CN117132829A - Multispectral camouflage identification imaging system and camouflage identification method - Google Patents

Abstract

The invention relates to a camouflage identification imaging system and a camouflage identification method, in particular to a multispectral camouflage identification imaging system and a camouflage identification method, which solve the technical problems that the existing camouflage identification imaging system is small in identification substance range and limited in application scene, or the camouflage identification method cannot image in real time and identify in real time, or the spectrum database is low in construction efficiency. The multispectral camouflage identification system comprises an image acquisition unit, a data processing unit and a display storage unit, wherein the image acquisition unit adopts a snapshot type spectrum imaging chip and can be used for real-time exposure and real-time data processing; the data processing unit can acquire continuous spectrum information by means of a spectrum fitting degree recognition technology, selects a characteristic database according to characteristic spectrums of different types of recognition targets, changes the recognition targets, and greatly increases the recognizable spectrum range. The multispectral camouflage identification method comprises spectrum correction, so that interference of environmental light factors can be removed, and the multispectral camouflage identification method works under different sunlight temperatures.

Description

Multispectral camouflage identification imaging system and camouflage identification method

Technical Field

The invention relates to a camouflage identification imaging system and a camouflage identification method, in particular to a multispectral camouflage identification imaging system and a camouflage identification method.

Background

The conventional camouflage identification imaging system mainly depends on the specific spectral reflectance, color and other characteristics of camouflage identification targets, the purpose of the camouflage identification imaging system is single, one imaging system can only identify camouflage of specific substances based on fixed characteristics, a plurality of imaging systems with different characteristics are needed to be built in for identifying various targets, and the identification cost is high.

The existing mature camouflage identification method based on multispectral imaging needs multiple exposure to construct a spectrum data cube, so that continuous shooting is usually needed when an unmanned aerial vehicle is used for carrying a multispectral imaging system for imaging identification, and a scene is kept relatively still in the shooting process, so that the camouflage identification method is difficult to adapt to the complex and changeable anti-fake scene requirements, and the application scene is severely limited. In addition, the application scenes of the camouflage identification method are mostly outdoor sunlight environments, sunlight color temperatures at different times in different weather are different, and reflection spectrums of objects are different, so that a spectrum database is required to be built under different color temperatures, the sample size is large, and the building efficiency is low.

Therefore, there is a need for a multispectral imaging system capable of real-time imaging and real-time recognition and a camouflage recognition method adapted to the multispectral imaging system to solve the above problems.

Disclosure of Invention

The invention aims to solve the technical problems that the existing camouflage identification imaging system is small in material identification range and limited in application scene, or the camouflage identification method cannot image in real time, identify in real time or is low in spectrum database construction efficiency, and provides a multispectral camouflage identification imaging system and a camouflage identification method.

In order to achieve the above purpose, the invention adopts the following technical scheme:

a multispectral camouflage identification imaging system is characterized in that: the device comprises a calibration white board, an image acquisition unit and a data processing unit which are electrically connected, and a display storage unit which is electrically connected with the data processing unit;

the calibration whiteboard is positioned in a shooting area of the camouflage identification target during calibration;

the image acquisition unit comprises a band-pass filter, an imaging lens group and a snapshot type spectrum imaging chip which are sequentially arranged along a light path of a camouflage identification target, and a sensor circuit group connected with the snapshot type spectrum imaging chip;

the data processing unit comprises a spectrum reconstruction module, a true color reduction module, a color recognition module, a spectrum classification recognition module, a recognition result output module and a spectrum correction module which are connected in sequence;

the input end of the spectrum correction module is connected with the output end of the spectrum reconstruction module, and the output end of the spectrum correction module is connected with the input end of the spectrum classification and identification module;

the snapshot type spectrum imaging chip is electrically connected with the spectrum reconstruction module through the sensor circuit group, and the identification result output module is electrically connected with the display storage unit;

the CIE-RGB1931 chromaticity curve is configured in the true color reduction module;

the color recognition module is internally provided with an RGB color database, and the RGB color database comprises a plurality of standard RGB channel data of camouflage recognition targets;

the spectral reflectance database comprises a plurality of camouflage object reflectance spectrums similar to the target camouflage object and camouflage material reflectance spectrums similar to the target camouflage material.

Further, the data processing unit is an external computer or an internal FPGA.

Further, the display storage unit comprises a digital image display device and a data storage device;

the digital image display device and the data storage device are respectively and electrically connected with the identification result output module.

The multispectral camouflage identification method is based on the multispectral camouflage identification imaging system and is characterized by comprising the following steps:

step 1, acquiring a spectral reflectivity database comprising a standard spectrum, and configuring the spectral reflectivity database into a spectrum classification and identification module; the standard spectrum comprises a plurality of camouflage object reflectivity spectrums of approximate target camouflage objects and a plurality of camouflage material reflectivity spectrums of approximate target camouflage materials;

step 2, an RGB color database comprising standard RGB channel data is obtained and configured into a color recognition module; the RGB color database comprises a plurality of standard RGB channel data of camouflage identification targets;

step 3, placing the calibration white board in a shooting area of a camouflage identification target, acquiring a spectrum image of the calibration white board by an image acquisition unit, and acquiring an environment spectrum of the calibration white board by a spectrum reconstruction module;

step 4, removing the calibration white board, and acquiring a gray image of a camouflage identification target shooting area by an image acquisition unit, wherein a spectrum reconstruction module converts the gray image into a spectrum data cube;

step 5, the true color reduction module converts the spectrum data cube into an RGB image;

step 6, the color recognition module selects corresponding standard RGB channel data from the RGB color database configured in the step 2 according to the camouflage recognition target, and performs preliminary color screening on the RGB image obtained in the step 5 according to the standard RGB channel data to obtain a camouflage recognition pixel area with the camouflage recognition target color;

