A method for generating simulated hyperspectral images based on aerial images

Abstract

The present invention relates to the technical field of simulated image generation, and specifically provides a method for generating simulated hyperspectral images based on aerial images, wherein the aerial images are obtained by drone photography, and the spectrum resolution is determined according to the flight altitude when the images are photographed, and the R, G, and B components are binarized, and the binarized components are spectrally decomposed according to the spectrum resolution to form R, G, and B simulated images of multiple spectrums; and the gray value Y of the RGB image is binarized, and the binarized gray image is spectrally decomposed according to the spectrum resolution to form a hyperspectral image with the R, G, and B simulated images of multiple spectrums. The present invention generates a hyperspectral image through aerial image simulation, which can be viewed and studied by beginners, avoiding the need for beginners to directly operate the hyperspectral imaging equipment, and significantly reducing the training cost of hyperspectral operators.

Description

Technical Field

[0001] The present invention relates to the technical field of simulated image generation, and specifically provides a method for generating simulated hyperspectral images based on aerial images. Background Art

[0002] Hyperspectral imaging technology can provide very high spectral resolution, which enables it to capture very subtle spectral changes and thus distinguish different substances and objects. Hyperspectral imaging technology integrates spectral information and image information, and can provide more comprehensive sample information. These advantages make hyperspectral imaging technology a very valuable tool in the field of modern science and technology. However, for most beginners of hyperspectral systems, the operation of hyperspectral systems is too complicated. If actual operations are performed directly, it is easy to damage the hyperspectral system. In addition, the actual installation equipment of the hyperspectral system is expensive, and the use of actual installations for training increases the cost of learning and training. Therefore, in the initial learning process, students should be prevented from directly performing actual operations to reduce the probability of equipment damage. A method that can obtain simulated hyperspectral images at low cost is needed. Summary of the invention

[0003] In order to solve the above problems, the present invention provides a method for generating simulated hyperspectral images based on aerial images, wherein RGB images are obtained by aerial photography with an unmanned aerial vehicle, and the visible light band and infrared band of the RGB image are decomposed into a hyperspectral image with a set spectral band resolution. The aerial images can be applied to the study and research of hyperspectral images, thus avoiding damage to the hyperspectral imaging equipment caused by improper operation by beginners.

[0004] The method for generating a simulated hyperspectral image based on an aerial image provided by the present invention comprises:

[0005] S1: Obtain RGB images through drone aerial photography and measure the flight height of the drone relative to the ground;

[0006] S2: Set the spectral resolution of the visible light band according to the flight altitude and infrared spectral resolution ;

[0007] S3: interpolate the R, G, and B components of the RGB image, and obtain the interpolation of the R and G components as the RG spectrum segment, the interpolation of the G and B components as the GB spectrum segment, and the interpolation of the R and B components as the RB spectrum segment;

[0008] Binarize the R component to obtain a binary image , the binarization threshold of the R component is the average brightness of the RG spectrum; binarization of the G component obtains a binary image The binarization threshold of the G component is the average brightness of the GB spectrum; the B component is binarized to obtain a binary image , the binarization threshold of the B component is the average brightness of the RB spectrum;

[0009] Binarize the image Decomposition Generation Image of spectrum ; Binarized image Decomposition Generation Image of spectrum ; Binarized image Decomposition Generation Image of spectrum ;

[0010] Binarize the gray value Y of the RGB image to obtain the binary image , the binarization threshold of the gray value Y is the average brightness of the RGB image, and the binarized image Decomposition Generation Image of spectrum .

[0011] Preferably, when the drone obtains the RGB image through aerial photography, the distance between the drone and the ground is obtained multiple times, and the average value of the distance is used as the flight altitude.

[0012] Preferably, the distance between the drone and the ground is obtained 10 times, and the average value of the 10 distances is used as the flight altitude.

[0013] Preferably, in S2, the spectral resolution of the visible light band is switched according to the flight altitude. and the number of spectral bands in the infrared band :

[0014] When the flight altitude is 1000m, , ;

[0015] When the flight altitude is 500m, , ;

[0016] When the flight altitude is 100m, , .

