Unmanned aerial vehicle remote sensing monitoring method based on camera response function calibration

Abstract

The invention belongs to the field of environment remote sensing monitoring, and particularly relates to an unmanned aerial vehicle remote sensing monitoring method based on camera response function calibration, which comprises the following steps: acquiring an image sequence of a monitored target area, and constructing a camera response function; the unmanned aerial vehicle obtains RGB image data of a monitoring target and splits the RGB image data into R, G and B single-band grayscale images; converting the DN values of the R, G and B single-band grayscale images into E values according to a camera response function; determining whether a reference plate is placed in the current imaging area or not, and if yes, converting the E value into reflectivity, and obtaining an index value or a band operation value of the monitoring target; otherwise, obtaining an index value or a wave band operation value of the monitoring target through the E value; finally, filtering is carried out, a linear graph is obtained, and remote sensing monitoring of the unmanned aerial vehicle on a monitoring target containing vegetation and water surface floating algae is completed. According to the method, indexes of vegetation, macroalgae and the like in an environment with relatively weak signals or relatively high scene radiance are calculated, and the remote sensing monitoring precision of the unmanned aerial vehicle is improved.

Description

Technical Field

[0001] The invention belongs to the field of environmental remote sensing monitoring, and in particular relates to an unmanned aerial vehicle remote sensing monitoring method based on camera response function calibration. Background Art

[0002] Unmanned aerial vehicle remote sensing (UAVRS) technology has the advantages of high spatial and temporal resolution of images, low cost of acquiring images, and flexibility. It is an important supplement to satellite remote sensing. At present, high-resolution UAV remote sensing technology plays an increasingly important role and has been widely used in marine environmental monitoring, land use surveys, crop growth monitoring and yield estimation, forest pest and disease protection and monitoring, and other fields. Green tide is a common type of algal bloom (Macroalgae blooms, MABs), which is mainly caused by the explosive growth of large floating green algae in seawater. In recent years, due to the influence of human activities and climate change, green tide events have become more frequent around the world. During the outbreak of green tide, local maritime traffic and fishery production activities are affected, and green tide accumulation on the beach destroys the sightseeing and tourism activities in coastal areas, causing huge economic losses to the local area. Taking the UAV remote sensing monitoring of the marine Enteromorpha green tide as an example, it is of great practical significance to actively carry out water environment monitoring and floating macroalgae monitoring based on UAV remote sensing. In particular, by using centimeter-level resolution drone images combined with satellite imagery to analyze information such as the spectrum of large algae, a large algae coverage inversion model based on satellite images can be established, providing technical support for high-precision estimation of large algae biomass.

[0003] However, the main imaging elements of the Red-Green-Blue (RGB) cameras carried by mainstream drones are charge coupled device (CCD) arrays or complementary metal oxide semiconductor (CMOS) arrays. When the camera uses the CCD array or CMOS array for imaging, although the charge collected on the CCD element or CMOS element is proportional to its irradiance, general digital cameras perform nonlinear mapping on the output of the CCD or CMOS before writing it to the storage medium. This nonlinear mapping of scene radiance to digital quantization value (Digital Number, DN) is called the camera response function. The camera response function imitates the film response characteristics, allowing the camera to better represent overly dark and overly bright scenes. However, the existence of the camera response function brings a series of problems to image processing: the DN value in the RGB image is generally nonlinearly related to the scene radiance, which is manifested as the low value of the DN value being raised and the high value being suppressed, that is, the DN value increases faster when the radiance is low, and increases slower when the radiance is high, resulting in the image data of different light intensities being incomparable. Especially when using drones to carry out water environment monitoring, it is easy to be disturbed by the sun glare on the water surface. The existence of glare makes the impact of nonlinear response on drone images more significant, which directly affects the working quality of drones in glare environments.

[0004] Reducing the nonlinear response of UAV RGB cameras has important application significance for quantitative analysis in UAV remote sensing. At present, image preprocessing through camera response function has been widely used in high dynamic range image synthesis, motion blurred image deblurring, image stitching and other fields, but its application in the field of UAV quantitative remote sensing is relatively small. Summary of the invention

[0005] The digital quantization value (DN) of the Red-Green-Blue (RGB) image acquired by the drone is nonlinearly related to the object radiance, resulting in the incomparability of drone image data under different light intensities, especially in environments with weak signals or high scene radiance, which makes it impossible to perform accurate analysis. In view of the above problems, the present invention provides a drone remote sensing monitoring method based on camera response function calibration to reduce nonlinear response and improve the accuracy of drone remote sensing monitoring.

