Multispectral zoom camera system

Abstract

The embodiment of the application provides a multispectral zoom camera system, the multispectral zoom camera system comprises a camera and an ambient light detection device; the camera comprises an image sensor, a disc filter, a variable zoom lens assembly, a motor and a processor; the ambient light detection device is used for acquiring light intensity data of light rays of the monitoring area and sending the light intensity data to the processor, wherein the light intensity data of the light rays comprises light intensity data corresponding to the target spectral band. By applying the embodiment of the application, the determination efficiency of the vegetation index can be improved.

Description

Technical Field

[0001] This application relates to the field of optical technology, and in particular to multispectral zoom camera systems. Background Technology

[0002] Since the reflectance spectrum of vegetation is different from that of the ground, and the reflectance spectrum of vegetation in different states is also different, the vegetation index can be determined by measuring the reflectance of a specified area. The vegetation index can reflect information such as vegetation coverage and vegetation growth status in that area.

[0003] In related technologies, a diffuse reflector with known reflectivity can be set up in advance near a specified area. A camera is used to take an image containing the specified area and the diffuse reflector from a high place. Based on the ratio of the brightness of the specified area in the image to the brightness of the diffuse reflector, the ratio of the reflectivity of the specified area to the reflectivity of the diffuse reflector is determined, thereby obtaining the reflectivity of the specified object.

[0004] However, this method requires a pre-set diffuser, so it is less efficient. Summary of the Invention

[0005] The purpose of this application is to provide a multispectral zoom camera system to improve the efficiency of vegetation index determination. The specific technical solution is as follows:

[0006] This application provides a multispectral zoom camera system, which includes: a camera and an ambient light detection device;

[0007] The camera includes an image sensor, a disc filter, a variable magnification lens assembly, a motor, and a processor;

[0008] The outer edge of the disc filter is provided with a gear, which, under the action of the motor, enables the disc filter to rotate about the optical axis. The disc filter includes at least four light-transmitting areas, and any one of the light-transmitting areas can transmit light of different spectral bands.

[0009] The processor is configured to drive the motor to rotate the disc filter so that light from the target light-transmitting area in the at least four light-transmitting areas located in the monitoring area is incident on the optical path of the image sensor; and to adjust the magnification of the lens assembly to the target magnification; wherein the target light-transmitting area is a light-transmitting area capable of transmitting the target spectral band, and the monitoring area includes vegetation;

[0010] The image sensor is used to sense the light in the monitored area, generate a first spectral image, and send the first spectral image to the processor;

[0011] The ambient light detection device is used to acquire light intensity data of the light in the monitored area and send the light intensity data to the processor, wherein the light intensity data includes light intensity data corresponding to the target spectral band;

[0012] The processor is further configured to receive the first spectral image and the light intensity data of the light; and to preprocess the first spectral image to obtain a second spectral image;

[0013] A target reflectance correction coefficient table corresponding to the target spectral band is determined from multiple preset reflectance correction coefficient tables, wherein the reflectance correction coefficient table is used to represent the correspondence between magnification and reflectance correction coefficient;

[0014] Based on the correspondence shown in the target reflectivity correction coefficient table, determine the target reflectivity correction coefficient corresponding to the target magnification;

[0015] Based on the target reflectance correction coefficient, the second spectral image and the light intensity data of the target spectral band are processed by a first preset function to generate reflectance image data of the monitored area;

[0016] The generated reflectance image data is processed by a second preset function to generate a normalized difference vegetation index image of the monitored area.

[0017] In one possible embodiment, the camera further includes an internal storage medium for storing multiple preset reflectance correction coefficient tables corresponding to multiple spectral bands. Each of the reflectance correction coefficient tables includes multiple sets of magnification and reflectance correction coefficients, wherein the magnification and the reflectance correction coefficients correspond one-to-one.

[0018] In one possible embodiment, in any of the preset reflectivity correction coefficient tables, the reflectivity correction coefficients are determined by adjacent magnifications and reflectivity correction coefficients via a linear interpolation function.

[0019] In one possible embodiment, the linear interpolation function includes:

[0020]

[0021] Where (Ka,Za) and (Kb,Zb) are a set of magnification and reflectance correction coefficients that are adjacent to each other before and after (k,Z).

[0022] In one possible embodiment, the preset reflectivity correction coefficient table includes two sets of specified magnification and reflectivity correction coefficients, and then other sets of magnification and reflectivity correction coefficients are generated via a linear interpolation function.