[0017] Preferably, the interpolation calculation formula of the RG spectrum segment, the GB spectrum segment and the RB spectrum segment is:

[0018] .

[0019] Preferably, in S3, the binary image is converted into Generate according to the following formula uint type image of spectral segments :

[0020] ;

[0021] in, Indicates that the values ​​are taken from 1 to integer;

[0022] Use a for loop to convert the uint type image Generate image :

[0023] ;

[0024] Binarize the image through a for loop Generate according to the following formula uint type image of spectral segments :

[0025] ;

[0026] Use a for loop to convert the uint type image Generate image :

[0027] ;

[0028] Binarize the image through a for loop Generate according to the following formula uint type image of spectral segments :

[0029] ;

[0030] Use a for loop to convert the uint type image Generate image :

[0031] .

[0032] Preferably, in S3, the gray value Y of the RGB image is calculated as:

[0033] .

[0034] Preferably, in S3, the binary image Carry out corrosion and dilation to different degrees in sequence, and decompose the dilated image to generate Image of spectrum .

[0035] Compared with the prior art, the present invention can achieve the following beneficial effects:

[0036] The present invention obtains RGB images through drone aerial photography, and simulates and generates hyperspectral images by adjusting the spectral resolution of the visible light band and infrared band of the RGB image, and can provide image data of different resolutions according to different flight altitudes. The simulated hyperspectral images can be applied to the study and research of hyperspectral images, and beginners can be trained without hyperspectral imaging equipment, avoiding beginners directly operating the hyperspectral imaging equipment, resulting in equipment damage, and significantly reducing the training cost of hyperspectral operators. In addition, the method is also helpful for the in-depth analysis and processing of drone aerial images, and has important application value in the fields of remote sensing and image analysis. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 is a flow chart of a method for generating a simulated hyperspectral image based on an aerial image provided in an embodiment of the present invention;

[0038] Figure 2 is an RGB image obtained by aerial photography of a drone provided by an embodiment of the present invention;

[0039] Figure 3 is a binary image provided according to an embodiment of the present invention ;

[0040] Figure 4 is a binary image provided according to an embodiment of the present invention ;

[0041] Figure 5 is a binary image provided according to an embodiment of the present invention ;

[0042] Figure 6 The first spectrum segment provided by the embodiment of the present invention image;

[0043] Figure 7 is the 10th spectrum segment provided according to an embodiment of the present invention image;

[0044] Figure 8 is the 20th spectrum segment provided according to an embodiment of the present invention image;

[0045] Fig. 9 is the 30th spectrum segment provided according to an embodiment of the present invention image;

[0046] Fig.10 The first spectrum segment provided by the embodiment of the present invention image;

[0047] Fig.11is the 10th spectrum segment provided according to an embodiment of the present invention image;

[0048] Fig.12 is the 20th spectrum segment provided according to an embodiment of the present invention image;

[0049] Fig.13 is the 30th spectrum segment provided according to an embodiment of the present invention image;

[0050] Fig.14 The first spectrum segment provided by the embodiment of the present invention image;

[0051] Fig.15 is the 10th spectrum segment provided according to an embodiment of the present invention image;

[0052] Fig.16 is the 20th spectrum segment provided according to an embodiment of the present invention image;

[0053] Fig.17 is the 30th spectrum segment provided according to an embodiment of the present invention image;

[0054] Fig.18 The first spectrum segment provided by the embodiment of the present invention image;

[0055] Fig.19 is the 10th spectrum segment provided according to an embodiment of the present invention image;

[0056] Fig. 20 is the 20th spectrum segment provided according to an embodiment of the present invention image;

[0057] Fig.21 is the 30th spectrum segment provided according to an embodiment of the present invention image;

[0058] Fig. 22 The first spectrum segment provided by the embodiment of the present invention image;

[0059] Fig.23 is the 10th spectrum segment provided according to an embodiment of the present invention image;

[0060] Fig.24 is the 20th spectrum segment provided according to an embodiment of the present invention image;

[0061] Fig.25 is the 30th spectrum segment provided according to an embodiment of the present invention image;