[0006] The technical solution adopted by the present invention to achieve the above-mentioned purpose is: a UAV remote sensing monitoring method based on camera response function calibration, comprising the following steps:

[0007] S1: Control the drone to fly at a fixed position, shoot the monitoring target area with different exposure times, obtain the image sequence of the monitoring target area, obtain the DN value to E value conversion table based on the Debevec algorithm, and construct the camera response function of the drone camera R, G, B bands;

[0008] S2: According to the actual situation of the monitoring target, determine whether there is a stable platform in the drone imaging area as a placement platform for the reference plate, so as to determine whether to place the reference plate in the drone imaging area;

[0009] S3: The drone adjusts the camera lens toward the ground to obtain RGB image data of the monitoring target including vegetation and algae floating on the water surface;

[0010] S4: Split the RGB image data into R, G, B single-band grayscale images according to the R, G, B bands;

[0011] S5: According to the camera response function in step S1, the DN value of the R, G, B single-band grayscale image is converted into an energy value linearly related to the object radiance, that is, an E value;

[0012] S6: Determine whether a reference plate is placed in the current drone imaging area. If a reference plate is placed, convert the E value into reflectivity and then execute step S7;

[0013] S7: If a reference plate is placed in the drone imaging area, the index value or band calculation value of the monitoring target is obtained through the reflectivity; conversely, if no reference plate is placed in the drone imaging area, the index value or band calculation value of the monitoring target is obtained through the E value;

[0014] S8: filtering the index value or band operation value of the monitoring target obtained in step S7 to remove abnormal values ​​including those higher than a set threshold and those lower than a set threshold;

[0015] S9 obtains a linear graph based on the filtered index value or band operation value to complete the remote sensing monitoring of the drone on the monitoring targets including vegetation and floating algae on the water surface.

[0016] The drone camera is a CCD camera or a CMOS camera.

[0017] The camera response function of the drone camera RGB band is constructed as follows:

[0018] The camera response function for the R band is constructed as:

[0019] y=5.542e-11*x 5 -3.763e-08*x 4 +9.692e-06*x 3 -0.001192*x 2+0.08768*x-3.852

[0021] Furthermore, the camera response function of the G band is constructed as:

[0022] y=3.565e-11*x 5 -2.414e-08*x 4 +6.265e-06*x 3 -0.0008077*x 2 +0.0717*x-3.857

[0024] Furthermore, the camera response function of the B band is constructed as:

[0025] y=3.953e-11*x 5 -2.55e-08*x 4 +6.431e-06*x 3 -0.000836*x 2 +0.07591*x-4.035

[0027] Among them, x is the DN value in the drone RGB image, and y is the E value.

[0028] The reference plate is a diffuse reflection reference plate with a known fixed reflectivity;

[0029] The spectral range of the reference plate is 400nm to 2500nm, and the hemispherical reflectance of the reference plate at a wavelength of 550nm is any one of 2%, 5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, and 99%.

[0030] If a reference plate is placed in the drone imaging area, the exposure time needs to be adjusted according to the DN value of the reference plate to prevent oversaturation of the reference plate. The specific adjustment method is: adjust the exposure time so that the DN value of the diffuse reflectance standard reference plate is no higher than 225.

[0031] In step S6, the E value is converted into reflectivity, specifically:

[0032]

[0033] Among them, R(λ) represents the reflectivity of the detection target, L(λ) represents the E value of the detection target, and L p (λ) represents the E value of the reference plate, and ρ(λ) is the reflectivity of the reference plate.

[0034] The index value or band operation value of the monitoring target is obtained, that is:

[0035] ZS=a 1 *r 1 +a 2 *r 2 +a 3 *r 3 ;

[0036] Among them, r is the reflectivity or E value, subscripts 1, 2, and 3 represent the R, G, and B bands respectively, and parameter a 1 、a 2 、a 3 It is the band operation coefficient value of the monitored target.