[0023] In one possible embodiment, the first preset function is defined as follows:

[0024] Where (i,j) represents the image pixel coordinates, B λ Represents the second spectral image, K λ I represents the target reflectivity correction coefficient. light, λ represents the light intensity data corresponding to the target spectral band, P λ This represents reflectance image data.

[0025] In one possible embodiment, the image sensor senses the light in the monitored area to generate a first spectral image, including:

[0026] The light in the monitored area is sensed to generate a first spectral image in RAW format;

[0027] The preprocessing includes:

[0028] The pixels of each color channel in the first spectral image in RAW format are weighted.

[0029] In one possible embodiment, the pixel weighting coefficients of each color channel are different when the target spectral band is different.

[0030] In one possible embodiment, the configuration of the weighting coefficients includes:

[0031] When the center wavelength of the target spectral band is 450 nm, the weight of the blue channel is 1;

[0032] When the center wavelength of the target spectral band is 560 nm, the weight of each green channel is 0.5;

[0033] When the center wavelength of the target spectral band is 850 nm, the weight of each color channel is 0.25.

[0034] In one possible embodiment, the second preset function is defined as follows:

[0035]

[0036] In the formula, (i,j) represents the image pixel coordinates, NDVI represents the normalized difference vegetation index data, and P nir P represents the reflectance image data generated in the near-infrared wavelength range of the target spectral band. red This refers to the reflectance image data generated when the target spectral band is in the red wavelength range.

[0037] Beneficial effects of the embodiments in this application:

[0038] The multispectral zoom camera system provided in this application embodiment converts the sensed light signal into a first spectral image using an image sensor. The pixel values ​​of the first spectral image are obtained by the image sensor sensing the light signal reflected from the monitored area. Simultaneously, because the light from the target's translucent area is incident on the image sensor's optical path under the motor's drive, the light signal sensed by the image sensor is obtained after filtering through the target's translucent area. Since the target's translucent area can pass through the target spectral band, the first spectral image and the second spectral image obtained by preprocessing the first spectral image can reflect the intensity of the target spectral band light reflected from the monitored area to the image sensor. Furthermore, since not all the light reflected from the monitored area illuminates the image sensor, and different proportions of light reflected from the monitored area will be incident on the image sensor due to different magnifications of the lens components, the intensity of the target spectral band light reflected from the target object can be determined by compensating the first spectral image with a target reflectivity correction coefficient determined based on the target magnification. The light reflected from the monitored area is ambient light, and its frequency does not change during reflection. Therefore, the light intensity data of the target spectral band acquired by the ambient light detection device can reflect the intensity of the light incident on the monitored area. According to the definition of reflectivity, the reflectivity of the monitored area can be determined given the known intensity of the incident light and the intensity of the reflected light. Therefore, based on the target reflectivity correction coefficient, processing the second spectral image and the light intensity data of the target spectral band using the first preset function yields the reflectivity image data of the monitored area. Furthermore, based on the reflectivity image data determined under different target spectral bands and the definition of vegetation values, the normalized difference vegetation index image of the monitored area can be determined. Since this process calculates the reflectivity image data based on the definition of reflectivity, there is no need to set up a diffuse reflection plate with known reflectivity for calibration, thus resulting in high efficiency.

[0039] Of course, implementing any product or method of this application does not necessarily require achieving all of the advantages described above at the same time. Attached Figure Description

[0040] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other embodiments can be obtained based on these drawings.

[0041] Figure 1 A schematic diagram of a multispectral zoom camera system provided in an embodiment of this application;

[0042] Figure 2 A schematic diagram of a disk filter in a multispectral zoom camera system provided in an embodiment of this application;

[0043] Figure 3 A schematic diagram illustrating the connection relationship between various components in a camera and an ambient light detection device provided in an embodiment of this application;

[0044] Figure 4 A schematic flowchart illustrating the processing of a first spectral image and light intensity data provided in an embodiment of this application;

[0045] Figure 5a This is a view of a camera in a multispectral zoom camera system provided in an embodiment of this application;

[0046] Figure 5b A cross-sectional view of a camera in a multispectral zoom camera system provided in this application embodiment;

[0047] Figure 6 This is a schematic diagram of an application scenario for the multispectral zoom camera system provided in an embodiment of this application. Detailed Implementation

[0048] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art based on this application are within the scope of protection of this application.

[0049] To provide a clearer explanation of the reflectance determination method provided in this application, the relevant terms used in this document will be explained below:

[0050] Vegetation indices: Based on the spectral characteristics of vegetation, various vegetation indices are calculated by combining image data of vegetation in different bands. Vegetation indices are commonly used in fields such as agricultural remote sensing, land cover, and mineral identification.