[0062] Fig.26 The first spectrum segment provided by the embodiment of the present invention image;

[0063] Fig. 27 is the 10th spectrum segment provided according to an embodiment of the present invention image;

[0064] Fig.28 is the 20th spectrum segment provided according to an embodiment of the present invention image;

[0065] Fig.29 is the 30th spectrum segment provided according to an embodiment of the present invention image;

[0066] Fig.30 The first spectrum segment provided by the embodiment of the present invention image;

[0067] Fig.31 The second spectrum segment provided by the embodiment of the present invention image;

[0068] Fig.32 The third spectrum provided by the embodiment of the present invention image;

[0069] Fig.33 The fourth spectrum segment provided according to the embodiment of the present invention image. DETAILED DESCRIPTION

[0070] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the following description, the same modules are represented by the same reference numerals. In the case of the same reference numerals, their names and functions are also the same. Therefore, the detailed description thereof will not be repeated.

[0071] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and do not constitute a limitation of the present invention.

[0072] like Figure 1 As shown, the method for generating a simulated hyperspectral image based on an aerial image provided by an embodiment of the present invention comprises the following specific steps:

[0073] S1: Use the drone to carry a camera payload to take pictures of the target area and obtain the following Figure 2 As shown in the RGB image, when the UAV takes the RGB image, the vertical distance between the UAV and the ground is measured multiple times by the photoelectric ranging device, and the average value of the multiple distances obtained by the measurement is calculated. The average value of the multiple measured distances is used as the flight height of the UAV relative to the ground. Usually, the number of measurements can be selected as 10 times.

[0074] S2: Set the spectral resolution of the simulated visible light band according to the flight altitude of the drone and infrared spectral resolution In the embodiment of the present invention, several spectrum resolution setting values ​​of classic flight altitudes are given, specifically:

[0075] When the calculated flight altitude is 1000m, the final simulated target in the visible light band , and Spectral resolution The value is 10, that is , and The final simulation target of the infrared band is Spectral resolution The value is 2, that is It is divided into 2 spectrums.

[0076] When the calculated flight altitude is 500m, the final simulated target in the visible light band , and Spectral resolution The value is 20, that is , and The final simulation target of the infrared band is Spectral resolution The value is 3, that is It is divided into 3 spectrums.

[0077] When the calculated flight altitude is 100m, the final simulated target in the visible light band , and Spectral resolution The value is 30, that is , and The final simulation target of the infrared band is Spectral resolution The value is 4, that is It is divided into 4 spectrum segments.

[0078] The spectral resolution can be set according to the specific value of the flight altitude. There is no specific value. You only need to follow that the higher the flight altitude, the lower the spectral resolution should be.

[0079] In the embodiment of the present invention, the flight altitude of the drone when acquiring the RGB image is 100m. Therefore, , .

[0080] S3: First, the simulation spectrum segment processing process of the visible light band of the RGB image is as follows:

[0081] The R, G, and B components of the RGB image are interpolated using the following formula, and the interpolation of the R and G components is the RG spectrum segment, the interpolation of the G and B components is the GB spectrum segment, and the interpolation of the R and B components is the RB spectrum segment:

[0082] .

[0083] The R component is binarized, and the binarization threshold of the R component is the brightness average mean1 of the RG spectrum in the RGB image. The pixels whose R component is greater than the brightness average mean1 are assigned a value of 1; the pixels whose R component is less than or equal to the brightness average mean1 are assigned a value of 0. After the assignment, the following is obtained: Figure 3 The binarized image shown for:

[0084] .

[0085] The G component is binarized. The binarization threshold of the G component is the brightness average mean2 of the GB spectrum in the RGB image. The pixels whose G component is greater than the brightness average mean2 are assigned a value of 1; the pixels whose G component is less than or equal to the brightness average mean2 are assigned a value of 0. After the assignment, the following is obtained: Figure 4 The binarized image shown for:

[0086] .