[0037] The step S8 is specifically as follows:

[0038] When the index value or band operation value of the monitored target is higher than the index value or band operation value of the non-monitored target:

[0039] By counting the index value or the maximum value of the band operation of the monitoring target in the relatively low radiance area or the non-glare area in the UAV image, the index value or the band operation value in the relatively high radiance area or the glare area is used as the low-pass filter threshold, and the index value or the band operation value above the threshold is changed to the low-pass filter threshold;

[0040] The index value or the minimum value of band operation of non-monitoring targets in relatively low radiance areas or non-glare areas in the UAV image is counted as the high-pass filter threshold of the index value or band operation value in relatively high radiance areas or glare areas, and the index value or band operation value below the threshold is changed to the high-pass filter threshold;

[0041] When the index value or band operation value of the monitored target is lower than the index value or band operation value of the non-monitored target:

[0042] By counting the index value or the minimum value of the band operation of the monitoring target in the relatively low radiance area or the non-glare area in the UAV image, the index value or the band operation value in the relatively high radiance area or the glare area is used as the high-pass filter threshold, and the index value or band operation value below the threshold is changed to the high-pass filter threshold;

[0043] The index value or band operation maximum value of non-monitoring targets in relatively low radiance areas or non-glare areas in UAV images is counted as the low-pass filter threshold of the index value or band operation value in relatively high radiance areas or glare areas, and the index value or band operation value above the threshold is changed to the low-pass filter threshold.

[0044] The present invention has the following beneficial effects and advantages:

[0045] Based on the camera response function, the present invention converts the DN value of the drone RGB image into an energy value (E value) that is linearly related to the object radiance. The E value has a good linear correlation with the exposure time and the scene radiance, so the E value can be used instead of the DN value for quantitative remote sensing. The E value is applied to the calculation of the reflectivity of objects in different light intensities, which can improve the comparability and consistency of drone RGB image data in different light and different exposure times; the E value is applied to the calculation of vegetation, large algae and other indexes in environments with weak signals or high scene radiance, which can improve the accuracy of drone remote sensing monitoring. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 The present invention is a flow chart of the method.

[0047] Figure 2 It is the response function of the R, G, and B bands of the drone RGB camera in the present invention.

[0048] Figure 3 The figure shows the change of vegetation reflectivity at different exposure times in the embodiment of the present invention.

[0049] Figure 4 It is the RGFAH index value of floating macroalgae calculated using the digital quantification value DN in the embodiment of the present invention.

[0050] Figure 5 It is the RGFAH index value of floating macroalgae calculated using the E value in the embodiment of the present invention.

[0051] Figure 6 It is the RGFAH index value of floating macroalgae at the position of the artificial section line in the embodiment of the present invention. DETAILED DESCRIPTION

[0052] The present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.

[0053] like Figure 1 As shown, it is a flow chart of the method of the present invention. The present invention is a UAV remote sensing monitoring method based on camera response function calibration, and its specific implementation method is as follows:

[0054] S1. Control the drone to fly at a fixed position, shoot the monitoring target area with different exposure times, obtain the image sequence of the monitoring target area, and construct the camera response function of the drone camera R, G, and B bands based on the Debevec algorithm;

[0055] Among them, based on the image sequence obtained by changing the exposure time when the drone is at a fixed position, the response function of the red (Red, R), green (Green, G), and blue (Blue, B) bands of the drone's RGB camera is constructed.

[0056] Further, the camera response function of the R, G, and B bands is constructed based on the DN value to E value conversion table, which is specifically obtained according to the Debevec algorithm; including the E value (R, G, and B bands) corresponding to DN values ​​of 0-255, as shown in Table 1. The present invention lists some conversion tables for DN values ​​of 0-9.

[0057] Table 1

[0058] DN 0 1 2 3 4 5 6 7 8 9 E_R -3.91826 -3.81168 -3.70511 -3.60015 -3.49814 -3.40024 -3.30677 -3.21749 -3.13131 -3.04815 E_G -4.00323 -3.89576 -3.78828 -3.68442 -3.58463 -3.48909 -3.39793 -3.31233 -3.23271 -3.15879 E_B -4.01891 -3.93549 -3.85208 -3.77066 -3.69196 -3.61613 -3.54298 -3.4722 -3.40327 -3.33628

[0059] Among them, the drone uses the camera carried by the drone itself for collection, and the drone camera is a CCD camera or a CMOS camera.

[0060] like Figure 2 As shown, the DN value is an integer value from 0 to 255, E_R, E_G, and E_B represent the E values ​​of the R, G, and B bands respectively, and the E value is an energy value linearly related to the radiance of the object.