[0051] A diffuse reflective plate with known reflectivity: a flat plate with diffuse reflectivity, and the reflectivity of each spectral band on the surface of the plate is known.

[0052] DN (Digital Number) value: The pixel value of an image directly captured by a digital camera, that is, a pixel value that has not been corrected to a unit with specific physical meaning.

[0053] ISP (Image Signal Processing): Image signal processing mainly involves a series of processing steps on the image signals output by the image sensor in the imaging system to obtain image results suitable for human eye observation.

[0054] RAW data: Raw data directly output by the image sensor, without undergoing non-linear processing such as the ISP image processing module. RAW data is usually in Bayer format and consists of four channels: one blue channel, one red channel, and two green channels (hereinafter referred to as green channel 1 and green channel 2, respectively).

[0055] Multispectral light intensity sensor: A sensor capable of sensing and acquiring the light intensity of incident light in multiple wavelength bands.

[0056] Multispectral imaging system: A camera that can acquire image information of an object in different wavelength ranges. Generally, an imaging system that can acquire images of an object in two or more different wavelength spectral ranges is called a multispectral imaging system.

[0057] The following will exemplify one possible application scenario of the multispectral zoom camera system provided in this application. It should be understood that the following example is only one possible application scenario of the multispectral zoom camera system provided in this application. In other possible embodiments, the multispectral zoom camera system provided in this application can also be applied to other possible application scenarios. The following example does not limit this in any way.

[0058] Crops are often planted over large areas, making it inefficient to determine their growth and pest / disease status through manual inspections. One relevant technology utilizes multispectral imaging systems to photograph crop areas, obtaining images of crops at different wavelengths. Based on these images, the reflectance of the crop area to different wavelengths of light is determined, resulting in a vegetation index. Because crops exhibit different reflectance spectra at different growth stages, differ from the ground's reflectance spectrum, and vary between healthy and diseased crops, the vegetation index can be used to determine crop coverage, growth status, and pest / disease status, effectively improving crop management efficiency.

[0059] Method 1:

[0060] A diffuse reflector with known reflectivity is pre-placed within the crop area. For each wavelength band of the crop image, the brightness of the crop area and the brightness of the diffuse reflector are determined from the crop image. The ratio of the crop area's brightness to the diffuse reflector's brightness, multiplied by the diffuse reflector's reflectivity for that wavelength band, gives the crop area's reflectivity for that wavelength band. This allows for the calculation of the crop area's reflectivity for different wavelength bands, thus enabling the calculation of the crop area's vegetation index.

[0061] Method 2:

[0062] For each band of crop image, the reflectance of the crop area to that band of light is directly calculated based on the pixel value of the pixel in the crop image, and the vegetation index of the crop area is calculated based on the calculated reflectance.

[0063] Method 1 requires pre-setting diffuse reflectors with known reflectivity within the crop area. When the crop area is large, this requires significant time and manpower, resulting in low efficiency. Method 2, however, relies heavily on DN values ​​(density numbers) for each pixel in the crop image, which are obtained through image ISP (Image Signal Processing). Different multispectral imaging systems employ different ISP methods; some systems significantly compensate for brightness during ISP, while others only slightly. Therefore, the calculation results of Method 2 are heavily influenced by the ISP process of the multispectral imaging system, making it difficult to calculate accurate reflectivity and consequently leading to inaccurate vegetation indices.

[0064] Based on this, this application provides a multispectral zoom camera system, such as... Figure 1 As shown, it includes a camera 100 and an ambient light detection device 200.

[0065] The camera 100 includes an image sensor, a disc filter, a variable lens assembly, a motor, and a processor.

[0066] A gear is circumferentially arranged on the outer edge of the disc filter. The gear is driven by a motor to enable the disc filter to rotate about the optical axis. The disc filter includes at least four light-transmitting areas, and any light-transmitting area can transmit light of different spectral bands.

[0067] The processor drives the motor to rotate the disc filter so that the light from the target light-transmitting area in at least four light-transmitting areas is located in the monitoring area and enters the optical path of the image sensor; and adjusts the magnification of the lens assembly to the target magnification; wherein, the target light-transmitting area is a light-transmitting area that can transmit the target spectral band, and the monitoring area includes vegetation;

[0068] An image sensor is used to sense the light in the monitored area, generate a first spectral image, and send the first spectral image to the processor;

[0069] An ambient light detection device 200 is used to acquire light intensity data of the light in the monitored area and send the light intensity data to a processor. The light intensity data includes light intensity data corresponding to the target spectral band.

