Moving target information detection method based on multi-slit hyperspectral data

Abstract

The invention discloses a moving target information detection method based on multi-slit hyperspectral data, and solves the problem that the form information and motion information of a moving target are difficult to accurately obtain in the prior art. Aiming at imaging position differences of a moving target in hyperspectral data obtained in different slits, a target motion area and motion information are obtained through data difference and template matching, and sampling correction is carried out on a target form through the motion information and iterative feedback operation is carried out on the form and the motion information. Accurate acquisition of motion information and form information of the moving target is realized.

Description

Technical Field

[0001] The invention relates to a moving target information detection method, and in particular to a moving target information detection method based on multi-slit hyperspectral data. Background Art

[0002] The multi-slit hyperspectral imaging system acquires the overall data of the entire target through push-scanning. Since the multi-slit hyperspectral imaging system adopts multi-slit spectrometry, it can acquire multiple sets of hyperspectral image data of the same target at different times through different slit spectrometry through push-scanning, thus realizing multiple sampling of the target.

[0003] The multiple sampling characteristics of the multi-slit hyperspectral imaging system enable it to detect information on moving targets. The multi-slit hyperspectral imaging system images the same area of ​​the target through different slits, and can obtain the imaging position changes of the moving target during the sampling interval. According to the changes in its imaging position, the speed and direction information of the target movement can be obtained.

[0004] At the same time, since the multi-slit hyperspectral imaging system is a push-scan imaging system, oversampling or undersampling will occur in the imaging of moving targets, and the detection of target morphological information and target motion information are coupled with each other, making it difficult for existing technologies to accurately obtain the morphological information and motion information of moving targets. Summary of the invention

[0005] In order to solve the technical problem that it is difficult to accurately obtain the morphological information and motion information of a moving target in the prior art, the present invention provides a moving target information detection method based on multi-slit hyperspectral data.

[0006] The inventive concept of the present invention:

[0007] The present invention aims at the imaging position difference of the moving target in the hyperspectral data obtained at different slits, obtains the target motion area and motion information through data differentiation and template matching, and realizes the accurate acquisition of the moving target motion information and morphology information through sampling correction of the target morphology by motion information and iterative feedback operation of morphology and motion information.

[0008] In order to achieve the above purpose and complete the above invention concept, the present invention adopts the following technical solutions:

[0009] A moving target information detection method based on multi-slit hyperspectral data, which is special in that it includes the following steps:

[0010] Step 1, obtain multiple hyperspectral image data of the same target pushed and scanned by a multi-slit hyperspectral system, and pre-process them separately, taking the hyperspectral image data obtained by the first slit as the benchmark, recorded as the first hyperspectral data, and the hyperspectral image data obtained by other slits are recorded as the second hyperspectral data; obtain the target motion area in the first hyperspectral data and the second hyperspectral data;

[0011] Step 2: Determine the difference array P in each target motion area 1 The maximum specific difference value in the spectrum segment N is obtained, and the pixel position coordinates corresponding to the maximum specific difference value are obtained. With the pixel position coordinates as the center, a rectangular matching template area K is selected in each target motion area. 2 ;

[0012] Step 3: Based on each matching template region K 2 In the corresponding difference array P 1 The position in the spectrum segment N is used to find the matching template area K in the spectrum segment N in the first hyperspectral data. 1 , with the matching template area K 1 The grayscale data in the image is used as a matching template, and the coordinates of the central pixel of the matching template are recorded as the first matching coordinates; and the matching template is used to perform template matching with the target motion area in each spectral segment N of the second hyperspectral data, and the coordinates of the central pixel of each matching area are recorded as the second matching coordinates, and the first matching coordinates are subtracted from each second matching coordinate to obtain a motion vector of the target;

[0013] Step 4, according to the angle between each motion vector and the push-scan motion vector of the multi-slit hyperspectral system in the image coordinate system, and the correspondence between the push-scan motion vector of the multi-slit hyperspectral system in the image coordinate system and in the world coordinate system, respectively calculate the motion direction of the target in the world coordinate system and take the average to obtain the average motion direction;