[0061] Furthermore, the camera response function of the R band is constructed as:

[0062] y=5.542e-11*x 5 -3.763e-08*x 4 +9.692e-06*x 3 -0.001192*x 2 +0.08768*x-3.852

[0064] Furthermore, the camera response function of the G band is constructed as:

[0065] y=3.565e-11*x 5 -2.414e-08*x 4 +6.265e-06*x 3 -0.0008077*x 2 +0.0717*x-3.857

[0067] Furthermore, the camera response function of the B band is constructed as:

[0068] y=3.953e-11*x 5 -2.55e-08*x 4 +6.431e-06*x 3 -0.000836*x 2 +0.07591*x-4.035

[0070] Among them, x is the DN value in the drone RGB image, and y is the E value.

[0071] S2. According to the actual situation of the monitoring target, determine whether there is a stable platform in the drone imaging area as a placement platform for the reference plate, so as to determine whether to place the reference plate in the drone imaging area;

[0072] Wherein, the reference plate is a diffuse reflection reference plate with a known fixed reflectivity;

[0073] In this embodiment, the diffuse reflection standard reference plate adopts the same reference plate developed by Anhui Institute of Optics and Precision Mechanics, Chinese Academy of Sciences. The reference plate is currently made of polytetrafluoroethylene, the most Lambertian material known, and has good diffuse reflection characteristics. The optical properties are similar to those of the high-reflection reference plate. Through the study of the spatial reflection distribution characteristics of the diffuse reflection reference body / target, the reference plate BRF (bidirectional reflectance factor) parameters are provided, and the target reflectance measurement is more accurate and more widely used. The reference plate adopts an aluminum alloy packaging box, which is suitable for field spectral measurement operations. At the same time, the direction-hemispherical reflectance and reflectance factor (BRF) parameters are also provided.

[0074] The spectral range of the reference plate is: 400nm to 2500nm, and the hemispherical reflectance of the reference plate at a wavelength of 550nm is: any one of 2%, 5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, and 99%.

[0075] S3, the drone adjusts the camera lens toward the ground to obtain RGB image data of the monitoring target including vegetation and algae floating on the water surface;

[0076] Furthermore, in step S3, if a reference plate with a known fixed reflectivity is placed in the drone imaging area, the exposure time must be adjusted to prevent oversaturation of the reference plate. The specific adjustment method is: adjust the exposure time so that the DN value of the diffuse reflectance standard reference plate is not higher than 225.

[0077] S4. Split the RGB image into single-band grayscale images according to the R, G, and B bands as input.

[0078] S5, according to the camera response function in step S1, converting the DN value of the R, G, B single-band grayscale image into an energy value linearly related to the object radiance, that is, an E value;

[0079] S6. If a reference plate is placed in the drone imaging area, the E value obtained in S5 is further converted into reflectivity: the nonlinear response of the drone RGB camera is manifested as the DN value increment decreases as the E value increases, resulting in the incomparability of the data obtained when the exposure time and light intensity change, which reduces the credibility of the reflectivity calculation results. Compared with the DN value, the E value has a better linear correlation with the exposure time, so the E value can be used instead of the DN value to calculate the reflectivity, thereby enhancing the comparability of image data with different exposure times and different light intensities.

[0080] Furthermore, in S6, the reflectivity is calculated according to the following formula:

[0081]

[0082] Among them, R(λ) represents the reflectivity of the detection target, L(λ) represents the E value of the detection target, and L p (λ) represents the E value of the reference plate, and ρ(λ) is the reflectivity of the reference plate.

[0083] S7. Index or band calculation: In the nonlinear mapping of the RGB camera, the DN value increment decreases as the radiance value increases. In this embodiment, the Enteromorpha in floating large algae is taken as an example. The radiance of the G band in its pixel is greater than that of the R band and the B band. Then the DN value increment of the G band of the Enteromorpha pixel in the glare area is smaller than that of the R and B bands, resulting in the index method of using band operations to enhance the Enteromorpha signal. The Enteromorpha index value in the glare area is low. In addition, when using drones to obtain high-resolution Enteromorpha green tide images, it is more susceptible to interference from the sun glare on the water surface, which directly affects the quantitative estimation of floating large algae in the glare environment. The E value has a better linear correlation with the scene radiance, which can reduce the impact of nonlinear response on the analysis of vegetation, large algae, etc. in high radiance areas or glare areas.