[0070] The processor is also used to receive the first spectral image and light intensity data; and to preprocess the first spectral image to obtain the second spectral image;

[0071] A target reflectance correction coefficient table corresponding to the target spectral band is determined from multiple preset reflectance correction coefficient tables. The reflectance correction coefficient table is used to show the correspondence between magnification and reflectance correction coefficient.

[0072] Based on the correspondence shown in the target reflectivity correction coefficient table, determine the target reflectivity correction coefficient corresponding to the target magnification.

[0073] Based on the target reflectance correction coefficient, the light intensity data of the second spectral image and the target spectral band are processed by a first preset function to generate reflectance image data of the monitored area.

[0074] The generated reflectance image data is processed by a second preset function to generate a normalized difference vegetation index image of the monitored area.

[0075] In this embodiment, since the image sensor in the multispectral zoom camera system is used to convert the sensed light signal into a first spectral image, the pixel value of the first spectral image is obtained by the image sensor sensing the light signal reflected from the monitoring area. At the same time, since the light from the target's transparent area located in the monitoring area is incident on the optical path of the image sensor under the drive of the motor, the light signal sensed by the image sensor is the light signal obtained after being filtered by the target's transparent area. Since the target's transparent area can pass through the target spectral band, the first spectral image and the second spectral image obtained by preprocessing the first spectral image can reflect the intensity of the target spectral band light in the light reflected from the monitoring area to the image sensor. Furthermore, since not all the light reflected from the monitoring area illuminates the image sensor, and the magnification of the lens components is different, different proportions of the light reflected from the monitoring area will be incident on the image sensor. Therefore, by compensating the first spectral image with the target reflectivity correction coefficient determined based on the target magnification, the intensity of the target band light in the light reflected from the target object can be determined. The light reflected from the monitored area is ambient light, and its frequency does not change during reflection. Therefore, the light intensity data of the target spectral band acquired by the ambient light detection device can reflect the intensity of the light incident on the monitored area. According to the definition of reflectivity, the reflectivity of the monitored area can be determined given the known intensity of the incident light and the intensity of the reflected light. Therefore, based on the target reflectivity correction coefficient, processing the second spectral image and the light intensity data of the target spectral band using the first preset function yields the reflectivity image data of the monitored area. Furthermore, based on the reflectivity image data determined under different target spectral bands and the definition of vegetation values, the normalized difference vegetation index image of the monitored area can be determined. Since this process calculates the reflectivity image data based on the definition of reflectivity, there is no need to set up a diffuse reflection plate with known reflectivity for calibration, thus resulting in high efficiency. This solves the problem mentioned in Method 1.

[0076] On the other hand, the multispectral zoom camera system provided in this application obtains the vegetation index based on the spectral image sensed by the image sensor. Unlike the DN value, the spectral image sensed by the image sensor is RAW data without ISP. Therefore, the pixel value of the pixel in the spectral image sensed by the image sensor is not affected by ISP, and can reflect the true brightness of the light reflected in the monitored area relatively accurately. Therefore, compared with the aforementioned method two, the vegetation index determined by the multispectral zoom camera system provided in this application is more accurate.

[0077] The camera 100 and the ambient light detection device 200 will be described separately below:

[0078] See Figure 2 , Figure 2The diagram shows a schematic of a disc filter in a multispectral zoom camera system provided in this application, including: a gear 111 and a light-transmitting area 112.

[0079] Figure 2 The disc filter shown includes four light-transmitting areas 112. In other possible embodiments, the disc filter may also include five or more light-transmitting areas 112. Each light-transmitting area 112 can transmit light of a specific spectral band, and the spectral bands of light transmitted by different light-transmitting areas 112 are different.

[0080] For example, in one possible embodiment, the four light-transmitting areas 112 are respectively capable of transmitting light with a center wavelength of 450nm, a center wavelength of 560nm, a center wavelength of 660nm, and a center wavelength of 850nm. In another possible embodiment, the four light-transmitting areas 112 are respectively capable of transmitting light with a center wavelength of 450nm, a center wavelength of 560nm, a center wavelength of 720nm, and a center wavelength of 850nm. In yet another possible embodiment, the four light-transmitting areas 112 are respectively capable of transmitting light with a center wavelength of 450nm, a center wavelength of 560nm, a center wavelength of 750nm, and a center wavelength of 850nm.

[0081] Gear 111 meshes with the transmission component of the motor, enabling the motor to drive the disc filter to rotate around the optical axis via gear 111. As the disc filter rotates, the positions of the light-transmitting areas 112 also change, so that the light from the different light-transmitting areas 112 located in the monitoring area is incident on the optical path of the image sensor.

