Sweet potato tuber growth status monitoring system based on spectral imaging analysis

Abstract

This invention relates to the field of data processing technology, and in particular to a sweet potato tuber growth status monitoring system based on spectral imaging analysis. The system includes a data acquisition module for acquiring spectral signal curves of target sweet potato tubers and smoothing the spectral signal curves to obtain target spectral signal curves; a curve segmentation module for acquiring at least two consecutive target spectral signal curves, mapping all target spectral signal curves onto the same two-dimensional plane, and dividing all target spectral signal curves in the two-dimensional plane into at least two sets of spectral signal sub-curves; a curve correction module for acquiring the degree of soil reflectance anomaly in the sets of spectral signal sub-curves and correcting the reflectance intensity of each spectral signal on each target spectral signal curve; and a status monitoring module for monitoring the growth status of target sweet potato tubers based on each corrected target spectral signal curve, thereby improving the accuracy of sweet potato tuber growth status monitoring.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a sweet potato tuber growth status monitoring system based on spectral imaging analysis. Background Technology

[0002] Water supply is a crucial factor influencing sweet potato growth and tuber shape development. Managing water supply can control sweet potato growth, leading to better tuber shape. Therefore, monitoring the growth status of sweet potatoes during their growth process and implementing different water supply strategies is of great significance. Currently, spectral imaging technology is used to monitor the growth status of sweet potatoes: infrared light is emitted from sweet potato tubers to obtain spectral signal curves between 1400nm and 1900nm. These spectral signal curves are compared with standard spectral signal curves during normal sweet potato growth to determine the degree of peaks. Significant differences between the spectral signal curves and standard spectral signal curves indicate abnormal water absorption function in the sweet potato tubers, suggesting a risk of poor growth. This method allows for rapid and non-destructive monitoring of sweet potato growth and health, helping to optimize water management strategies and improve tuber quality.

[0003] However, when using spectral imaging technology to monitor the growth status of sweet potato tubers, the reflectance spectral signal is easily affected by the soil moisture and particle structure above the sweet potato tubers, resulting in inaccurate spectral signals collected from the sweet potato tubers. Consequently, the results of judging the growth status of sweet potatoes based on spectral signals are also biased.

[0004] Therefore, ensuring the accuracy of spectral signal acquisition of sweet potato tubers in order to reduce the error in monitoring the growth status of sweet potato tubers has become an urgent problem to be solved. Summary of the Invention

[0005] In view of this, embodiments of the present invention provide a sweet potato tuber growth status monitoring system based on spectral imaging analysis to solve the problem of how to ensure the accuracy of spectral signal acquisition of sweet potato tubers, so as to reduce the error in monitoring the growth status of sweet potato tubers.

[0006] This invention provides a sweet potato tuber growth status monitoring system based on spectral imaging analysis, the system comprising:

[0007] The data acquisition module is used to acquire the infrared spectral signal of the target sweet potato tuber within a preset period and construct the corresponding spectral signal curve. The horizontal axis of the spectral signal curve is the wavelength, and the vertical axis is the reflection intensity. Based on the peaks and troughs in the spectral signal curve, the spectral signal curve is smoothed to obtain the target spectral signal curve.

[0008] The curve division module is used to acquire target spectral signal curves within at least two consecutive preset periods, map all target spectral signal curves onto the same two-dimensional plane, where the horizontal axis of the two-dimensional plane is wavelength and the vertical axis is reflection intensity, and divide all target spectral signal curves in the two-dimensional plane into at least two sets of spectral signal sub-curves based on the inflection point of each target spectral signal curve in the two-dimensional plane.

[0009] The curve correction module is used to obtain the degree of soil reflectance anomaly of the corresponding spectral signal sub-curve set based on the difference between each spectral signal sub-curve in each spectral signal sub-curve set, and to correct the reflection intensity of each spectral signal on each target spectral signal curve according to the degree of soil reflectance anomaly of each spectral signal sub-curve set, so as to obtain the corrected target spectral signal curve.

[0010] The status monitoring module is used to monitor the growth status of the target sweet potato tuber based on each of the corrected target spectral signal curves.

[0011] Preferably, the data acquisition module smooths the spectral signal curve based on the peaks and troughs in the spectral signal curve to obtain the target spectral signal curve, including:

[0012] Obtain the peak and trough points in the spectral signal curve. For any one of the peak and trough points, obtain the scattering anomaly degree of the point based on the adjacent data points on the spectral signal curve.