step 7, selecting a spectrum data cube corresponding to the camouflage identification pixel area from the spectrum data cubes obtained in the step 4, and correcting the spectrum data cube corresponding to the camouflage identification pixel area by the spectrum correction module by adopting the environment spectrum obtained in the step 3 to obtain a correction spectrum of each pixel in the camouflage identification pixel area;

step 8, selecting a corresponding camouflage object reflectivity spectrum and camouflage material reflectivity spectrum in the spectrum reflectivity database configured in the step 1 according to the camouflage identification target by the spectrum classification identification module, respectively comparing each pixel correction spectrum with the camouflage object reflectivity spectrum and the camouflage material reflectivity spectrum in the camouflage identification pixel area, selecting the pixel if the pixel contains the spectrum characteristic of the camouflage identification target, and otherwise, continuously comparing the next pixel; the region formed by the pixels containing the spectrum characteristics of the camouflage identification target in the camouflage identification pixel region is recorded as a camouflage pixel region;

and 9, marking the camouflage pixel area in the RGB image obtained in the step 5, obtaining a color image with camouflage marks, transmitting the color image to a display storage unit through a recognition result output module, and displaying and storing the color image with the camouflage marks by the display storage unit to finish camouflage recognition.

Further, the step 8 specifically includes:

step 8.1, selecting a corresponding camouflage object reflectivity spectrum and a camouflage material reflectivity spectrum in the spectral reflectivity database configured in the step 1 by the spectral classification recognition module according to the camouflage recognition target;

step 8.2, selecting a wave band at a characteristic spectrum part of the correction spectrum of each pixel in the camouflage identification pixel area, and normalizing the spectrum of the corresponding pixel by taking the wave band at the characteristic spectrum part as a standard to obtain a normalized spectrum of each pixel in the camouflage identification pixel area;

selecting a wave band at a characteristic spectrum part of the camouflage object reflectivity spectrum and a wave band at a characteristic spectrum part of the camouflage material reflectivity spectrum, and respectively normalizing the camouflage object reflectivity spectrum and the camouflage material reflectivity spectrum by taking the wave band at the characteristic spectrum part as a standard to obtain a normalized camouflage object reflectivity spectrum and a normalized camouflage material reflectivity spectrum;

step 8.3, respectively calculating Euclidean distance between the normalized spectrum of each pixel and the normalized camouflage object reflectivity spectrum in the characteristic spectrum section and Euclidean distance between the normalized spectrum of each pixel and the camouflage material reflectivity spectrum to obtain two groups of Euclidean distances of each pixel;

step 8.4, setting a spectrum recognition threshold, comparing the two Euclidean distances of each pixel obtained in the step 8.3 with the spectrum recognition threshold respectively, and classifying the pixel as other pixels if the two Euclidean distances of the pixel are both larger than the spectrum recognition threshold;

if at least one group of Euclidean distances of the pixels is smaller than or equal to the spectrum recognition threshold value, and the Euclidean distance between the normalized spectrum of the pixels and the reflectivity spectrum of the camouflage object is smaller than the Euclidean distance between the normalized spectrum and the reflectivity spectrum of the camouflage material, classifying the pixels as the camouflage object;

if at least one group of Euclidean distances of the pixels is smaller than or equal to a spectrum recognition threshold value, and the Euclidean distance between the normalized spectrum of the pixels and the reflectivity spectrum of the camouflage object is larger than the Euclidean distance between the normalized spectrum and the reflectivity spectrum of the camouflage material, classifying the pixels as camouflage material;

step 8.5, marking pixels classified as camouflage material, wherein the area formed by the pixels is camouflage pixel area.

Further, the step 6 specifically includes:

step 6.1, the color recognition module selects corresponding standard RGB channel data from the RGB color database configured in the step 2 according to a camouflage recognition target, normalizes the standard RGB channel data by taking the gray value of the channel G in the standard RGB channel data as a standard to obtain normalized standard RGB channel data, wherein the normalized standard RGB channel data comprises standard R channel gray data, standard G channel gray data and standard B channel gray data;

step 6.2, extracting RGB channel data of each pixel in the RGB image obtained in the step 5, and normalizing the RGB channel data by taking the gray value of a channel G in the RGB channel data of each pixel as a standard according to the normalization method of the step 6.1 to obtain normalized RGB channel data of each pixel, wherein the normalized RGB channel data comprises R channel gray data, G channel gray data and B channel gray data;

step 6.3, respectively calculating Euclidean distance between normalized RGB channel data and normalized standard RGB channel data of each pixel according to the following formula:

wherein L represents the euclidean distance between the normalized RGB channel data and the normalized standard RGB channel data, and a smaller euclidean distance represents a smaller difference between the two sets of data;

r ₀ 、g ₀ 、b ₀ respectively representing standard R channel gray scale data, standard G channel gray scale data and standard B channel gray scale data;

r, G and B respectively represent R channel gray data, G channel gray data and B channel gray data of pixels in the RGB image;

and 6.4, setting a color recognition threshold, comparing the Euclidean distance between the normalized RGB channel data of each pixel and the normalized standard RGB channel data with the color recognition threshold, and recording the pixel corresponding to the normalized RGB channel data with the Euclidean distance smaller than the color recognition threshold as a target pixel, wherein the region formed by the target pixels is a camouflage recognition pixel region with camouflage recognition target colors.