[0082] Because different light-transmitting areas 112 can transmit light with different spectral bands, when light from different light-transmitting areas 112 located in the monitoring area is incident on the optical path of the image sensor, light with different spectral bands in the monitoring area can pass through the light-transmitting area 112 and enter the image sensor. Furthermore, at most one light-transmitting area 112 located in the monitoring area can transmit light onto the optical path of the image sensor at any given time. Therefore, by driving the rotation of the disc filter via a drive motor, the image sensor can generate spectral images with different spectral bands.

[0083] The monitoring area in this article is the area where the vegetation index is to be determined. For example, it can be a crop area as in the previous example, or a forest area, etc., and the monitoring area should be within the field of view of the camera.

[0084] The connection relationships between the components of camera 100 and ambient light detection device 200 can be as follows: Figure 3As shown, the disc filter 110 is meshed with the motor 140, the motor 140 is electrically connected to the processor 150, the image sensor 120 is electrically connected to the processor 150, the lens assembly 130 is electrically connected to the processor 150, and the ambient light detection device 200 is electrically connected to the processor 150.

[0085] The processor 150 can send electrical signals to the motor 140 to drive the disc filter 110 to rotate, so that light from a specific light-transmitting area 112 on the disc filter 110 located in the monitoring area is incident on the optical path of the image sensor 120. Furthermore, the processor 150 can send electrical signals to the lens assembly 130 to drive the lens assembly 130 to adjust to a specific magnification. Here, the specific light-transmitting area 112 is referred to as the target light-transmitting area, and the specific magnification is referred to as the target magnification. The target light-transmitting area should be able to transmit light in the target spectral band. The target spectral band depends on the spectral bands involved in calculating the vegetation index. For example, the calculation of the normalized difference vegetation index involves the near-infrared wavelength range and the red wavelength range; therefore, the target spectral band must be at least two spectral bands, with at least one target spectral band located in the near-infrared wavelength range and at least one target spectral band located in the red wavelength range.

[0086] Image sensor 120 sends a first spectral image to processor 150 via an electrical connection. This first spectral image is a spectral image generated by image sensor 120 sensing light in the monitored area. Since the first spectral image has not undergone ISP processing, it can be a RAW format image as described above. Furthermore, ambient light detection device 200 sends light intensity data of the monitored area to processor 150 via an electrical connection. This light intensity data includes at least the light intensity data corresponding to the target spectral band.

[0087] The following will explain how processor 150 processes the first spectral image and light intensity data. The processing can be divided into the following steps: Figure 4 The steps shown include:

[0088] S401, preprocess the first spectral image to obtain the second spectral image.

[0089] Preprocessing may consist of only weighting the pixels of each color channel (hereinafter referred to as channel weighting), or it may include weighting the pixels of each color channel as well as dark noise correction and scaling to a preset resolution (hereinafter referred to as scaling).

[0090] Regarding dark noise correction, this application does not impose any limitations on the method of dark noise correction, and any dark noise correction method can be used. It is understood that the light signal sensed by the image sensor may include some noise light signals in addition to the light reflected from the monitoring area. These noise light signals may cause the first spectral image obtained to fail to accurately reflect the brightness of the light reflected from the monitoring area, thereby resulting in an inaccurate vegetation index. Therefore, the influence of these dark noise light signals on the vegetation index can be reduced or even eliminated by dark noise correction, thereby improving the accuracy of the vegetation index obtained.

[0091] For scaling, the first spectral image can be scaled to a preset size, such as 800*600, to facilitate subsequent unified processing.

[0092] For channel weighting, the pixel value of each pixel in the processed image obtained through channel weighting is the brightness. In one possible example, we first perform dark noise correction on the first spectral image, denoting the resulting image data as Data1, and then denoting the blue channel data of Data1 as Data... 1,blue The data for the two green channels are respectively denoted as Data. 1,green1 Data 1,green2 The data in the red channel is denoted as Data. 1,red The channel weighting method can be performed according to formula (1):

[0093] Data2=β blue *Data 1,blue +β green1 *Data 1,green1 +β green2 *Data 1,green2 +β red *Data 1,red …Formula (1)

[0094] Where Data2 is the image data obtained through channel weighting, β blue β green1 β green2 β red This refers to the pre-set weights for each color channel. These weights can be set based on user experience and / or actual needs, or they can be set based on the signal-to-noise ratio (SNR) of each color channel, with higher SNR channels receiving higher weights. For example, if the SNR of the red channel > the SNR of green channel 1 > the SNR of green channel 2 > the SNR of the blue channel, then β... red >β green1 >β green2 >β blue .