[0014] Step 5: Get the target's moving speed. The moving speed and the average moving direction obtained in step 4 are the current moving information of the target, recorded as vector

[0015] Step 6: According to the average motion direction calculated in step 4 and the motion speed obtained in step 5, the target motion region in the spectrum segment N in the first hyperspectral data and the target motion region in the spectrum segment N in each second hyperspectral data are sampled and corrected to obtain a corrected target motion region. The corrected target motion region corresponding to the spectrum segment N in the first hyperspectral data is recorded as the first motion region T. 1 , the correction target motion region corresponding to the spectrum segment N in the second hyperspectral data is the second motion region;

[0016] Step 7: Compare the grayscale values ​​in each second motion area with the grayscale values ​​in the first motion area T 1 The grayscale values ​​in the second motion region are differentiated and the absolute values ​​are taken to obtain a differential array P of the second motion region. 2 ; Use the method from step 2 to step 5 to calculate the corresponding difference array P 2 , the first motion area T 1 The grayscale data in the inner and second motion areas are processed to obtain vector

[0017] Step 8. Calculate the vector Comparison vector The speed change ratio ρ is compared with the second preset threshold value b:

[0018] If ρ≤b, then the vector is the target's motion information, executing step 9;

[0019] If ρ>b, use the vector Alternative vector First sports area T 1 The target motion region in the spectrum segment N in the first hyperspectral data is replaced by the second motion region, and the target motion region in the spectrum segment N in each second hyperspectral data is replaced by the second motion region, and the process returns to step 6 until ρ≤b;

[0020] Step 9: Use the method in step 6 to obtain the vector For the first motion region T 1 The grayscale data in the first motion area T is sampled and corrected to obtain the corresponding first motion area T 2 ;

[0021] Step 10: Set the first motion area T 2 The length and width pixel values ​​of the target are multiplied by the imaging spatial resolution of the multi-slit hyperspectral system to obtain the morphological information of the target, wherein the morphological information of the target includes the length and width of the target;

[0022] Complete the detection of the target's motion information and morphological information.

[0023] Furthermore, step 1 specifically includes:

[0024] 1.1. Acquire multiple hyperspectral image data of the same target by a multi-slit hyperspectral system, and pre-process them separately. The hyperspectral image data acquired by the first slit is used as the benchmark and recorded as the first hyperspectral data. The hyperspectral image data acquired by other slits are all recorded as the second hyperspectral data. The first slit is the slit that is first imaged in the multi-slit hyperspectral system.

[0025] 1.2. Differentiate each second hyperspectral data with the first hyperspectral data and take the absolute value to obtain the differential array P of each second hyperspectral data 1 ;

[0026] 1.3. For the difference array P 1 The specific difference values ​​in each spectral segment are averaged, and the spectral segment with the largest average is determined, which is recorded as spectral segment N;

[0027] 1.4. Binarizing the specific differential values ​​in each spectrum segment N based on the first preset threshold a to obtain binary data;

[0028] 1.5. Obtain the boundary range of the non-zero area in the binary data, translate the X-axis and Y-axis of its image coordinate system to the edge of the boundary range, and each obtained rectangular area is recorded as the target motion area in the first hyperspectral data and the corresponding second hyperspectral data.

[0029] Furthermore, step 5 is specifically to obtain the target's motion speed according to the modulus of each of the motion vectors, the imaging spatial resolution of the multi-slit hyperspectral system, and the imaging time difference between the other slits and the first slit. The motion speed and the average motion direction obtained in step 4 are the current motion information of the target, which is recorded as vector

[0030] Further, step 5 is specifically to multiply the imaging spatial resolution of the multi-slit hyperspectral system by the modulus of each of the motion vectors to obtain the target's motion distance, and to divide each motion distance by the corresponding imaging time difference and take the average to obtain the target's motion speed. The motion speed and the average motion direction obtained in step 4 are the current motion information of the target, recorded as vector

[0031] Furthermore, step 1.4 is specifically to binarize the specific differential values ​​in each spectral segment N based on the first preset threshold a, assigning a value of 1 to the differential data that is greater than or equal to the first preset threshold a, and assigning a value of 0 to the differential data that is less than the first preset threshold a, to obtain binary data.