[0087] The B component is binarized. The binarization threshold of the B component is the brightness average mean3 of the RB spectrum in the RGB image. The pixels whose B component is greater than the brightness average mean3 are assigned a value of 1; the pixels whose B component is less than or equal to the brightness average mean3 are assigned a value of 0. After the assignment, the following is obtained: Figure 5 The binarized image shown for:

[0088] .

[0089] Binarized image , and Decompose the spectrum into Spectrum, specifically:

[0090] Binarize the image through a for loop Regenerate a uint type image of 30 spectral segments according to the following formula , the generation rules are:

[0091] ;

[0092] in, Integers from 1 to 30.

[0093] Among them, the first spectrum obtained by the for loop like Figure 6 As shown; the 10th spectrum obtained by the for loop like Figure 7 As shown; the 20th spectrum obtained by the for loop like Figure 8 As shown; the 30th spectrum obtained by for loop like Fig. 9 shown.

[0094] Then use the for loop to convert the uint type image Generate the final image to be simulated , the generation rules are:

[0095] .

[0096] Among them, the first spectrum obtained by the for loop like Fig.10 As shown; the 10th spectrum obtained by the for loop like Fig.11 As shown; the 20th spectrum obtained by the for loop like Fig.12 As shown; the 30th spectrum obtained by for loop like Fig.13 shown.

[0097] Binarize the image through a for loop Regenerate a uint type image of 30 spectral segments according to the following formula , the generation rules are:

[0098] .

[0099] Among them, the first spectrum obtained by the for loop like Fig.14 As shown; the 10th spectrum obtained by the for loop like Fig.15 As shown; the 20th spectrum obtained by the for loop like Fig.16 As shown; the 30th spectrum obtained by for loop like Fig.17 shown.

[0100] Then use the for loop to convert the uint type image Generate the final image to be simulated , the generation rules are:

[0101] .

[0102] Among them, the first spectrum obtained by the for loop like Fig.18 As shown; the 10th spectrum obtained by the for loop like Fig.19 As shown; the 20th spectrum obtained by the for loop like Fig. 20 As shown; the 30th spectrum obtained by for loop like Fig.21 shown.

[0103] Binarize the image through a for loop Regenerate a uint type image of 30 spectral segments according to the following formula , the generation rules are:

[0104] .

[0105] Among them, the first spectrum obtained by the for loop like Fig. 22 As shown; the 10th spectrum obtained by the for loop like Fig.23 As shown; the 20th spectrum obtained by the for loop like Fig.24 As shown; the 30th spectrum obtained by for loop like Fig.25 shown.

[0106] Then use the for loop to convert the uint type image Generate the final image to be simulated , the generation rules are:

[0107] .

[0108] Among them, the first spectrum obtained by the for loop like Fig.26 As shown; the 10th spectrum obtained by the for loop like Fig. 27 As shown; the 20th spectrum obtained by the for loop like Fig.28 As shown; the 30th spectrum obtained by for loop like Fig.29 shown.

[0109] Through the above process, the R, G and B components of the RGB image are spectrally decomposed to obtain 30 simulated images with different spectral bands.

[0110] The simulation sub-spectral processing process of the infrared band of the RGB image is as follows:

[0111] The gray value Y of the RGB image is calculated according to the following formula:

[0112] .

[0113] The gray value Y is binarized. The binarization threshold of the gray value Y is the overall brightness average mean of the RGB image. The pixels whose gray value Y is greater than the brightness average mean are assigned a value of 1; the pixels whose gray value Y is less than or equal to the brightness average mean are assigned a value of 0. After the assignment, the binary image is obtained. for:

[0114] .

[0115] Binarized image Erosion processing is performed, and the image after corrosion is recorded as fushi. The binary image fushi after corrosion is expanded to different degrees to achieve regional expansion. To ensure the expansion effect, the embodiment of the present invention performs four expansion processes of different degrees. The binary images after expansion are obtained in order from small to large according to the expansion degree: pengzhang1, pengzhang2, pengzhang3 and pengzhang4. During the expansion process, if the brightness value of the pixel is greater than 255, it is assigned to 225 to prevent the pixel brightness from escaping.