[0095] By using this embodiment, the influence of noise light signals in the weighted results can be reduced or even eliminated as much as possible through channel weighting, thereby obtaining a more accurate vegetation index based on the weighted results.

[0096] In another possible embodiment, channel weighting is performed directly without dark noise correction of the first spectral image. The channel weighting can be performed according to formula (2), and the pixel values ​​of the two green channels in the first spectral image are denoted as Data. 0,green1 Data 0,green2 The pixel value of the blue channel is denoted as Data. 0,blue The pixel value of the red channel is denoted as Data. 0,red ,:

[0097] Light = α blue *Data 0,blue +α green1 *Data 0,green1 +α green2 *Data 0,green2 +α red *Data 0,red …Formula (2)

[0098] Where, α blue α green1 α green2 and α red Preset weights.

[0099] It is understandable that the signal-to-noise ratio (SNR) of the data from each channel sensed by the image sensor is related to the target spectral band. For example, if the target spectral band is the first spectral band, the blue channel in the first spectral image sensed by the image sensor has a higher SNR, while the other color channels have a lower SNR. However, if the target spectral band is the second spectral band, the red channel in the original image data sensed by the image sensor has a higher SNR, while the other color channels have a lower SNR.

[0100] Therefore, in one possible embodiment, the pixel weighting coefficients for each color channel can be different depending on the target spectral band. The configuration of the weighting coefficients can be as shown in Table 1:

[0101] 450nm 560nm 660, 720, 750nm 850nm Blue Channel 1 0 0 0.25 Green Channel 1 0 0.5 0 0.25 Green Channel 2 0 0.5 0 0.25 Red Channel 0 0 1 0.25

[0102] Table 1 shows the mapping relationship between bands and weights for different color channels.

[0103] The 1 in the second row and second column of Table 1 indicates that for the blue channel, the weight corresponding to the 450nm band is 1. Taking the aforementioned formula (1) as an example, that is, when the center wavelength of the target spectral band is 450nm, β in formula (1) is...blue =1. The 0 in the third row and second column of Table 1 indicates that for green channel 1, the weight corresponding to the 450nm band is 0. Taking the aforementioned formula (1) as an example, that is, when the center wavelength of the target spectral band is 450nm, β in formula (1) is 0. green1 =0, and so on. It is understood that Table 1 is only an example of the correspondence provided in this application. In other possible embodiments, the division of bands, the number of color channels, and the weights corresponding to each band can be different. Table 1 does not impose any restrictions on this.

[0104] For situations requiring multiple preprocessing steps, the order of each preprocessing step can be set according to actual needs. For example, assuming the preprocessing includes dark noise correction, channel weighting, and scaling, the first spectral image can be dark noise corrected first, then the dark noise correction result can be channel weighted, and then the channel weighted result can be scaled. Alternatively, the first spectral image can be scaled first, the scaled result can be dark noise corrected, and then the dark noise correction result can be channel weighted. Other orders are also possible, and this application does not impose any restrictions on them.

[0105] S402, determine the target reflectance correction coefficient based on the target magnification and target spectral band.

[0106] In this application, the camera 100 stores multiple reflectance correction coefficient tables, each of which records a magnification and a reflectance correction coefficient. For example, it can be shown in Table 2:

[0107] Magnification Z1 Z2 ... Zn Reflectivity correction coefficient k1 k2 … kn

[0108] Table 2. Correspondence between magnification and reflectance correction coefficient

[0109] The second column in Table 2 indicates that the reflectance correction factor corresponding to the magnification Z1 is k1, the third column indicates that the reflectance correction factor corresponding to the magnification Z1 is k2, and so on.

[0110] Each reflectance correction coefficient table corresponds to a spectral band, and different reflectance correction coefficient tables correspond to different spectral bands. When determining the target reflectance correction coefficient, firstly, the target reflectance correction coefficient table corresponding to the target spectral band is determined from multiple reflectance correction coefficient tables. Then, the reflectance correction coefficient corresponding to the target magnification is determined from the target reflectance correction coefficient table. This reflectance correction coefficient is the target reflectance correction coefficient.

[0111] For ease of description, assume that the target reflectivity correction coefficient table is as shown in Table 2 above, and the target magnification is Z1, then the target reflectivity correction coefficient is k1.