[0084] S71. If a reference plate is placed in the drone imaging area, the index value or band calculation value of the monitoring target is obtained through the reflectivity;

[0085] S72. If no reference plate is placed in the drone imaging area, the index value or band calculation value of the monitoring target is obtained through the E value;

[0086] Furthermore, in S7, the index value or band operation value of the monitoring target is obtained, that is:

[0087] ZS=a 1 *r 1 +a 2 *r 2 +a 3 *r 3 ;

[0088] Among them, r is the reflectivity or E value, subscripts 1, 2, and 3 represent the R, G, and B bands respectively, and parameter a 1 、a 2 、a 3 is the band operation coefficient value of the monitoring target, a 1 、a 2 、a 3 It is a coefficient for adjusting the operation of the R, G, and B bands. In this embodiment, the default values ​​for Enteromorpha aquatica are 0.2581, 0.7419, and -1, respectively.

[0089] S8, index or band calculation value filtering: remove abnormally high and abnormally low index or band calculation values;

[0090] S81. When the index value or band operation value of the monitored target is higher than the index value or band operation value of the non-monitored target:

[0091] Furthermore, by counting the index value or the maximum value of the band operation of the monitoring target in the relatively low radiance area or the non-glare area in the UAV image, the index value or the band operation value in the relatively high radiance area or the glare area is used as the low-pass filter threshold, and the index value or the band operation value above the threshold is changed to the low-pass filter threshold;

[0092] Furthermore, the index values ​​or band operation minimum values ​​of non-monitoring targets in relatively low radiance areas or non-glare areas in the UAV image are counted as high-pass filter thresholds for the index values ​​or band operation values ​​in relatively high radiance areas or glare areas, and the index values ​​or band operation values ​​below the threshold are changed to the high-pass filter thresholds;

[0093] S82. When the index value or band operation value of the monitored target is lower than the index value or band operation value of the non-monitored target:

[0094] Furthermore, by counting the index value or the minimum value of the band operation of the monitoring target in the relatively low radiance area or the non-glare area in the UAV image, the index value or the band operation value in the relatively high radiance area or the glare area is used as the high-pass filter threshold, and the index value or the band operation value below the threshold is changed to the high-pass filter threshold;

[0095] Furthermore, the index values ​​or the maximum values ​​of band operations of non-monitored targets in relatively low radiance areas or glare-free areas in the UAV image are counted as low-pass filter thresholds for the index values ​​or band operation values ​​in relatively high radiance areas or glare areas, and the index values ​​or band operation values ​​above the threshold are changed to the low-pass filter threshold.

[0096] S9. According to the filtered index value or band operation value, a linear graph is obtained to complete the remote sensing monitoring of the monitoring targets including vegetation and floating algae on the water surface by the UAV, and then carry out biomass calculation and other tasks.

[0097] In order to verify the performance of the method proposed in the present invention, two examples are used to illustrate the calculation of vegetation reflectance in different light intensities and the monitoring of green tide of Enteromorpha based on UAV remote sensing.

[0098] Example 1: Calculation of vegetation reflectance at different light intensities;

[0099] By fixing the object's radiance and changing the exposure time quantitatively while keeping the camera's focal length, aperture, ISO and other parameters unchanged, the imaging results of a fixed exposure time and quantitatively changing the radiance can be simulated. Under clear weather conditions, the drone's flight position is fixed, the camera is set to take images at different exposure times, and a reference plate (gray plate) with a known fixed reflectivity is placed. During the data acquisition process, the imaging interval between data is short, so it can be assumed that the object's radiance and reflectivity remain unchanged. Based on Figure 2 The camera response function shown in FIG. 1 and the lookup table based on the camera response function are used to convert the DN value into the E value. Vegetation samples are selected, and the reflectance of the vegetation samples is calculated using the DN value and the E value respectively. The variation range of the reflectance at different exposure times is compared and analyzed to verify the effectiveness of the method proposed in the present invention. The variation range is based on the minimum value, and the calculation formula is as follows:

[0100] Change range = (band maximum value – band minimum value) / band minimum value

[0101] Figure 3 Columns 1 to 3 show the changes of DN value, reflectivity calculated by DN value, and reflectivity calculated by E value with exposure time. Figure 3 It can be seen that as the exposure time increases from 1 / 8000 second to 1 / 2500 second, the reflectivity result curve calculated by the DN value is more discrete, while the reflectivity curve calculated by the E value is more concentrated. The change amplitudes of the vegetation reflectivity R, G, and B bands calculated by the E value decreased by 45.57%, 64.61%, and 13.22%, respectively. Compared with the reflectivity result calculated by the DN value, the reflectivity result calculated by the E value has the smallest change amplitude, which shows that the method of using the E value to calculate the reflectivity of water bodies is effective, and verifies that the method proposed in the present invention can improve the comparability of drone image data under different exposure times and different light intensities.