Further, in step 6.1, the specific normalization method is as follows:

A. dividing the gray value of the channel G in the standard RGB channel data by 1 to obtain a coefficient a;

B. the gray value R of the channel R in the standard RGB channel data ₁ Gray value G of channel G ₁ Channel B gray value B ₁ Dividing the channel gray value by the coefficient a to obtain normalized channel gray value r,g、b；

C. R, g, b are integrated into normalized standard RGB channel data.

Further, the step 5 specifically includes:

step 5.1, multiplying the spectrum dimension matrix of each pixel in the spectrum data cube obtained in the step 4 by a CIE-RGB1931 chromaticity coordinate curve to obtain the RGB channel gray value of each pixel:

RGB(x,y)＝CIE*S(x,y)

wherein S (x, y) represents a spectral dimension matrix of the pixel at the spatial coordinates (x, y), RGB (x, y) represents RGB channel gray values of the pixel at the spatial coordinates (x, y), and CIE represents a CIE-RGB1931 chromaticity coordinate curve;

and 5.2, arranging the RGB channel gray values of each pixel according to the spectral data cube space coordinate sequence to obtain an RGB image.

Further, the step 1 specifically includes:

step 1.1: obtaining the linear difference of characteristic spectrum segments of the approximate target camouflage objects and the approximate target camouflage materials, and preparing a certain amount of approximate target camouflage objects and approximate target camouflage materials as samples constructed by a spectral reflectance database;

step 1.2, placing the calibration white board under sunlight, acquiring a spectrum image of the calibration white board by an image acquisition unit, and acquiring a sunlight environment spectrum by a spectrum reconstruction module;

step 1.3, an image acquisition unit respectively acquires spectrum images of an approximate target camouflage object and an approximate target camouflage material, and a spectrum reconstruction module is adopted to acquire a camouflage object spectrum and a camouflage material spectrum;

step 1.4, a spectrum correction module corrects the camouflage object spectrum and the camouflage material spectrum obtained in the step 1.3 according to the sunlight environment spectrum obtained in the step 1.2 to obtain a camouflage object reflectivity spectrum and a camouflage material reflectivity spectrum, and a characteristic spectrum section of the camouflage object reflectivity spectrum and the camouflage material reflectivity spectrum is selected and recorded;

step 1.5, replacing the approximate target camouflage object and the approximate target camouflage material, and repeating the steps 1.3-1.4 to obtain a plurality of groups of camouflage object reflectivity spectrums and characteristic spectrum segments of camouflage material reflectivity spectrums of all the approximate target camouflage objects and the approximate target camouflage materials prepared in the step 1.1;

step 1.6, judging whether the characteristic spectrum section of the plurality of groups of camouflage object reflectivity spectrums and camouflage material reflectivity spectrums obtained in the step 1.5 has the rule of linear difference of the characteristic spectrum section obtained in the step 1.1 or not, thereby verifying the effectiveness of the characteristic spectrum section and obtaining effective camouflage object reflectivity spectrums and characteristic spectrum sections of camouflage material reflectivity spectrums;

and 1.7, storing the effective camouflage object reflectivity spectrum and camouflage material reflectivity spectrum corresponding to the characteristic spectrum of the camouflage object reflectivity spectrum and the camouflage material reflectivity spectrum as standard spectrums in a spectral reflectivity database, and configuring the spectral reflectivity database into a spectral classification and identification module.

Further, the step 2 specifically includes:

step 2.1, placing a camouflage identification target in sunlight, and recording RGB three-channel gray information of the camouflage identification target by using a colorless RGB color camera;

step 2.2, replacing the camouflage identification targets, repeating the step 2.1 to obtain RGB three-channel gray scale information of various camouflage identification targets, and forming standard RGB channel data by the RGB three-channel gray scale information of various camouflage identification targets;

and 2.3, storing the standard RGB channel data in an RGB color database, and configuring the RGB color database into a color recognition module.

Compared with the prior art, the invention has the following beneficial technical effects:

1. in the multispectral camouflage identification imaging system provided by the invention, the snapshot type spectral imaging chip is adopted, so that the imaging system has the advantages of real-time exposure and real-time data processing, and continuous identification can be carried out in a moving scene;

2. in the multispectral camouflage identification imaging system provided by the invention, the spectrum classification identification module acquires continuous spectrum information by means of a spectrum fitting degree identification technology;

3. in the multispectral camouflage identification imaging system provided by the invention, the RGB color database is configured in the color identification module, the spectral reflectivity database is configured in the spectral classification identification module, and the characteristic database can be switched according to the characteristic spectrum of different types of identification targets, so that the identification targets are changed, the functions of color image imaging, camouflage identification, region marking and the like can be realized at the same time by means of a single detector without adjusting the hardware level, and the camouflage region is displayed clearly and simultaneously has lower cost;

4. according to the multispectral camouflage identification method provided by the invention, the spectrum correction module is arranged, so that interference of environmental light factors can be removed, the multispectral camouflage identification method works under different sunlight color temperatures, the sample size for constructing the spectral reflectance database can be reduced, and the construction efficiency of the spectral reflectance database can be improved.

Drawings

FIG. 1 is a schematic diagram of an embodiment of a multispectral camouflage identification imaging system of the invention;

FIG. 2 is a schematic diagram of a multi-spectral camouflage identification imaging system according to an embodiment of the invention;

the reference numerals are explained as follows:

the system comprises a 101-image acquisition unit, a 111-band-pass filter, a 121-imaging lens group, a 131-snapshot spectrum imaging chip, a 132-sensor circuit group, a 141-calibration whiteboard and a 142-calibration whiteboard storage clamp;

201-data processing unit, 211-spectrum reconstruction module, 221-true color reduction module, 231-spectrum correction module, 241-color recognition module, 251-RGB color database, 261-spectrum classification recognition module, 271-spectrum reflectivity database, 281-recognition result output module;

301-display storage unit, 311-digital image display device, 321-data storage device.

