8-400-001, 400nm～850nm The waveband range is divided into 9 wavebands, the size is 3mm×3mm, the thickness is 100μm, and the data acquisition time is 1ms;

Step S2. Turn on the color temperature standard lamp, the light wave emitted by it irradiates the standard color panel, and the color temperature standard lamp presents a uniform spot on the standard color panel; at the same time, the spectrum chip 2 is activated to obtain a uniform brightness image of the white square of the standard color panel. Test once every 1ms, continuously collect 50 times, and obtain the multi-spectral data of each pixel of the white square of the standard color plate after 2 spectroscopy chips, that is, 50 sets of multi-spectral data matrices are obtained;

Step S3: Perform temporal and spatial noise reduction processing on 50 sets of multispectral data matrices. The temporal noise reduction processing method is to average the corresponding positions of the 50 sets of multispectral data matrices to finally obtain an averaged multispectral data matrix; spatial noise reduction The processing is to average the multi-spectral data of the same unit T _n in the white square area, and finally obtain 9 multi-spectral data, which are used as the standard multi-spectral data of the white square at the color temperature value and stored in the standard data module 4; this step The obtained multi-spectral data corresponding to the color temperature after noise reduction processing are: the multi-spectral data corresponding to the color temperature value of 2800K are 8.042842, 9.380254, 14.2396, 40.57888, 64.85493, 95.96654, 70.46638, 40.36318, 39.55659; the color temperature value of 5000K corresponds to many The spectral data is 8.819268, 8.646168, 26.13805, 45.86027, 52.01659, 92.57915, 69.0926, 56.85547, 45.25743; the multi-spectral data corresponding to the color temperature value of 6500K is 9.199462, 8.471258, 30.61717, 47.96308, 47.45459, 91.54122, 68.79153, 63.18177, 47.18177 channels As the abscissa, the multi-spectral intensity corresponding to the 9 channels is plotted as the ordinate, as shown in Figure 4, and the data is stored in the standard data module 4;

Step S4: Repeat steps S2 and S3 to obtain standard multi-spectral data under all standard colors and store them in the standard data module 4;

Step S5: Use a mobile phone or camera with a color temperature test system to take a scene picture, the spectrum chip 2 obtains the scene picture and multispectral data, the data processing module takes the multispectral data of a white area in the scene picture, and corresponds to the _{same unit T n} The average multi-spectral data x _et (9) is: 7.243209, 9.035698, 14.36263, 38.57236, 59.42341, 90.43591, 65.83675, 39.70332, 37.84984. The color temperature matching module 5 calculates the multi-spectral data and the standard multi-spectral data. Data x _st (9,3) is calculated for standard deviation;

The color temperature value calculation method is the standard variance method, the color temperature value standard deviation threshold S _t =5 is preset, and the ambient light multispectral data x _et (9) of the pixel points are sequentially compared with the standard multispectral data x _{st in the standard data module.} (9,3) To calculate the standard deviation, the calculation formula is:

The variance calculation results of this embodiment are respectively S ₆₅₀₀ = 11.32462, S ₅₀₀₀ = 8.256537, S ₂₈₀₀ = 3.156483, S ₂₈₀₀ <S _t , then the color temperature value of the pixel is 2800K, and the color temperature matching module 5 transmits the test result to the control Module 1, the control module 1 only needs to perform the main white balance calibration of the image according to the color temperature value; because this embodiment only collects spectral data once, no time noise reduction processing is needed, and only spatial noise reduction processing is required.

According to the test method of step 5, the present invention detects the color temperature under sunlight (respectively 40 minutes after sunrise on a sunny day (3000K), noon (5400K), and noon (6500K) on a cloudy day) under sunlight to determine the detection of the detection method The result is accurate.

The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited to this. Any person skilled in the art can easily conceive of changes or substitutions within the technical scope disclosed by the present invention, which shall cover Within the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims

An environmental color temperature testing method based on multispectral image detection technology, which is characterized in that it specifically includes the following steps:

Step S1. Place a standard color plate containing m standard colors under a color temperature standard lamp, and adjust the distance between the color temperature standard lamp and the standard color plate, and the distance between the standard color plate and the spectrum chip; wherein, the color temperature output by the color temperature standard lamp The value is 1; the spectroscopic chip includes a spectroscopic element and an image sensor, and the spectroscopic element is used to divide incident light into n wavebands, and transmit radiation information on different spectral wavebands to the light sensing surface of the underlying image sensor, Make the captured image include n photos of different spectral bands, so that for a certain pixel area in the image, corresponding to n spectral data, that is, the multi-spectral data collection of a certain area of the target is realized;

Step S2. Start the color temperature standard lamp and the spectrum chip. The light waves emitted by the color temperature standard lamp irradiate the standard color plate. The color temperature standard lamp presents a uniform spot on the standard color plate. The light reflected by the standard color plate enters the spectrum chip, and the spectrum chip Obtain images and multispectral data with uniform brightness for the 1.2.3....m color squares in the standard color palette, and continuously collect N times, that is, N sets of multispectral data matrices are obtained for each color square;

Step S3: Perform temporal and spatial noise reduction processing on the N sets of multispectral data matrices of each color square. The temporal noise reduction processing method is to average the corresponding positions of the N sets of multispectral data matrices, and finally obtain an averaged multispectral data Matrix; spatial noise reduction processing is to average the multispectral data of the same band in the same color block area, and finally obtain n multispectral data, as the standard multispectral data of the m-th color block under the color temperature value, and store it in In the standard data module; in this way, determine the standard multispectral data x st (n, I CCT ) of the m-th color square under the 1.2.3....I color temperature value in turn;