[0112] It is understandable that the target magnification may be different in different application scenarios. As shown in Table 2 above, the table can only represent the correspondence between discrete magnification and reflectivity correction coefficient. If the correspondence is to include reflectivity correction coefficients corresponding to all magnifications, then a large number of magnifications and corresponding reflectivity correction coefficients need to be recorded in the reflectivity correction coefficient table, which will consume a lot of system resources.

[0113] Based on this, in one possible embodiment, multiple standard magnifications and their corresponding reflectance correction coefficients are recorded only in the reflectance correction coefficient table. When the target magnification is a standard magnification, the reflectance correction coefficient corresponding to the standard magnification equal to the target magnification is determined according to the reflectance correction coefficient table and used as the target reflectance correction coefficient. When the target magnification is not a standard magnification, the reflectance correction coefficients corresponding to each standard magnification are interpolated to obtain the reflectance correction coefficient corresponding to the target magnification.

[0114] The interpolation method can vary depending on the application scenario. For example, it can be linear interpolation or nonlinear interpolation. For instance, in one possible embodiment, linear interpolation can be performed according to formula (3) to obtain the target reflectivity correction coefficient:

[0115]

[0116] Where k is the target reflectance correction coefficient, Z is the target magnification, Za is the magnification preceding the target magnification in the reflectance correction coefficient table, ka is the reflectance correction coefficient corresponding to Za, Zb is the magnification following the target magnification in the reflectance correction coefficient table, and kb is the reflectance correction coefficient corresponding to Zb. Taking Table 2 as an example, assuming Z1 < Z2 < ... < Zn, and the target magnification is between Z2 and Z3, then Za is Z2, and Zb is Z3.

[0117] S403, based on the target reflectance correction coefficient, performs the first preset function processing on the light intensity data of the second spectral image and the target spectral band to generate reflectance image data of the monitored area.

[0118] According to the definition of reflectivity, if the brightness of the reflected light is denoted as I... c The incident light intensity is denoted as I. e The reflectance can then be calculated using formula (4):

[0119]

[0120] As analyzed above, the pixel value of the second spectral image is the brightness, and this brightness, after being corrected by the target reflectivity correction coefficient, can be regarded as the brightness of the light reflected from the monitoring area. Taking the case where the correction method is to multiply the pixel value of the second spectral image by the target reflectivity correction coefficient as an example, and considering that the reflectivity is different at different locations in the monitoring area, formula (4) can be rewritten as formula (5):

[0121]

[0122] Where k is the aforementioned target compensation coefficient, B λ Let (i, j) represent the second spectral image, and B represent the pixel coordinates of the image. λ (i, j) represents the pixel value of the pixel located at image coordinates (i, j) in the second spectral image, P λ P represents reflectance image data. λ (i, j) represents the reflectance at the spatial location corresponding to the image coordinates (i, j) in the reflectance image data.

[0123] As explained above, the light incident on the monitoring area can be considered as the light intensity of the target spectral band, therefore formula (5) can be rewritten as formula (6):

[0124]

[0125] Among them, I light, λ represents the light intensity data corresponding to the target spectral band. That is, the reflectance image data can be calculated according to formula (6). In other words, the aforementioned first preset function can be limited to the aforementioned formula (6).

[0126] S404, the generated reflectance image data is processed by the second preset function to generate a normalized difference vegetation index image of the monitored area.

[0127] The second preset function is formula (7):

[0128]

[0129] Where (I,j) represents the image pixel coordinates, NDVI represents the normalized difference vegetation index data, NDVI(I,j) represents the normalized difference vegetation index at the spatial location corresponding to pixel (I,j), and P nir P represents the reflectance image data generated in the near-infrared wavelength range of the target spectral band. red This represents reflectance image data generated when the target spectral band is in the red wavelength range.

[0130] See Figure 5a and Figure 5b , Figure 5a The image shown is a view of a camera in the multispectral zoom camera system provided in this application. Figure 5b This is a cross-sectional view of a camera in the multispectral zoom camera system provided in this application.

[0131] See Figure 6 , Figure 6 The diagram illustrates an application scenario of the multispectral zoom camera system provided in this application. In this example, the multispectral zoom camera system is mounted on a high frame to position it at a high altitude, capable of capturing images of the monitored area. "Higher" in this context refers to a height exceeding a preset height threshold. By adjusting the multispectral zoom camera system so that the light inlet of the lens assembly is aligned with the monitored area and the light inlet of the ambient light detection device is aligned with the sky, the vegetation index of the monitored area can be measured.

[0132] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a solid-state drive (SSD), etc.