Detailed Description

The multispectral camouflage identification imaging system and the camouflage identification method provided by the invention are further described in detail below with reference to the accompanying drawings and the detailed description. It should be understood by those skilled in the art that these embodiments are merely for explaining the technical principles of the present invention, and are not intended to limit the scope of the present invention.

A multispectral camouflage identification imaging system, as shown in fig. 1 and 2, comprises a calibration whiteboard 141, an image acquisition unit 101 and a data processing unit 201 which are electrically connected, and a display storage unit 301 which is electrically connected with the data processing unit 201.

The calibration white board 141 is positioned in a photographing region of the camouflage recognition target when calibrated. In this embodiment, a calibration whiteboard storage jig 142 for storing the calibration whiteboard 141 may be further provided, and the calibration whiteboard 141 is stored in the calibration whiteboard storage jig 142 for the rest of the time except for calibration. The calibration white board 141 can calibrate the light source spectrum at any time, so as to obtain the material reflectivity spectrum.

The image acquisition unit 101 includes a band-pass filter 111, an imaging lens group 121, and a snapshot-type spectral imaging chip 131, which are sequentially disposed along an optical path of a camouflage recognition target, and a sensor circuit group 132 connected to the snapshot-type spectral imaging chip 131. The snapshot spectrum imaging chip 131 can obtain mosaic images and has the functions of real-time exposure and real-time data processing.

The data processing unit 201 includes a spectrum reconstruction module 211, a true color restoration module 221, a color recognition module 241, a spectrum classification recognition module 261, and a recognition result output module 281, and a spectrum correction module 231, which are sequentially connected. The input end of the spectrum correction module 231 is connected with the output end of the spectrum reconstruction module 211, and the output end is connected with the input end of the spectrum classification recognition module 261. The true color restoration module 221 is configured with a CIE-RGB1931 chromaticity curve, and can convert the spectrum data into RGB three-channel gray data. The color recognition module 241 is configured with an RGB color database 251, and the RGB color database 251 includes a plurality of standard RGB channel data of camouflage recognition targets. The spectral classification recognition module 261 is provided with a spectral reflectance database 271, and the spectral reflectance database 271 includes a plurality of camouflage object reflectance spectra of the approximate target camouflage object and a plurality of camouflage material reflectance spectra of the approximate target camouflage material.

The data processing unit 201 is an external computer or an internal FPGA, can acquire continuous spectrum information by means of spectrum fitting degree recognition technology, can change recognition targets by switching characteristic databases according to different types of recognition target characteristic spectrums, and greatly increases the recognizable spectrum range. The spectral reconstruction module 211 incorporates an algorithm that is compatible with the snapshot spectral imaging chip 131, which converts the mosaic image into a spectral data cube. To eliminate the spectral variation generated by the light source variation, a spectral correction module 231 is added to the data processing unit 201, and the spectral correction needs to be performed in combination with the light source spectral calibration data. The RGB color database 251 is constructed in advance according to the recognition target color, and the spectral reflectance database 271 is constructed in advance according to the spectral characteristics of the recognition target.

The display storage unit 301 includes a digital image display device 311 and a data storage device 321, the snapshot-type spectral imaging chip 131 is electrically connected to the spectral reconstruction module 211 through the sensor circuit group 132, and the recognition result output module 281 is electrically connected to the digital image display device 311 and the data storage device 321, respectively. The digital image display device 311 is an integrated display, the data storage device 321 is an internal memory, and the display storage unit 301 may be an external computer in other embodiments.

The data processing unit 201 may be integrated on an electronic device to complete reading, processing and storage of data. The data processing unit 201 is operative to read the pre-selected characteristic spectrum and RGB database according to the recognition target, and the working environment whiteboard calibration image. The image acquisition unit 201 records gray information of the target area to be detected, the data processing unit 201 performs data processing analysis processes such as spectrum reconstruction and the like to finish camouflage identification of the target area, and the identification result is stored and displayed by the display storage unit 301. The data processing unit 201 and the display storage unit 301 can generate a color image with a recognition result by means of the true color restoration module 221, and the operation is simple.

The multispectral camouflage identification method is based on the multispectral camouflage identification imaging system and is characterized by comprising the following steps:

step 1, a spectral reflectance database 271 including a standard spectrum including a plurality of camouflage object reflectance spectra approximating a target camouflage object and a plurality of camouflage material reflectance spectra approximating a target camouflage material is acquired and configured into a spectral classification recognition module 261. The specific method comprises the following steps:

step 1.1: obtaining the linear difference of the characteristic spectrum segments of the approximate target camouflage objects and the approximate target camouflage materials, and preparing a certain amount of approximate target camouflage objects and approximate target camouflage materials as samples constructed by the spectral reflectivity database 271;

step 1.2, placing the calibration white board 141 under sunlight, and acquiring a spectrum image of the calibration white board 141 by the image acquisition unit 101 and acquiring a sunlight environment spectrum by the spectrum reconstruction module 211;

step 1.3, the image acquisition unit 101 acquires spectrum images of an approximate target camouflage object and an approximate target camouflage material respectively, and acquires a camouflage object spectrum and a camouflage material spectrum by adopting the spectrum reconstruction module 211;

step 1.4, the spectrum correction module 231 corrects the camouflage object spectrum and the camouflage material spectrum obtained in step 1.3 according to the sunlight environment spectrum obtained in step 1.2 to obtain a camouflage object reflectivity spectrum and a camouflage material reflectivity spectrum, and selects and records a characteristic spectrum segment of the camouflage object reflectivity spectrum and the camouflage material reflectivity spectrum;