Step S4. Use a mobile phone or camera with a color temperature test system to obtain multi-spectral data x et (n) in a variety of specific color regions in the scene picture, average the multi-spectral data of the same wavelength band, and then use the color temperature matching module to reduce the amount of ambient light Spectral data and standard multi-spectral data x st (n, I CCT ) are calculated for standard deviation. If the calculated standard deviation value is less than the standard deviation threshold St, it means that the color temperature of the ambient light in the area is equal to the standard multi-spectral data Corresponding color temperature value; if it is greater than St, then calculate the next column of data;

The standard deviation calculation formula is:
An environmental color temperature test method based on multispectral image detection technology according to claim 1, wherein in step S4, when selecting a specific color area in the scene picture, white, blue, green, and red are the main ones.
The test system used in the environmental color temperature test method based on the multispectral image detection technology of claim 1, wherein the system includes a control module, a spectrum chip, a data processing module, a standard data module, and a color temperature matching module; wherein , The control module is connected with the spectrum chip and the data processing module, and is used for the spectrum data collection of the spectrum chip and the start of the data processing module for command control;

The spectrum chip is connected to the data processing module, the images and multi-spectral data collected by the spectrum chip are sent to the data processing module, and the data processing module performs noise reduction processing; the spectrum chip includes a spectroscopic element and an image sensor, and the spectroscopic element is used to The incident light is divided into n bands to realize the function of spectral light splitting; the light splitting element transmits the radiation information of the same shooting target in different spectral bands to the light sensing surface of the underlying image sensor, so that the captured image includes n different spectra A photo of the band, so that for a certain pixel area in the image, it corresponds to n spectral data, that is, the multi-spectral data collection of a certain area of the target is realized; for obtaining the multi-spectral data of a certain area of the target image, all the pixels in the area are used The corresponding band is obtained by averaging, that is, the number of multispectral data corresponding to a certain area of the target image is n;

The data processing module is connected with the standard data module and the color temperature matching module;

The standard data module is connected to the color temperature matching module; the standard data module is used to pre-store the multi-spectral data corresponding to the standard color temperature value. The multi-spectral data includes m categories, corresponding to m standard colors, that is, one category represents one type Standard color, each category contains the standard multi-spectral number sequence corresponding to the I set of standard color temperature values, the number of standard multi-spectral data in each column is the total number of bands of the spectrum chip n, and the standard multi-spectral data of each category is x st (n , I CCT );

The color temperature matching module is connected to the control module, and is used to calculate the color temperature value of the ambient light in the captured image and transmit it to the control module. The calculation method is the standard variance method, that is, the color temperature value standard deviation threshold S t is preset, Calculate the standard deviation of the ambient light multispectral data x et (n) of each area in the captured image according to the category and the standard multispectral data x st (n, I CCT ) in the category according to the following formula, if the calculated standard If the variance value is less than St, it means that the color temperature value of the ambient light in the area is equal to the color temperature value corresponding to the standard multispectral data; if it is greater than St, proceed to the next column of data calculation;

The standard deviation calculation formula is:
The environmental color temperature test system based on the multispectral image detection technology of claim 3, wherein the control module is further configured to perform image white balance calibration according to the color temperature value calculated by the color temperature matching module.
The environmental color temperature test system based on multispectral image detection technology according to claim 3, wherein the light splitting element is a filter type light splitting element, a dispersive light splitting element, an interference type light splitting element or a diffraction type light splitting element.
The environmental color temperature test system based on multi-spectral image detection technology according to claim 3, wherein the image sensor is a silicon-based image sensor, specifically a CMOS image sensor or a CCD image sensor, which is used to divide the light The light signal is converted into an electrical signal and output as a digital signal or code, and its exposure time is on the order of milliseconds to seconds.
The environmental color temperature test system based on multispectral image detection technology of claim 3, wherein the light splitting element is a filter film, and the filter film has a single-layer structure including M cycles, each The cycle includes T 1 , T 2 ......T n units, all of which are composed of n kinds of materials with known and different light transmittances that are spliced after being coated and etched one by one, and each unit covers the image sensor In this way, the filter film corresponding to each pixel has the same or different spectral transmittance; the calculation method of spectral data is shown in formula (1),

S i =∫I(λ)T i (λ)η(λ)dλ, (1)

Among them, S is the intensity value of the optical signal output by the image sensor, I is the incident spectrum, which is the signal to be solved, T is the spectral transmittance of the filter film, η is the quantum efficiency of the image sensor, and λ is the incident wavelength.
The environmental color temperature test system based on multispectral image detection technology according to claim 7, wherein the preparation method of the filter film is: selecting n kinds of polyimide filters with different spectral transmittances. Thin film material, first coat the first filter thin film material on the image sensor, and then coat an etching layer, according to the corresponding relationship with the image sensor pixels, the necessary places are reserved, and the unnecessary places are etched away ; After coating the second filter film material, and then coating an etching layer, according to the corresponding relationship with the image sensor pixels, the necessary places are reserved, and the unneeded places are etched away; cycle in turn until the All n kinds of filter film materials are coated on the image sensor image. After the above n kinds of filter film materials are coated and etched one by one, a complete layer of filter film with M periodicity is finally formed, each The cycle includes T 1 , T 2 ... T n units.