[0032] Furthermore, in step 1.4, the first preset threshold a is 1.2 to 3 times the dark current fluctuation range of the detector in the multi-slit hyperspectral system.

[0033] Furthermore, in step 1.4, the first preset threshold a is 1.5 times the dark current fluctuation range of the detector in the multi-slit hyperspectral system.

[0034] Furthermore, step 6 specifically includes performing sampling correction on the target motion region in the spectrum segment N in the first hyperspectral data and the target motion region in the spectrum segment N in each second hyperspectral data respectively:

[0035] If the average motion direction calculated in step 4 is the same as the push-scan direction of the multi-slit hyperspectral system, the push-scan speed of the multi-slit hyperspectral system is subtracted from the motion speed obtained in step 5 to obtain the first imaging push-scan absolute speed of the target, and oversampling correction is performed according to the first imaging push-scan absolute speed;

[0036] If the average motion direction calculated in step 4 is opposite to the push-scan direction of the multi-slit hyperspectral system, then the push-scan speed of the multi-slit hyperspectral system is added to the motion speed obtained in step 5 to obtain the second imaging push-scan absolute speed of the target, and undersampling correction is performed according to the second imaging push-scan absolute speed;

[0037] The corrected target motion region is obtained, and the corrected target motion region corresponding to the spectrum segment N in the first hyperspectral data is recorded as the first motion region T 1 , the correction target motion region corresponding to the spectrum segment N in the second hyperspectral data is the second motion region.

[0038] Furthermore, step 8 specifically includes:

[0039] 8.1. Calculating vectors Comparison vector The speed change ratio ρ:

[0040]

[0041] 8.2. Compare the speed change ratio ρ with the second preset threshold value b:

[0042] If ρ≤b, then the vector is the motion information of the target, executing step 9;

[0043] If ρ>b, use the vector Alternative vector First sports area T 1 The target motion region in the spectrum segment N in the first hyperspectral data is replaced by the second motion region, and the target motion region in the spectrum segment N in each second hyperspectral data is replaced by the second motion region, and the process returns to step 6 until ρ≤b.

[0044] Furthermore, in step 2, the matching template area K 2 The length and width of the target motion area are 1 / 3 of the length and width respectively, and the minimum is 3 pixels;

[0045] In step 1.1, the preprocessing includes dark current correction, relative radiation correction, bad pixel correction and image shift correction.

[0046] Beneficial effects of the present invention:

[0047] The present invention aims at the characteristics of multi-slit hyperspectral image data, utilizes the time difference between multi-slit hyperspectral data, and provides a moving target information detection method based on multi-slit hyperspectral data, and obtains the target motion area by using the difference method; in view of the coupling influence of the target morphological information and motion information, multiple iterative calculations are used to obtain the accurate motion information and morphological information of the target, and while obtaining the target hyperspectral information, the detection of its motion and morphological information is realized, which is of great significance for expanding the use of multi-slit hyperspectral data in navigation target monitoring and aviation aircraft monitoring in the low-altitude economic field. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of an embodiment of a moving target information detection method based on multi-slit hyperspectral data of the present invention;

[0049] Figure 2 It is a schematic diagram of obtaining the target motion area in step 5 in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] The technical solution of the present invention will be clearly and completely described below in conjunction with the accompanying drawings and embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0051] A moving target information detection method based on multi-slit hyperspectral data. In this embodiment, a three-slit hyperspectral imaging system is selected, the spatial dimension along the track direction × vertical track direction is 20×10, and the spectral dimension is 5 spectral segments.