[0013] If the scattering anomaly at any point is greater than or equal to a preset anomaly threshold, then the average reflection intensity between the left and right adjacent data points of any point is obtained, and the average reflection intensity is used to replace the reflection intensity of any point; if the scattering anomaly at any point is less than the preset anomaly threshold, then the reflection intensity of any point remains unchanged.

[0014] By traversing each of the stated peaks and troughs, the target spectral signal curve is obtained.

[0015] Preferably, obtaining the scattering anomaly degree of any point based on adjacent data points on the spectral signal curve includes:

[0016] Calculate the absolute value of the difference in reflection intensity between any point and its left and right adjacent data points respectively, obtain the average absolute value of the difference, and normalize the average absolute value of the difference to obtain the local difference degree of any point.

[0017] On the spectral signal curve, obtain the left adjacent peak or left adjacent trough of any point, count the number of left interval data points between any point and the left adjacent peak or left adjacent trough, obtain the number of right interval data points between any point and the right adjacent peak or right adjacent trough, and use a preset exponential function to inversely normalize the sum of the number of left interval data points and the number of right interval data points to obtain the fluctuation continuity of any point;

[0018] The degree of scattering anomaly at any given point is obtained based on the degree of local difference at that point and the degree of fluctuation duration.

[0019] Preferably, the curve division module divides all target spectral signal curves in the two-dimensional plane into at least two sets of spectral signal sub-curves based on the inflection points of each target spectral signal curve in the two-dimensional plane, including:

[0020] Cluster the inflection points of all target spectral signal curves in the two-dimensional plane to obtain at least two inflection point clusters. Obtain the centroid of each inflection point cluster. Based on the position of each centroid in the two-dimensional plane, draw a vertical line through each centroid. Use the vertical line to divide all target spectral signal curves in the two-dimensional plane into at least two sets of spectral signal sub-curves.

[0021] Preferably, the curve correction module obtains the degree of soil reflectance anomaly of the corresponding spectral signal sub-curve set based on the differences between the various spectral signal sub-curves in each spectral signal sub-curve set, including:

[0022] For any set of spectral signal sub-curves, obtain the wavelength range corresponding to the set of spectral signal sub-curves, take any wavelength in the wavelength range as the target wavelength, calculate the absolute value of the difference in reflection intensity between every two adjacent spectral signal sub-curves in the set of spectral signal sub-curves at the target wavelength, and obtain the sum of the absolute values ​​of the differences.

[0023] The sum of the absolute values ​​of the differences corresponding to each wavelength in the wavelength range is obtained, and the mean of the sum of the absolute values ​​of the differences is used as the degree of signal change difference of any set of spectral signal sub-curves.

[0024] The degree of signal variation difference for each of the spectral signal sub-curve sets is obtained, and the degree of soil reflectance anomaly for any spectral signal sub-curve set is obtained based on the degree of signal variation difference for each of the spectral signal sub-curve sets.

[0025] Preferably, obtaining the soil reflectance anomaly degree of any spectral signal sub-curve set based on the degree of signal change difference of each spectral signal sub-curve set includes:

[0026] Calculate the average value of the differences in signal variation among all spectral signal sub-curve sets. Calculate the absolute value of the difference between the difference in signal variation of any spectral signal sub-curve set and the average value, and record it as the difference feature value. Normalize the product of the difference feature value and the difference in signal variation of any spectral signal sub-curve set to obtain the degree of soil reflectance anomaly of any spectral signal sub-curve set.

[0027] Preferably, the curve correction module corrects the reflection intensity of each spectral signal on each target spectral signal curve according to the degree of soil reflectance anomaly in each set of spectral signal sub-curves, to obtain the corrected target spectral signal curve, including:

[0028] The degree of soil reflection anomaly in each set of spectral signal sub-curves is compared with a preset threshold for soil reflection anomaly. If the degree of soil reflection anomaly in any set of spectral signal sub-curves is less than the threshold for soil reflection anomaly, the reflection intensity of each spectral signal on each spectral signal sub-curve in each set of spectral signal sub-curves remains unchanged.

[0029] If the soil reflectance anomaly degree of any set of spectral signal sub-curves is greater than or equal to the soil reflectance anomaly degree threshold, then the reflectance intensity of each spectral signal on each spectral signal sub-curve in the set of any set of spectral signal sub-curves is corrected to obtain the corrected set of spectral signal sub-curves.