[0102] Example 2: Monitoring of Enteromorpha green tide based on UAV remote sensing

[0103] Using a DJI Mavic 2 drone, we acquired RGB image data of Enteromorpha floating on the water surface, and selected typical images containing glare for analysis. Figure 2The camera response function shown in FIG. 1 and the lookup table based on the camera response function are used to convert the DN value into the E value. The DN value and the E value are used to perform index or band calculation respectively. In this embodiment, the values ​​of parameters a1, a2, and a3 are adjustable, and the parameters for Enteromorpha aquatica are set to 0.2581, 0.7419, and -1 respectively. The index or band calculation model degenerates into the RGFAH index. The RGFAH index value is normalized to 0 to 1, and the changes in the RGFAH index values ​​calculated by the two methods under the influence of glare are compared and analyzed. The RGFAH value calculated by the DN value is shown in FIG. Figure 4 As shown, the RGFAH value calculated by E value is as follows Figure 5 shown.

[0104] Depend on Figure 4 and Figure 5 It can be seen that in the glare area, the RGFAH index value calculated by the DN value is lower, which is caused by the nonlinear response of the drone RGB camera. Compared with the RGFAH index value calculated by the DN value, the RGFAH index value of the glare area calculated by the E value is consistent with the size distribution of the non-glare area.

[0105] exist Figure 4 and Figure 5 Make an artificial cut line on the RGFAH value diagram shown in the figure (see the cut line position for details). Figure 4 and Figure 5 The white solid line in the middle) is used to analyze the changes in the RGFAH value at the section line position. The results are as follows Figure 6 As shown. Since the Enteromorpha region has a higher RGFAH value, it appears as a peak on the profile of the RGFAH value, and the position of the peak in the curve in the RGFAH diagram is corresponded by the numbers 1# and 2#. Figure 6 From the solid black line (RGFAH value calculated by DN value), we can see that the peak on the far right (the peak indicated by number 2#) is smaller than the other peaks, that is, the RGFAH value of Enteromorpha in the glare area calculated by DN value is smaller; analysis Figure 6 From the black dotted line in the middle (RGFAH value calculated by E value), we can see that its rightmost peak (the peak indicated by number 2#) is consistent with the other peaks in size, that is, the RGFAH value of Enteromorpha calculated by E value is relatively stable under the influence of glare, and it has a strong ability to resist glare interference.

[0106] In addition, it should be understood that although this specification is described according to implementation methods, not every implementation method contains an independent technical solution. This narrative method of the specification is only for clarity of description. Those skilled in the art should regard this specification as a whole. The technical solutions in each embodiment can also be appropriately combined to form other implementation methods that can be understood by those skilled in the art.

Claims

1. A UAV remote sensing monitoring method based on camera response function calibration, characterized in that: The following steps are involved: S1: Control the drone to fly at a fixed position, shoot the monitoring target area with different exposure times, obtain the image sequence of the monitoring target area, obtain the DN value to E value conversion table based on the Debevec algorithm, and construct the camera response function of the drone camera R, G, B bands; S2: According to the actual situation of the monitoring target, determine whether there is a stable platform in the drone imaging area as a placement platform for the reference plate, so as to determine whether to place the reference plate in the drone imaging area; S3: The drone adjusts the camera lens toward the ground to obtain RGB image data of the monitoring target including vegetation and algae floating on the water surface; S4: Split the RGB image data into R, G, B single-band grayscale images according to the R, G, B bands; S5: According to the camera response function in step S1, the DN value of the R, G, B single-band grayscale image is converted into an energy value linearly related to the object radiance, that is, an E value; S6: Determine whether a reference plate is placed in the current drone imaging area. If a reference plate is placed, convert the E value into reflectivity and then execute step S7; S7: If a reference plate is placed in the drone imaging area, the index value or band calculation value of the monitoring target is obtained through the reflectivity; conversely, if no reference plate is placed in the drone imaging area, the index value or band calculation value of the monitoring target is obtained through the E value; S8: filtering the index value or band operation value of the monitoring target obtained in step S7 to remove abnormal values ​​including those higher than a set threshold and those lower than a set threshold; S9: According to the filtered index value or band operation value, a linear graph is obtained to complete the remote sensing monitoring of the monitoring targets including vegetation and floating algae on the water surface by the UAV.