[0133] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0134] The various embodiments in this specification are described in a related manner. For the same or similar parts between the various embodiments, please refer to each other. Each embodiment focuses on describing the differences from other embodiments.

[0135] The above description is merely a preferred embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application are included within the scope of protection of this application.

Claims

1. A multispectral zoom camera system, characterized in that, The multispectral zoom camera system includes: a camera and an ambient light detection device; The camera includes an image sensor, a disc filter, a variable magnification lens assembly, a motor, and a processor; The outer edge of the disc filter is provided with a gear, which, under the action of the motor, enables the disc filter to rotate about the optical axis. The disc filter includes at least four light-transmitting areas, and any one of the light-transmitting areas can transmit light of different spectral bands. The processor is configured to drive the motor to rotate the disc filter so that light from the target light-transmitting area in the at least four light-transmitting areas located in the monitoring area is incident on the optical path of the image sensor; and to adjust the magnification of the lens assembly to the target magnification; wherein the target light-transmitting area is a light-transmitting area capable of transmitting the target spectral band, and the monitoring area includes vegetation; The image sensor is used to sense the light in the monitored area, generate a first spectral image, and send the first spectral image to the processor; The ambient light detection device is used to acquire light intensity data of the light in the monitored area and send the light intensity data to the processor, wherein the light intensity data includes light intensity data corresponding to the target spectral band; The processor is further configured to receive the first spectral image and the light intensity data of the light; and to preprocess the first spectral image to obtain a second spectral image; the preprocessing includes: weighting the pixels of each color channel in the first spectral image; A target reflectance correction coefficient table corresponding to the target spectral band is determined from multiple preset reflectance correction coefficient tables, wherein the reflectance correction coefficient table is used to represent the correspondence between magnification and reflectance correction coefficient; Based on the correspondence shown in the target reflectivity correction coefficient table, determine the target reflectivity correction coefficient corresponding to the target magnification; Based on the target reflectance correction coefficient, the second spectral image and the light intensity data of the target spectral band are processed by a first preset function to generate reflectance image data of the monitored area; The generated reflectance image data is processed by a second preset function to generate a normalized difference vegetation index image of the monitored area. Wherein, the first preset function is limited to: ,in, Represents the pixel coordinates of the image. This represents the second spectral image. This represents the target reflectivity correction coefficient. This represents the light intensity data corresponding to the target spectral band. This represents reflectance image data.

2. The multispectral zoom camera system according to claim 1, wherein, The camera also includes an internal storage medium for storing multiple preset reflectance correction coefficient tables corresponding to multiple spectral bands. Each of the reflectance correction coefficient tables includes multiple sets of magnification and reflectance correction coefficients, wherein the magnification and the reflectance correction coefficients correspond one-to-one.

3. The multispectral zoom camera system according to claim 2, wherein, In any of the preset reflectivity correction coefficient tables, the reflectivity correction coefficients are determined by adjacent magnifications and reflectivity correction coefficients via a linear interpolation function.

4. The multispectral zoom camera system according to claim 3, wherein, The linear interpolation function includes: ; Where (Ka,Za) and (Kb,Zb) are a set of magnification and reflectance correction coefficients that are adjacent to each other before and after (k,Z).

5. The multispectral zoom camera system according to claim 4, wherein, The preset reflectivity correction coefficient table includes two sets of specified magnification and reflectivity correction coefficients, and then other sets of magnification and reflectivity correction coefficients are generated through a linear interpolation function.

6. The multispectral zoom camera system according to claim 1, wherein the image sensor senses the light in the monitored area to generate a first spectral image, comprising: The light in the monitored area is sensed to generate a first spectral image in RAW format; The preprocessing includes: The pixels of each color channel in the first spectral image in RAW format are weighted.

7. The multispectral zoom camera system according to claim 6, wherein, The pixel weighting coefficients for each color channel are different depending on the target spectral band.

8. The multispectral zoom camera system according to claim 7, wherein, The configuration of the weighting coefficients includes: When the center wavelength of the target spectral band is 450 nm, the weight of the blue channel is 1; When the center wavelength of the target spectral band is 560 nm, the weight of each green channel is 0.5; When the center wavelength of the target spectral band is 850 nm, the weight of each color channel is 0.

25.

9. The multispectral zoom camera system according to claim 1, wherein, The second preset function is limited to: ; In the formula, Represents the pixel coordinates of the image. This represents the normalized difference vegetation index data. This represents the reflectance image data generated in the near-infrared wavelength range for the target spectral band. This refers to the reflectance image data generated when the target spectral band is in the red wavelength range.

Images