[0030] Using the spectral signal sub-curves in each of the modified spectral signal sub-curve sets, the corresponding spectral signal sub-curves in each of the target spectral signal curves are replaced to obtain each modified target spectral signal curve.

[0031] Preferably, the step of correcting the reflection intensity of each spectral signal on each spectral signal sub-curve in the arbitrary spectral signal sub-curve set to obtain the corrected spectral signal sub-curve set includes:

[0032] For any spectral signal on each spectral signal sub-curve in the arbitrary spectral signal sub-curve set, obtain the product between a preset multiple and the degree of soil reflectance anomaly in the arbitrary spectral signal sub-curve set. Use the sum of the constant 1 and the product as the correction coefficient of the arbitrary spectral signal. Based on the product of the correction coefficient and the reflectance intensity of the arbitrary spectral signal, obtain the corrected reflectance intensity of the arbitrary spectral signal.

[0033] The beneficial effects of the embodiments of the present invention compared with the prior art are as follows:

[0034] This invention analyzes the signal fluctuation characteristics of the spectral signal curves of the target sweet potato tuber within each preset period. Based on the peak characteristics of the spectral signal curves, it adaptively smooths each spectral signal curve to obtain the target spectral signal curve. By separating possible soil scattering interference, it smooths out scattering interference anomalies. Furthermore, by combining the curve variation differences of multiple target spectral signal curves, it obtains the possibility that the spectral curve fluctuations may be affected by soil reflection interference, i.e., the degree of soil reflection anomaly. Thus, it corrects the anomalies that may be affected by soil reflection interference, obtaining relatively accurate spectral curve data (corrected target spectral signal curve) that is not affected by the soil above the sweet potato tuber. This improves the accuracy of monitoring the growth status of the target sweet potato tuber based on the corrected target spectral signal curve. Attached Figure Description

[0035] To more clearly illustrate the technical solutions in the embodiments of the present invention, 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0036] Figure 1 This is a structural block diagram of a sweet potato tuber growth status monitoring system based on spectral imaging analysis provided in Embodiment 1 of the present invention;

[0037] Figure 2 This is a schematic diagram of a two-dimensional plane provided in an embodiment of the present invention;

[0038] Figure 3 This is a schematic diagram of the division of a target spectral signal curve in a two-dimensional plane provided by an embodiment of the present invention. Detailed Implementation

[0039] Embodiments of this disclosure are described in detail below, with examples of these embodiments illustrated in the accompanying drawings. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this disclosure, and should not be construed as limiting it.

[0040] It should be noted that the terms "first," "second," etc., used in this disclosure and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure.

[0041] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0042] See Figure 1 This is a structural block diagram of a sweet potato tuber growth status monitoring system based on spectral imaging analysis provided in Embodiment 1 of the present invention, as shown below. Figure 1 As shown, the system may include:

[0043] The data acquisition module 11 is used to acquire the infrared spectral signal of the target sweet potato tuber within a preset period and construct the corresponding spectral signal curve. The horizontal axis of the spectral signal curve is the wavelength, and the vertical axis is the reflection intensity. Based on the peaks and troughs in the spectral signal curve, the spectral signal curve is smoothed to obtain the target spectral signal curve.

[0044] The specific scenario addressed by this invention is as follows: When using an infrared light source and a spectral detector to monitor sweet potato tubers, the spectral data collected from the sweet potato tubers is subject to scattering and reflection of the infrared light by the soil above the tubers, which masks the signal from the sweet potato tubers. As a result, the spectral signal actually received by the spectral detector includes the reflected signal from the soil, making the acquisition of the spectral signal from the sweet potato tubers inaccurate and leading to a large error in monitoring the growth status of the sweet potato tubers based on the spectral signal.

[0045] Therefore, to solve the above problems, after irrigating the target sweet potato, infrared light is emitted to the target sweet potato tubers using infrared spectroscopy. Then, a spectrometer (such as an FTIR or fiber optic spectrometer) is used to receive the reflected infrared spectral signals. This is used to obtain the material characteristics of the target sweet potato tubers by detecting the spectral signals reflected from them. The received infrared spectral signals are separated into reflection intensities at different wavelengths, forming a spectral signal curve of the target sweet potato tubers. In this curve, the horizontal axis represents wavelength, and the vertical axis represents reflection intensity.