[0116] Decompose the dilated binary image to generate an image of type uint8 , specifically:

[0117] like Fig.30 The first spectrum shown The image is:

[0118] ;

[0119] like Fig.31 The second spectrum shown The image is:

[0120] ;

[0121] like Fig.32 The third spectrum shown The image is:

[0122] ;

[0123] like Fig.33 The 4th spectrum shown The image is:

[0124] .

[0125] According to the spectral resolution of the infrared band , Selective Output ~ In this embodiment, , so the infrared spectrum output ~ .like , then the infrared spectrum should output ~ ;like , then the infrared spectrum should output and .

[0126] Through the above method, the RGB image obtained by aerial photography can be simulated to obtain images with multiple spectral resolutions. , , and The composed hyperspectral images can be directly used for analysis and research by trainees, which can significantly reduce the training cost of hyperspectral operators.

[0127] Although the embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and cannot be understood as limiting the present invention. Those skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention.

[0128] The above specific implementations of the present invention do not constitute a limitation on the protection scope of the present invention. Any other corresponding changes and modifications made based on the technical concept of the present invention should be included in the protection scope of the claims of the present invention.

Claims

1. A method for generating a simulated hyperspectral image based on an aerial image, characterized in that: include: S1: Obtain RGB images through drone aerial photography and measure the flight height of the drone relative to the ground; S2: Setting the number of spectral bands of the simulated visible light band according to the flight altitude and the number of infrared bands ; S3: performing weighted summation on the R, G, and B components of the RGB image, obtaining a weighted summation of the R and G components as an RG spectrum segment, a weighted summation of the G and B components as a GB spectrum segment, and a weighted summation of the R and B components as an RB spectrum segment; Binarize the R component to obtain a binary image , the binarization threshold of the R component is the average brightness of the RG spectrum; binarization of the G component obtains a binary image The binarization threshold of the G component is the average brightness of the GB spectrum; the B component is binarized to obtain a binary image , the binarization threshold of the B component is the average brightness of the RB spectrum; Binarize the image Decomposition Generation Image of spectrum ; Binarized image Decomposition Generation Image of spectrum ; Binarized image Decomposition Generation Image of spectrum ; Binarize the gray value Y of the RGB image to obtain a binary image , the binarization threshold of the gray value Y is the brightness average of the RGB image, and the binarized image Decomposition Generation Image of spectrum ; S3 is specifically: Binarize the image through a for loop Generate according to the following formula uint type image of spectral segments : ； in, Indicates that the values ​​are taken from 1 to integer; Use a for loop to convert the uint type image Generate image : ； Binarize the image through a for loop Generate according to the following formula uint type image of spectral segments : ； Use a for loop to convert the uint type image Generate image : ； Binarize the image through a for loop Generate according to the following formula uint type image of spectral segments : ； Use a for loop to convert the uint type image Generate image : 。 2. The method for generating a simulated hyperspectral image based on an aerial image according to claim 1, characterized in that: When the drone is used to obtain an RGB image, the distance between the drone and the ground is obtained multiple times, and the average value of the distance is used as the flight height.

3. The method for generating a simulated hyperspectral image based on an aerial image according to claim 2, characterized in that: The distance between the drone and the ground is obtained 10 times, and the average value of the 10 distances is used as the flight height.

4. The method for generating a simulated hyperspectral image based on an aerial image according to claim 1, characterized in that: In S2, the number of spectral bands of the visible light band is switched according to the flight altitude. and the number of infrared bands : When the flight altitude is 1000m, , ; When the flight altitude is 500m, , ; When the flight altitude is 100m, , .

5. The method for generating a simulated hyperspectral image based on an aerial image according to claim 1, characterized in that: The weighted summation formula of the RG spectrum segment, the GB spectrum segment and the RB spectrum segment is: 。 6. The method for generating a simulated hyperspectral image based on an aerial image according to claim 1, characterized in that: In S3, the gray value Y of the RGB image is calculated as: 。 7. The method for generating a simulated hyperspectral image based on an aerial image according to claim 6, characterized in that: In S3, the binary image Carry out corrosion and dilation to different degrees in sequence, and decompose the dilated image to generate Image of spectrum .

Images