[0052] like Figure 1 As shown, the detection method comprises the following steps:

[0053] Step 1: Obtain multiple hyperspectral image data of the same target by a multi-slit hyperspectral system, and pre-process them separately. The hyperspectral image data obtained by the first slit is used as the benchmark and recorded as the first hyperspectral data. The hyperspectral image data obtained by other slits are recorded as the second hyperspectral data. Obtain the target motion area in the first hyperspectral data and the second hyperspectral data. Specifically include:

[0054] 1.1. Acquire multiple hyperspectral image data of the same target by a multi-slit hyperspectral system, and pre-process them separately. The hyperspectral image data acquired by the first slit is taken as a reference, recorded as the first hyperspectral data D1, and the hyperspectral image data acquired by the other two slits are recorded as the second hyperspectral data, represented by D2 and D3 respectively; the first slit is the slit that is first imaged in the multi-slit hyperspectral system; in this embodiment, the pre-processing includes dark current correction, relative radiation correction, bad pixel correction and image shift correction.

[0055] 1.2. Difference the second hyperspectral data D2 and D3 with the first hyperspectral data D1 and take the absolute value to obtain the differential array P of the second hyperspectral data 11 , P 12 :

[0056] P 11 =|D2-D1|

[0057] P 12 =|D3-D1|

[0058] 1.3. For the difference array P 11 , P 12 The specific difference values ​​in each spectral segment are averaged to determine the spectral segment with the largest average value, which is recorded as spectral segment N; then each second hyperspectral data corresponds to a spectral segment N; in this embodiment, P 11 , P 12 The selected spectral segments are the same, which are both spectral segment 5.

[0059] 1.4. Based on the first preset threshold a, the difference array P 11 , P 12 The specific differential values ​​in each spectral band 5 are binarized, and the differential data greater than or equal to the first preset threshold value a is assigned a value of 1, and the differential data less than the first preset threshold value a is assigned a value of 0, thereby obtaining binary data; the first preset threshold value a is 1.2 to 3 times the dark current fluctuation range of the detector in the multi-slit hyperspectral system, and 1.5 times is preferably used in this embodiment.

[0060] 1.5. Obtain the boundary range of the non-zero area in the binary data, translate the X-axis and Y-axis of its image coordinate system to the edge of the boundary range, and record the obtained rectangular area Q1 as the target motion area in the first hyperspectral data D1 and the second hyperspectral data D2, and record the obtained rectangular area Q2 as the target motion area in the first hyperspectral data D1 and the second hyperspectral data D3; set the size of the rectangular area Q1 to be 10 length × 5 width; set the size of the rectangular area Q2 to be 15 length × 9 width.

[0061] Step 2: Determine the difference array P in each target motion area 11 , P 12The maximum specific difference value in the spectrum segment 5 is obtained, and the pixel position coordinates corresponding to the maximum specific difference value are obtained. With the pixel position coordinates as the center, a rectangular matching template area K is selected in each target motion area. 21 , K 22 ; Matching template area K 21 , K 22 The length i and width j of the target motion area Q1 and Q2 are respectively 1 / 3 of the length and width and are rounded up to a minimum of 3 pixels. In this embodiment, the matching template area K 21 The length i and width j of are both 3. Matching template area K 22 The length i and width j are 5 pixels and 3 pixels respectively.

[0062] Step 3: Based on the matching template area K 21 , K 22 In the difference array P 11 , P 12 The position in the spectrum segment 5 is respectively found to match the template area K in the spectrum segment 5 in the first hyperspectral data. 11 , K 12 , to find the matching template area K 11 , K 12 The grayscale data in are used as matching templates, and the coordinates of the central pixel of the matching template are recorded as the first matching coordinates; and the matching template is used to perform template matching with the target motion area in the spectral segment 5 of the second hyperspectral data D2 and D3, and the coordinates of the central pixel of the matching area are recorded as the second matching coordinates. The first matching coordinates are subtracted from each second matching coordinate to obtain the motion vectors v11 and v12 of the target.

[0063] Step 4: According to the included angle between the motion vectors v11 and v12 and the push-scan motion vector of the multi-slit hyperspectral system in the image coordinate system, and the corresponding relationship between the push-scan motion vector of the multi-slit hyperspectral system in the image coordinate system and the world coordinate system, the motion direction of the target in the world coordinate system is calculated and the average is taken to obtain the average motion direction.