[0046] In this embodiment of the invention, following the method for acquiring spectral signal curves described above, the spectral signal curves of the target sweet potato tuber are acquired multiple times continuously, resulting in multiple spectral signal curves over a monitoring period of 24 hours. The acquisition frequency of the spectral signal curves is once every 10 minutes, i.e., the preset period is 10 minutes.

[0047] Because soil particles are uneven in size, infrared spectral signals may be scattered when they enter the soil due to the uneven density distribution and complex soil particles. Therefore, in this embodiment of the invention, abnormal spectral signals that may be affected by scattering are first monitored in each spectral signal curve. The abnormal spectral signals are smoothed to eliminate the influence of abnormal spectral signals caused by infrared scattering. Then, after smoothing the abnormal spectral signals, the degree of soil scattering influence on each spectral signal curve is analyzed to correct the actual collected spectral signals and obtain spectral data that can more realistically and accurately characterize the growth status of the target sweet potato tubers.

[0048] In smoothing abnormal spectral signals in each spectral signal curve, considering the influence of soil scattering, the spectrometer will receive unconventional spectral wavelengths, resulting in abrupt changes in the reflection intensity of some wavelengths. These abrupt changes manifest as peaks or troughs in the spectral signal curve that show significant differences in reflection intensity compared to adjacent wavelengths, distinct from the fluctuations of normal peaks or troughs. Secondly, since scattering can cause abrupt changes in the reflection intensity of a certain wavelength in a normal spectral signal curve, these abrupt changes appear as peaks or troughs composed of only a single coordinate point on the spectral signal curve. Furthermore, the waveform at these peaks or troughs is relatively narrow, with weak fluctuation continuity. Therefore, in this embodiment of the invention, taking a single spectral signal curve as an example, the spectral signal curve is smoothed based on its peaks and troughs to obtain the target spectral signal curve.

[0049] The AMPD peak finding algorithm is used to obtain the peaks and troughs in the spectral signal curve. Based on the differences between the adjacent data points of each peak and trough, each peak and trough is traversed sequentially to obtain the corresponding scattering anomaly degree. Specifically, for any one of the peaks and troughs, the absolute value of the difference in reflection intensity between the point and its left and right adjacent data points is calculated to obtain the average absolute value of the difference. The average absolute value of the difference is then normalized to obtain the local difference degree of the point.

[0050] On the spectral signal curve, obtain the left adjacent peak or left adjacent trough of any point, count the number of left interval data points between any point and the left adjacent peak or left adjacent trough, obtain the number of right interval data points between any point and the right adjacent peak or right adjacent trough, and use a preset exponential function to inversely normalize the sum of the number of left interval data points and the number of right interval data points to obtain the fluctuation continuity of any point;

[0051] The degree of scattering anomaly at any given point is obtained based on the degree of local difference at that point and the degree of fluctuation duration.

[0052] In one embodiment, taking the i-th wave crest as an example, the expression for calculating the scattering anomaly degree of the i-th wave crest is:

[0053]

[0054] Among them, S i The scattering anomaly at the i-th peak is represented by norm(), which is the normalization function, and || represents the absolute value sign. i x represents the reflection intensity at the i-th wave crest. i,l x represents the reflection intensity of the left adjacent data point of the i-th wave crest. i,r The reflection intensity of the right adjacent data point of the i-th wave crest is represented by exp(), which represents an exponential function with the natural constant as the base. i+1 n represents the number of data points between the i-th peak and the (i+1)-th peak, which is also the number of right-interval data points. i-1 This represents the number of data points between the i-th peak and the (i-1)-th peak, which is also the number of data points in the left interval.

[0055] It should be noted that, Used to characterize the local difference of the i-th wave crest, the greater the difference in reflection intensity between the i-th wave crest and its left and right adjacent data points, the more obvious the abnormal prominence of the i-th wave crest compared to normal wave crests, the greater the risk of anomalies caused by scattering, and the greater the degree of scattering anomaly of the i-th wave crest; (n i+1 +n i-1 ) is used to characterize the degree of fluctuation continuity from the i-th wave crest to its left and right adjacent wave crests, (n i+1 +n i-1 The larger the value of ), the larger the waveform of the i-th peak point, the stronger the extension, and the more likely it is to be a sudden peak point caused by scattering. The greater the degree of scattering anomaly corresponding to the i-th peak point.