2. The method for remote sensing monitoring of unmanned aerial vehicles based on camera response function calibration according to claim 1, characterized in that: The drone camera is a CCD camera or a CMOS camera.

3. The method for remote sensing monitoring of unmanned aerial vehicles based on camera response function calibration according to claim 1, characterized in that: The camera response function of the drone camera RGB band is constructed as follows: The camera response function for the R band is constructed as: y=5.542e-11*x 5 -3.763e-08*x 4 +9.692e-06*x 3 -0.001192*x 2 +0.08768*x-3.852 Furthermore, the camera response function of the G band is constructed as: y=3.565e-11*x 5 -2.414e-08*x 4 +6.265e-06*x 3 -0.0008077*x 2 +0.0717*x-3.857 Furthermore, the camera response function of the B band is constructed as: y=3.953e-11*x 5 -2.55e-08*x 4 +6.431e-06*x 3 -0.000836*x 2 +0.07591*x-4.035 Among them, x is the DN value in the drone RGB image, and y is the E value.

4. The method for remote sensing monitoring of unmanned aerial vehicles based on camera response function calibration according to claim 1, characterized in that: The reference plate is a diffuse reflection reference plate with a known fixed reflectivity; The spectral range of the reference plate is 400nm to 2500nm, and the hemispherical reflectance of the reference plate at a wavelength of 550nm is any one of 2%, 5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, and 99%.

5. The method for remote sensing monitoring of unmanned aerial vehicles based on camera response function calibration according to claim 1, characterized in that: If a reference plate is placed in the drone imaging area, the exposure time needs to be adjusted according to the DN value of the reference plate to prevent oversaturation of the reference plate. The specific adjustment method is: adjust the exposure time so that the DN value of the diffuse reflectance standard reference plate is no higher than 225.

6. The method for remote sensing monitoring of unmanned aerial vehicles based on camera response function calibration according to claim 1, characterized in that: In step S6, the E value is converted into reflectivity, specifically: Among them, R(λ) represents the reflectivity of the detection target, L(λ) represents the E value of the detection target, and L p (λ) represents the E value of the reference plate, and ρ(λ) is the reflectivity of the reference plate.

7. The method for remote sensing monitoring of unmanned aerial vehicles based on camera response function calibration according to claim 1, characterized in that: The index value or band operation value of the monitoring target is obtained, that is: ZS=a1*r1+a2*r2+a3*r3; Where r is the reflectivity or E value, subscripts 1, 2, and 3 represent the R, G, and B bands, respectively, and parameters a1, a2, and a3 are the band operation coefficient values ​​of the monitored target.

8. The method for remote sensing monitoring of unmanned aerial vehicles based on camera response function calibration according to claim 1, characterized in that: The step S8 is specifically as follows: When the index value or band operation value of the monitored target is higher than the index value or band operation value of the non-monitored target: By counting the index value or the maximum value of the band operation of the monitoring target in the relatively low radiance area or the non-glare area in the UAV image, the index value or the band operation value in the relatively high radiance area or the glare area is used as the low-pass filter threshold, and the index value or the band operation value above the threshold is changed to the low-pass filter threshold; The index value or the minimum value of band operation of non-monitoring targets in relatively low radiance areas or non-glare areas in the UAV image is counted as the high-pass filter threshold of the index value or band operation value in relatively high radiance areas or glare areas, and the index value or band operation value below the threshold is changed to the high-pass filter threshold; When the index value or band operation value of the monitored target is lower than the index value or band operation value of the non-monitored target: By counting the index value or the minimum value of the band operation of the monitoring target in the relatively low radiance area or the non-glare area in the UAV image, the index value or the band operation value in the relatively high radiance area or the glare area is used as the high-pass filter threshold, and the index value or band operation value below the threshold is changed to the high-pass filter threshold; The index value or band operation maximum value of non-monitoring targets in relatively low radiance areas or non-glare areas in UAV images is counted as the low-pass filter threshold of the index value or band operation value in relatively high radiance areas or glare areas, and the index value or band operation value above the threshold is changed to the low-pass filter threshold.

Images