[0064] Step 5: Compare the imaging spatial resolution d of the multi-slit hyperspectral system with the modulus of each motion vector. Multiply them respectively to get the target's moving distance, and divide each moving distance by the corresponding imaging time difference t 1 ,t 2 And take the average to get the target's movement speed. Then the movement speed and the average movement direction obtained in step 4 are the current movement information of the target, recorded as vector The formula used is as follows:

[0065]

[0066] Step 6: Average movement direction calculated according to step 4 The target motion region in the spectrum segment 5 in the first hyperspectral data and the target motion region in the spectrum segment 5 in the second hyperspectral data D2 and D3 are sampled and corrected respectively according to the motion speed obtained in step 5:

[0067] If the average motion direction calculated in step 4 The push-scan direction is the same as that of the multi-slit hyperspectral system, and there is an oversampling phenomenon. The push-scan speed of the multi-slit hyperspectral system is subtracted from the target movement speed to obtain the first imaging push-scan absolute speed of the target. According to the first imaging push-scan absolute speed, oversampling correction is performed based on the first detection time of the target.

[0068] If the average motion direction calculated in step 4 is opposite to the push-sweep direction of the multi-slit hyperspectral system, and there is undersampling, then the push-sweep speed of the multi-slit hyperspectral system is added to the target motion speed to obtain the second imaging push-sweep absolute speed of the target, and undersampling correction is performed based on the second imaging push-sweep absolute speed. In this embodiment, undersampling correction is performed based on the first target detection time as a reference to obtain the corrected target motion area. The corrected target motion area corresponding to the spectrum segment 5 in the first hyperspectral data is recorded as the first motion area T 1 , the correction target motion region corresponding to the spectrum segment 5 in the second hyperspectral data is the second motion region.

[0069] Step 7: Compare the grayscale values ​​in each second motion area with the grayscale values ​​in the first motion area T 1 The grayscale values ​​in the second motion area are differentiated and the absolute values ​​are taken to obtain the differential array P of the second motion area. 21 , P 22 ; Use the method from step 2 to step 5 to calculate the difference array P 21 , P 22 , the first motion area T 1 The grayscale data in the inner and second motion areas are processed to obtain vector Because the average has been taken, the vector is a numerical value.

[0070] When using the method from step 2 to step 5, the difference array P 21 , P 22 Corresponding to the difference array P in steps 2 to 5 21 , P 22 Segment 5, first motion region T 1 The inner grayscale data corresponds to the grayscale data of the spectral segment 5 in the first hyperspectral data in steps 2 to 5, and the grayscale data in the second motion area corresponds to the grayscale data of the spectral segment 5 in the second hyperspectral data in steps 2 to 5.

[0071] Step 8. Calculate the vector Comparison vector The speed change ratio ρ:

[0072]

[0073] Compare the speed change ratio ρ with the second preset threshold b:

[0074] If ρ≤b, then the vector is the motion information of the target, and step 14 is executed.

[0075] If ρ>b, use the vector Alternative vector First sports area T 1 The target motion region in the spectrum segment N in the first hyperspectral data is replaced by the second motion region, and the target motion region in the spectrum segment N in each second hyperspectral data is replaced by the second motion region, and the process returns to step 6 until ρ≤b.

[0076] The second preset threshold b is set according to the required detection accuracy.

[0077] Step 9: Use the method in step 6 to obtain the vector For the first motion region T 1 The grayscale data in the first motion area T is sampled and corrected to obtain the corresponding first motion area T 2 .

[0078] Step 10: Set the first motion area T 2 The length pixel value T 2x , width pixel value T 2y The target morphological information is obtained by multiplying the imaging spatial resolution d of the multi-slit hyperspectral system. The target morphological information includes the length G2 of the target x Hekuan G2 y ; The formula used is as follows:

[0079] G2 x =T 2x *d

[0080] G2 y =T 2y *d

[0081] Complete the detection of the target's motion information and morphological information.