step 1.5, replacing the approximate target camouflage object and the approximate target camouflage material, and repeating the steps 1.3-1.4 to obtain a plurality of groups of camouflage object reflectivity spectrums and characteristic spectrum segments of camouflage material reflectivity spectrums of all the approximate target camouflage objects and the approximate target camouflage materials prepared in the step 1.1;

step 1.6, judging whether the characteristic spectrum section of the plurality of groups of camouflage object reflectivity spectrums and camouflage material reflectivity spectrums obtained in the step 1.5 has the rule of linear difference of the characteristic spectrum section obtained in the step 1.1 or not, thereby verifying the effectiveness of the characteristic spectrum section and obtaining effective camouflage object reflectivity spectrums and characteristic spectrum sections of camouflage material reflectivity spectrums;

step 1.7, storing the effective camouflage object reflectivity spectrum and camouflage material reflectivity spectrum corresponding to the characteristic spectrum of the camouflage object reflectivity spectrum and camouflage material reflectivity spectrum as standard spectrum in a spectrum reflectivity database 271, and configuring the spectrum reflectivity database 271 into a spectrum classification and identification module 261.

Step 2, an RGB color database 251 including standard RGB channel data is acquired and configured into a color recognition module 241; wherein the RGB color database 251 includes a plurality of standard RGB channel data of camouflage recognition targets. The specific method comprises the following steps:

step 2.1, placing a camouflage identification target in sunlight, and recording RGB three-channel gray information of the camouflage identification target by using a colorless RGB color camera;

step 2.2, replacing the camouflage identification targets, repeating the step 2.1 to obtain RGB three-channel gray scale information of various camouflage identification targets, and forming standard RGB channel data by the RGB three-channel gray scale information of various camouflage identification targets;

step 2.3, storing the standard RGB channel data in the RGB color database 251, and configuring the RGB color database 251 into the color recognition module 241.

Step 3, the calibration whiteboard 141 is placed in a shooting area of a camouflage recognition target, the image acquisition unit 101 acquires a spectrum image of the calibration whiteboard 141, and the spectrum reconstruction module 211 acquires an environment spectrum of the calibration whiteboard. After single calibration, the device can continuously work for a period of time until the color temperature of the ambient light changes obviously, and then the next calibration is carried out.

Step 4, the calibration whiteboard 141 is retracted into the calibration whiteboard storage fixture 142, the image acquisition unit 101 acquires a gray image of the camouflage identification target shooting area, and the spectrum reconstruction module 211 converts the gray image into a spectrum data cube.

Step 5, the true color restoration module 221 converts the spectrum data cube into an RGB image, specifically:

step 5.1, multiplying the spectrum dimension matrix of each pixel in the spectrum data cube obtained in the step 4 by a CIE-RGB1931 chromaticity coordinate curve to obtain the RGB channel gray value of each pixel:

RGB(x,y)＝CIE*S(x,y)

wherein S (x, y) represents a spectral dimension matrix of the pixel at the spatial coordinates (x, y), RGB (x, y) represents RGB channel gray values of the pixel at the spatial coordinates (x, y), and CIE represents a CIE-RGB1931 chromaticity coordinate curve;

and 5.2, arranging the RGB channel gray values of each pixel according to the spectral data cube space coordinate sequence to obtain an RGB image.

Step 6, the color recognition module 241 selects the corresponding standard RGB channel data from the RGB color database 251 configured in step 2 according to the camouflage recognition target, and performs preliminary color screening on the RGB image obtained in step 5 according to the standard RGB channel data, so as to obtain a camouflage recognition pixel area with the camouflage recognition target color. The specific method comprises the following steps:

step 6.1, the color recognition module 241 selects the corresponding standard RGB channel data from the RGB color database 251 configured in step 2 according to the camouflage recognition target, normalizes the standard RGB channel data by taking the gray value of the channel G in the standard RGB channel data as a standard, and obtains normalized standard RGB channel data, where the normalized standard RGB channel data includes standard R channel gray data, standard G channel gray data, and standard B channel gray data.

The specific normalization method comprises the following steps:

A. dividing the gray value of the channel G in the standard RGB channel data by 1 to obtain a coefficient a;

B. the gray value R of the channel R in the standard RGB channel data ₁ Gray value G of channel G ₁ Channel B gray value B ₁ Dividing the channel gray values by the coefficient a to obtain normalized channel gray values r, g and b;

C. r, g, b are integrated into normalized standard RGB channel data.

And 6.2, extracting RGB channel data of each pixel in the RGB image obtained in the step 5, and normalizing the RGB channel data by taking the gray value of a channel G in the RGB channel data of each pixel as a standard according to the normalization method of the step 6.1 to obtain normalized RGB channel data of each pixel, wherein the normalized RGB channel data comprises R channel gray data, G channel gray data and B channel gray data.

Step 6.3, respectively calculating Euclidean distance between normalized RGB channel data and normalized standard RGB channel data of each pixel according to the following formula:

wherein L represents the euclidean distance between the normalized RGB channel data and the normalized standard RGB channel data, and a smaller euclidean distance represents a smaller difference between the two sets of data;

r ₀ 、g ₀ 、b ₀ respectively representing standard R channel gray scale data, standard G channel gray scale data and standard B channel gray scale data;

r, G and B respectively represent R channel gray scale data, G channel gray scale data and B channel gray scale data of pixels in the RGB image.

And 6.4, setting a color recognition threshold, comparing Euclidean distance L between normalized RGB channel data of each pixel and normalized standard RGB channel data with the color recognition threshold, and recording pixels corresponding to normalized RGB channel data with L smaller than the color recognition threshold as target pixels, wherein the region formed by the target pixels is a camouflage recognition pixel region with camouflage recognition target colors. The color recognition threshold is not too small, and basic color distinction is realized.