[0082] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A moving target information detection method based on multi-slit hyperspectral data, characterized in that: The following steps are involved: Step 1, obtain multiple hyperspectral image data of the same target pushed and scanned by a multi-slit hyperspectral system, and pre-process them separately, taking the hyperspectral image data obtained by the first slit as the benchmark, recorded as the first hyperspectral data, and the hyperspectral image data obtained by other slits are recorded as the second hyperspectral data; obtain the target motion area in the first hyperspectral data and the second hyperspectral data; Step 2: Determine the maximum specific differential value in the spectrum segment N of the differential array P1 in each target motion area, and obtain the pixel position coordinates corresponding to the maximum specific differential value, and select a rectangular matching template area K2 in each target motion area with the pixel position coordinates as the center; Step 3, based on the position of each matching template region K2 in the corresponding spectrum segment N of the differential array P1, respectively find the matching template region K1 in the spectrum segment N of the first hyperspectral data, use the grayscale data in the matching template region K1 as the matching template, and record the coordinates of the central pixel of the matching template as the first matching coordinates; and use the matching template to perform template matching with the target motion regions in the spectrum segments N of each second hyperspectral data, respectively record the coordinates of the central pixel of each matching region as the second matching coordinates, and subtract the first matching coordinates from each second matching coordinate to obtain a motion vector of a target; Step 4, according to the angle between each motion vector and the push-scan motion vector of the multi-slit hyperspectral system in the image coordinate system, and the correspondence between the push-scan motion vector of the multi-slit hyperspectral system in the image coordinate system and in the world coordinate system, respectively calculate the motion direction of the target in the world coordinate system and take the average to obtain the average motion direction; Step 5: Get the target's moving speed. The moving speed and the average moving direction obtained in step 4 are the current moving information of the target, recorded as vector Step 6: According to the average motion direction calculated in step 4 and the motion speed obtained in step 5, the target motion region in the spectrum segment N in the first hyperspectral data and the target motion region in the spectrum segment N in each second hyperspectral data are sampled and corrected respectively to obtain a corrected target motion region. The corrected target motion region corresponding to the spectrum segment N in the first hyperspectral data is recorded as the first motion region T1, and the corrected target motion region corresponding to the spectrum segment N in the second hyperspectral data is recorded as the second motion region; Step 7: Difference the grayscale values ​​in each second motion region with the grayscale values ​​in the first motion region T1 and take the absolute value to obtain a difference array P2 of the second motion region; use the method of steps 2 to 5 to process the corresponding difference array P2, the grayscale data in the first motion region T1 and the grayscale data in the second motion region to obtain vector Step 8. Calculate vector Comparison vector The speed change ratio ρ is compared with the second preset threshold value b: If ρ≤b, then the vector is the target's motion information, executing step 9; If ρ>b, use the vector Alternative vector The first motion region T1 replaces the target motion region in the spectrum segment N in the first hyperspectral data, and the second motion region replaces the target motion region in the spectrum segment N in each second hyperspectral data, and returns to step 6 until ρ≤b; Step 9: Use the method in step 6 to obtain the vector Sampling and correcting the grayscale data in the first motion region T1 to obtain the corresponding first motion region T2; Step 10, multiplying the length and width pixel values ​​of the first motion area T2 by the imaging spatial resolution of the multi-slit hyperspectral system to obtain the morphological information of the target, wherein the morphological information of the target includes the length and width of the target; Complete the detection of the target's motion information and morphological information.

2. The moving target information detection method based on multi-slit hyperspectral data according to claim 1 is characterized in that: Step 1 specifically includes: 1.

1. Acquire multiple hyperspectral image data of the same target by a multi-slit hyperspectral system, and pre-process them separately. The hyperspectral image data acquired by the first slit is used as the benchmark and recorded as the first hyperspectral data. The hyperspectral image data acquired by other slits are all recorded as the second hyperspectral data. The first slit is the slit that is first imaged in the multi-slit hyperspectral system. 1.

2. Differentiate each second hyperspectral data from the first hyperspectral data and take the absolute value to obtain a differential array P1 of each second hyperspectral data; 1.

3. Take the average of the specific difference values ​​in each spectrum segment of each difference array P1, and determine the spectrum segment with the largest average value, which is recorded as spectrum segment N; 1.

4. Binarizing the specific differential values ​​in each spectrum segment N based on the first preset threshold a to obtain binary data; 1.