And 7, selecting a spectrum data cube corresponding to the camouflage identification pixel area from the spectrum data cubes obtained in the step 4, and correcting the spectrum data cube corresponding to the camouflage identification pixel area by using the environment spectrum obtained in the step 3 by the spectrum correction module 231 to obtain a corrected spectrum of each pixel in the camouflage identification pixel area. The specific method for correcting the optical data cube comprises the following steps:

the spectrum data cubes corresponding to the camouflage identification pixel areas are divided by the environmental spectrum one by one.

Step 8, the spectrum classification recognition module 261 selects the corresponding camouflage object reflectivity spectrum and camouflage material reflectivity spectrum from the spectrum reflectivity database 271 configured in step 1 according to the camouflage recognition target, and compares each pixel correction spectrum in the camouflage recognition pixel area with the camouflage object reflectivity spectrum and the camouflage material reflectivity spectrum respectively, if the pixel contains the spectrum characteristic of the camouflage recognition target, the pixel is selected, otherwise, the next pixel is continuously compared; and (3) marking the region comprising the pixels with the camouflage identification target spectral characteristics in the camouflage identification pixel region as a camouflage pixel region.

The specific method of the step 8 is as follows:

step 8.1, the spectrum classification recognition module 261 selects a corresponding camouflage object reflectivity spectrum and a camouflage material reflectivity spectrum from the spectrum reflectivity database 271 configured in step 1 according to the camouflage recognition target;

step 8.2, selecting a wave band at a characteristic spectrum part of the correction spectrum of each pixel in the camouflage identification pixel area, and normalizing the spectrum of the corresponding pixel by taking the wave band at the characteristic spectrum part as a standard to obtain a normalized spectrum of each pixel in the camouflage identification pixel area;

selecting a wave band at a characteristic spectrum part of the camouflage object reflectivity spectrum and a wave band at a characteristic spectrum part of the camouflage material reflectivity spectrum, and respectively normalizing the camouflage object reflectivity spectrum and the camouflage material reflectivity spectrum by taking the wave band at the characteristic spectrum part as a standard to obtain a normalized camouflage object reflectivity spectrum and a normalized camouflage material reflectivity spectrum;

step 8.3, respectively calculating Euclidean distance between the normalized spectrum of each pixel and the normalized camouflage object reflectivity spectrum in the characteristic spectrum section and Euclidean distance between the normalized spectrum of each pixel and the camouflage material reflectivity spectrum to obtain two groups of Euclidean distances of each pixel;

step 8.4, setting a spectrum recognition threshold, comparing the two Euclidean distances of each pixel obtained in the step 8.3 with the spectrum recognition threshold respectively, and classifying the pixel as other pixels if the two Euclidean distances of the pixel are both larger than the spectrum recognition threshold;

if at least one group of Euclidean distances of the pixels is smaller than or equal to the spectrum recognition threshold value, and the Euclidean distance between the normalized spectrum of the pixels and the reflectivity spectrum of the camouflage object is smaller than the Euclidean distance between the normalized spectrum and the reflectivity spectrum of the camouflage material, classifying the pixels as the camouflage object;

if at least one group of Euclidean distances of the pixels is smaller than or equal to a spectrum recognition threshold value, and the Euclidean distance between the normalized spectrum of the pixels and the reflectivity spectrum of the camouflage object is larger than the Euclidean distance between the normalized spectrum and the reflectivity spectrum of the camouflage material, classifying the pixels as camouflage material;

step 8.5, marking pixels classified as camouflage material, wherein the area formed by the pixels is camouflage pixel area.

And 9, marking a camouflage pixel area in the RGB image obtained in the step 5, obtaining a color image with camouflage marks, sending the color image to the display storage unit 301 through the identification result output module 281, and displaying and storing the color image with camouflage marks by the display storage unit 301 to finish camouflage identification.

Claims (10)

1. A multispectral camouflage identification imaging system, characterized in that: comprises a calibration whiteboard (141), an image acquisition unit (101) and a data processing unit (201) which are electrically connected, and a display storage unit (301) which is electrically connected with the data processing unit (201);

the calibration whiteboard (141) is positioned in a shooting area of the camouflage identification target during calibration;

the image acquisition unit (101) comprises a band-pass filter (111), an imaging lens group (121) and a snapshot type spectrum imaging chip (131) which are sequentially arranged along a light path of a camouflage identification target, and a sensor circuit group (132) connected with the snapshot type spectrum imaging chip (131);

the data processing unit (201) comprises a spectrum reconstruction module (211), a true color restoration module (221), a color recognition module (241), a spectrum classification recognition module (261) and a recognition result output module (281) which are sequentially connected, and a spectrum correction module (231);

the input end of the spectrum correction module (231) is connected with the output end of the spectrum reconstruction module (211), and the output end of the spectrum correction module is connected with the input end of the spectrum classification and identification module (261);

the snapshot type spectrum imaging chip (131) is electrically connected with the spectrum reconstruction module (211) through the sensor circuit group (132), and the identification result output module (281) is electrically connected with the display storage unit (301);

the true color reduction module (221) is internally provided with CIE-RGB1931 chromaticity curve;

the color recognition module (241) is internally provided with an RGB color database (251), and the RGB color database (251) comprises a plurality of standard RGB channel data of camouflage recognition targets;

the spectral classification identification module (261) is internally provided with a spectral reflectivity database (271), and the spectral reflectivity database (271) comprises a plurality of camouflage object reflectivity spectrums of approximate target camouflage objects and a plurality of camouflage material reflectivity spectrums of approximate target camouflage materials.

2. A multispectral camouflage identification imaging system as set forth in claim 1 wherein: the data processing unit (201) is an external computer or an internal FPGA.

3. A multispectral camouflage identification imaging system according to claim 1 or 2, wherein: the display storage unit (301) includes a digital image display device (311) and a data storage device (321);

the digital image display device (311) and the data storage device (321) are respectively and electrically connected with the identification result output module (281).

4. A multispectral camouflage identification method based on the multispectral camouflage identification imaging system as claimed in any one of claims 1 to 3, comprising the steps of:

step 1, acquiring a spectrum reflectivity database (271) comprising standard spectrums, and configuring the spectrum reflectivity database into a spectrum classification and identification module (261); the standard spectrum comprises a plurality of camouflage object reflectivity spectrums of approximate target camouflage objects and a plurality of camouflage material reflectivity spectrums of approximate target camouflage materials;

step 2, an RGB color database (251) comprising standard RGB channel data is acquired and configured into a color recognition module (241); the RGB color database (251) includes standard RGB channel data of a plurality of camouflage recognition targets;

step 3, placing the calibration white board (141) in a shooting area of a camouflage identification target, acquiring a spectrum image of the calibration white board (141) by an image acquisition unit (101), and acquiring an environment spectrum of the calibration white board (141) by a spectrum reconstruction module (211);

step 4, removing the calibration white board (141), and acquiring a gray image of a camouflage identification target shooting area by an image acquisition unit (101), wherein a spectrum reconstruction module (211) converts the gray image into a spectrum data cube;

step 5, a true color reduction module (221) converts the spectrum data cube into an RGB image;

step 6, a color recognition module (241) selects corresponding standard RGB channel data from an RGB color database (251) configured in the step 2 according to a camouflage recognition target, and performs preliminary color screening on the RGB image obtained in the step 5 according to the standard RGB channel data to obtain a camouflage recognition pixel area with a camouflage recognition target color;

step 7, selecting a spectrum data cube corresponding to the camouflage identification pixel area from the spectrum data cubes obtained in the step 4, and correcting the spectrum data cube corresponding to the camouflage identification pixel area by a spectrum correction module (231) by adopting the environment spectrum obtained in the step 3 to obtain a corrected spectrum of each pixel in the camouflage identification pixel area;

step 8, a spectrum classification and identification module (261) selects a corresponding camouflage object reflectivity spectrum and a corresponding camouflage material reflectivity spectrum from a spectrum reflectivity database (271) configured in the step 1 according to a camouflage identification target, and compares each pixel correction spectrum in a camouflage identification pixel area with the camouflage object reflectivity spectrum and the camouflage material reflectivity spectrum respectively, if a pixel contains the spectrum characteristics of the camouflage identification target, the pixel is selected, otherwise, the next pixel is continuously compared; the region formed by the pixels containing the spectrum characteristics of the camouflage identification target in the camouflage identification pixel region is recorded as a camouflage pixel region;

and 9, marking a camouflage pixel area in the RGB image obtained in the step 5, obtaining a color image with camouflage marks, sending the color image to a display storage unit (301) through a recognition result output module (281), and displaying and storing the color image with the camouflage marks by the display storage unit (301) to finish camouflage recognition.

5. The multispectral camouflage identification method according to claim 4, wherein the step 8 is specifically:

step 8.1, a spectrum classification and identification module (261) selects a corresponding camouflage object reflectivity spectrum and a camouflage material reflectivity spectrum from a spectrum reflectivity database (271) configured in step 1 according to a camouflage identification target;

step 8.2, selecting a wave band at a characteristic spectrum part of the correction spectrum of each pixel in the camouflage identification pixel area, and normalizing the spectrum of the corresponding pixel by taking the wave band at the characteristic spectrum part as a standard to obtain a normalized spectrum of each pixel in the camouflage identification pixel area;

selecting a wave band at a characteristic spectrum part of the camouflage object reflectivity spectrum and a wave band at a characteristic spectrum part of the camouflage material reflectivity spectrum, and respectively normalizing the camouflage object reflectivity spectrum and the camouflage material reflectivity spectrum by taking the wave band at the characteristic spectrum part as a standard to obtain a normalized camouflage object reflectivity spectrum and a normalized camouflage material reflectivity spectrum;

step 8.3, respectively calculating Euclidean distance between the normalized spectrum of each pixel and the normalized camouflage object reflectivity spectrum in the characteristic spectrum section and Euclidean distance between the normalized spectrum of each pixel and the camouflage material reflectivity spectrum to obtain two groups of Euclidean distances of each pixel;

step 8.4, setting a spectrum recognition threshold, comparing the two Euclidean distances of each pixel obtained in the step 8.3 with the spectrum recognition threshold respectively, and classifying the pixel as other pixels if the two Euclidean distances of the pixel are both larger than the spectrum recognition threshold;

if at least one group of Euclidean distances of the pixels is smaller than or equal to the spectrum recognition threshold value, and the Euclidean distance between the normalized spectrum of the pixels and the reflectivity spectrum of the camouflage object is smaller than the Euclidean distance between the normalized spectrum and the reflectivity spectrum of the camouflage material, classifying the pixels as the camouflage object;

if at least one group of Euclidean distances of the pixels is smaller than or equal to a spectrum recognition threshold value, and the Euclidean distance between the normalized spectrum of the pixels and the reflectivity spectrum of the camouflage object is larger than the Euclidean distance between the normalized spectrum and the reflectivity spectrum of the camouflage material, classifying the pixels as camouflage material;

step 8.5, marking pixels classified as camouflage material, wherein the area formed by the pixels is camouflage pixel area.