5. Obtain the boundary range of the non-zero area in the binary data, translate the X-axis and Y-axis of its image coordinate system to the edge of the boundary range, and each obtained rectangular area is recorded as the target motion area in the first hyperspectral data and the corresponding second hyperspectral data.

3. The moving target information detection method based on multi-slit hyperspectral data according to claim 1 or 2, characterized in that: Step 5 is specifically to obtain the target's motion speed according to the modulus of each of the motion vectors, the imaging spatial resolution of the multi-slit hyperspectral system, and the imaging time difference between the other slits and the first slit. The motion speed and the average motion direction obtained in step 4 are the current motion information of the target, which is recorded as vector 4. The moving target information detection method based on multi-slit hyperspectral data according to claim 3 is characterized in that: Step 5 is specifically to multiply the imaging spatial resolution of the multi-slit hyperspectral system by the modulus of each of the motion vectors to obtain the target's motion distance, and to divide each motion distance by the corresponding imaging time difference and take the average to obtain the target's motion speed. The motion speed and the average motion direction obtained in step 4 are the current motion information of the target, recorded as vector 5. The moving target information detection method based on multi-slit hyperspectral data according to claim 4 is characterized in that: Specifically, step 1.4 is to binarize the specific differential values ​​in each spectral segment N based on the first preset threshold a, assigning a value of 1 to the differential data that is greater than or equal to the first preset threshold a, and assigning a value of 0 to the differential data that is less than the first preset threshold a, thereby obtaining binary data.

6. The moving target information detection method based on multi-slit hyperspectral data according to claim 5 is characterized in that: In step 1.4, the first preset threshold a is 1.2 to 3 times the dark current fluctuation range of the detector in the multi-slit hyperspectral system.

7. The moving target information detection method based on multi-slit hyperspectral data according to claim 6 is characterized in that: In step 1.4, the first preset threshold a is 1.5 times the dark current fluctuation range of the detector in the multi-slit hyperspectral system.

8. The moving target information detection method based on multi-slit hyperspectral data according to claim 7 is characterized in that: Step 6 specifically includes performing sampling correction on the target motion region in the spectrum segment N in the first hyperspectral data and the target motion region in the spectrum segment N in each second hyperspectral data respectively: If the average motion direction calculated in step 4 is the same as the push-scan direction of the multi-slit hyperspectral system, the push-scan speed of the multi-slit hyperspectral system is subtracted from the motion speed obtained in step 5 to obtain the first imaging push-scan absolute speed of the target, and oversampling correction is performed according to the first imaging push-scan absolute speed; If the average motion direction calculated in step 4 is opposite to the push-scan direction of the multi-slit hyperspectral system, then the push-scan speed of the multi-slit hyperspectral system is added to the motion speed obtained in step 5 to obtain the second imaging push-scan absolute speed of the target, and undersampling correction is performed according to the second imaging push-scan absolute speed; The corrected target motion region is obtained, and the corrected target motion region corresponding to the spectrum segment N in the first hyperspectral data is recorded as the first motion region T1, and the corrected target motion region corresponding to the spectrum segment N in the second hyperspectral data is recorded as the second motion region.

9. The moving target information detection method based on multi-slit hyperspectral data according to claim 8, characterized in that: Step 8 specifically includes: 8.

1. Calculating vectors Comparison vector The speed change ratio ρ: 8.

2. Compare the speed change ratio ρ with the second preset threshold b: If ρ≤b, then the vector is the motion information of the target, executing step 9; If ρ>b, use the vector Alternative vector The first motion region T1 replaces the target motion region in the spectrum segment N in the first hyperspectral data, the second motion region replaces the target motion region in the spectrum segment N in each second hyperspectral data, and the process returns to step 6 until ρ≤b.

10. The moving target information detection method based on multi-slit hyperspectral data according to claim 9, characterized in that: In step 2, the length and width of the matching template area K2 are respectively 1 / 3 of the length and width of the target motion area and are rounded up to a minimum of 3 pixels; In step 1.1, the preprocessing includes dark current correction, relative radiation correction, bad pixel correction and image shift correction.

Images