6. The multispectral camouflage identification method according to claim 4 or 5, wherein the step 6 is specifically:

step 6.1, a color recognition module (241) selects corresponding standard RGB channel data from an RGB color database (251) configured in the step 2 according to a camouflage recognition target, and normalizes the standard RGB channel data by taking the gray value of a channel G in the standard RGB channel data as a standard to obtain normalized standard RGB channel data, wherein the normalized standard RGB channel data comprises standard R channel gray data, standard G channel gray data and standard B channel gray data;

step 6.2, extracting RGB channel data of each pixel in the RGB image obtained in the step 5, and normalizing the RGB channel data by taking the gray value of a channel G in the RGB channel data of each pixel as a standard according to the normalization method of the step 6.1 to obtain normalized RGB channel data of each pixel, wherein the normalized RGB channel data comprises R channel gray data, G channel gray data and B channel gray data;

step 6.3, respectively calculating Euclidean distance L between normalized RGB channel data of each pixel and normalized standard RGB channel data according to the following formula:

wherein L represents the euclidean distance between the normalized RGB channel data and the normalized standard RGB channel data, and a smaller euclidean distance represents a smaller difference between the two sets of data;

r ₀ 、g ₀ 、b ₀ respectively representing standard R channel gray scale data, standard G channel gray scale data and standard B channel gray scale data;

r, G and B respectively represent R channel gray data, G channel gray data and B channel gray data of pixels in the RGB image;

and 6.4, setting a color recognition threshold, comparing Euclidean distance L between normalized RGB channel data of each pixel and normalized standard RGB channel data with the color recognition threshold, and recording pixels corresponding to normalized RGB channel data with L smaller than the color recognition threshold as target pixels, wherein the region formed by the target pixels is a camouflage recognition pixel region with camouflage recognition target colors.

7. The method for identifying multispectral camouflage according to claim 6, wherein in step 6.1, the specific normalization method is as follows:

A. dividing the gray value of the channel G in the standard RGB channel data by 1 to obtain a coefficient a;

B. the gray value R of the channel R in the standard RGB channel data ₁ Gray value G of channel G ₁ Channel B gray value B ₁ Dividing the channel gray values by the coefficient a to obtain normalized channel gray values r, g and b;

C. r, g, b are integrated into normalized standard RGB channel data.

8. The multispectral camouflage identification method according to claim 7, wherein the step 5 is specifically:

step 5.1, multiplying the spectrum dimension matrix of each pixel in the spectrum data cube obtained in the step 4 by a CIE-RGB1931 chromaticity coordinate curve to obtain the RGB channel gray value of each pixel:

RGB(x,y)＝CIE*S(x,y)

wherein S (x, y) represents a spectral dimension matrix of the pixel at the spatial coordinates (x, y), RGB (x, y) represents RGB channel gray values of the pixel at the spatial coordinates (x, y), and CIE represents a CIE-RGB1931 chromaticity coordinate curve;

and 5.2, arranging the RGB channel gray values of each pixel according to the spectral data cube space coordinate sequence to obtain an RGB image.

9. The multispectral camouflage identification method according to claim 8, wherein the step 1 is specifically:

step 1.1: obtaining a characteristic spectrum section linear difference between an approximate target camouflage object and an approximate target camouflage material, and preparing a certain amount of approximate target camouflage objects and approximate target camouflage materials as samples constructed by a spectral reflectivity database (271);

step 1.2, placing the calibration white board (141) under sunlight, acquiring a spectrum image of the calibration white board (141) by an image acquisition unit (101), and acquiring a sunlight environment spectrum by a spectrum reconstruction module (211);

step 1.3, an image acquisition unit (101) respectively acquires spectrum images of an approximate target camouflage object and an approximate target camouflage material, and a spectrum reconstruction module (211) is adopted to acquire a camouflage object spectrum and a camouflage material spectrum;

step 1.4, a spectrum correction module (231) corrects the camouflage object spectrum and the camouflage material spectrum obtained in the step 1.3 according to the sunlight environment spectrum obtained in the step 1.2 to obtain a camouflage object reflectivity spectrum and a camouflage material reflectivity spectrum, and a characteristic spectrum section of the camouflage object reflectivity spectrum and the camouflage material reflectivity spectrum is selected and recorded;

step 1.5, replacing the approximate target camouflage object and the approximate target camouflage material, and repeating the steps 1.3-1.4 to obtain a plurality of groups of camouflage object reflectivity spectrums and characteristic spectrum segments of camouflage material reflectivity spectrums of all the approximate target camouflage objects and the approximate target camouflage materials prepared in the step 1.1;

step 1.6, judging whether the characteristic spectrum section of the plurality of groups of camouflage object reflectivity spectrums and camouflage material reflectivity spectrums obtained in the step 1.5 has the rule of linear difference of the characteristic spectrum section obtained in the step 1.1 or not, thereby verifying the effectiveness of the characteristic spectrum section and obtaining effective camouflage object reflectivity spectrums and characteristic spectrum sections of camouflage material reflectivity spectrums;

and 1.7, storing the effective camouflage object reflectivity spectrum and camouflage material reflectivity spectrum corresponding to the characteristic spectrum of the camouflage object reflectivity spectrum and the camouflage material reflectivity spectrum as standard spectrums in a spectrum reflectivity database (271), and configuring the spectrum reflectivity database (271) into a spectrum classification and identification module (261).

10. The multispectral camouflage identification method according to claim 9, wherein the step 2 is specifically:

step 2.1, placing a camouflage identification target in sunlight, and recording RGB three-channel gray information of the camouflage identification target by using a colorless RGB color camera;

step 2.2, replacing the camouflage identification targets, repeating the step 2.1 to obtain RGB three-channel gray scale information of various camouflage identification targets, and forming standard RGB channel data by the RGB three-channel gray scale information of various camouflage identification targets;

step 2.3, storing the standard RGB channel data in an RGB color database (251), and configuring the RGB color database (251) into a color recognition module (241).