[0056] After determining the degree of scattering anomaly, an anomaly threshold of 0.8 is set. If the scattering anomaly degree at any point is greater than or equal to the preset threshold, the average reflection intensity between the left and right adjacent data points of that point is obtained, and this average reflection intensity is used to replace the reflection intensity of that point. If the scattering anomaly degree at any point is less than the preset threshold, the reflection intensity of that point remains unchanged. Similarly, the scattering anomaly degree at each peak and trough in the spectral signal curve is obtained, and the reflection intensity at each peak and trough is smoothed according to the scattering anomaly degree, resulting in a smoothed spectral signal curve, which is denoted as the target spectral signal curve, thus smoothing out the influence of scattering-induced abnormal signals.

[0057] At this point, the smoothing process for each spectral signal curve is complete, resulting in multiple target spectral signal curves.

[0058] The curve division module 12 is used to acquire target spectral signal curves within at least two consecutive preset periods, map all target spectral signal curves onto the same two-dimensional plane, where the horizontal axis of the two-dimensional plane is wavelength and the vertical axis is reflection intensity, and divide all target spectral signal curves in the two-dimensional plane into at least two sets of spectral signal sub-curves based on the inflection point of each target spectral signal curve in the two-dimensional plane.

[0059] Considering that the moisture characteristics of sweet potato tubers are generally displayed between 1400nm and 1900nm in the spectrum, but after irrigation, the soil layer above the sweet potato tubers contains a large amount of water and the soil particles have a high water content. Before the infrared light reaches the sweet potato tubers, some of the larger soil particles with water content may reflect the infrared light to the spectrometer in advance. They will also be present in the 1400nm to 1900nm range on the spectrometer. As a result, the spectrum of the moisture characteristic band actually monitored is not the manifestation of the sweet potato tubers, but is mixed with the characterization of the soil.

[0060] Therefore, in this embodiment of the invention, the differences in water loss rates between soil and sweet potato tubers are analyzed by combining the water loss characteristics of soil and sweet potato tubers. Specifically, the soil above the sweet potato tubers is directly exposed to the air and is subject to solar radiation and wind, resulting in significant water loss through evaporation and a faster water loss rate after irrigation. However, the water-storing cells in the sweet potato tubers lose water due to dry external environments or pressure changes, and the tuber cell structure has a certain water retention capacity. Therefore, when the surrounding soil moisture is sufficient, the water loss rate of the sweet potato tubers is lower. Correspondingly, after irrigation, as time progresses, the water content in the sweet potato tubers gradually decreases in the normally collected spectral signal curves, forming multiple spectral signal curves with gradually decreasing reflectance. Based on the differences in water loss characteristics between soil and sweet potato, within the water characteristic band range, if there are spectral signal fluctuations caused by soil reflection, the intervals between different soil spectral signal curves should be larger than those of the sweet potato tuber spectral signal curves, and the waveform changes between adjacent monitored spectral signal curves should be more pronounced.

[0061] Based on the above features, embodiments of the present invention acquire target spectral signal curves within at least two consecutive preset periods, and map all target spectral signal curves onto the same two-dimensional plane, referring to... Figure 2 It is a schematic diagram of a two-dimensional plane, with wavelength A on the horizontal axis and reflection intensity B on the vertical axis. The two-dimensional plane includes the target spectral signal curves under multiple monitoring of the target sweet potato tuber, with different colors representing different target spectral signal curves.

[0062] Although each target spectral signal curve differs slightly in reflection intensity and waveform, their fluctuation trends are similar. Therefore, the locations of their inflection points are also similar. This allows the clustered target spectral signal curves to be divided into multiple segments by using adjacent inflection points, each segment representing a characteristic fluctuation. Specifically, firstly, the inflection points on each target spectral signal curve are obtained and marked in a two-dimensional plane using the second derivative method. Inflection points on a curve are points where the curve's concavity / convexity changes, i.e., the boundary between concavity and convexity. The second derivative method is existing technology and will not be elaborated upon here. Then, based on the inflection points of each target spectral signal curve in the two-dimensional plane, all target spectral signal curves in the two-dimensional plane are divided into at least two sets of spectral signal sub-curves.

[0063] The partitioning method is as follows: cluster the inflection points of all target spectral signal curves in the two-dimensional plane to obtain at least two inflection point clusters, obtain the centroid of each inflection point cluster, draw a vertical line through each centroid according to the position of each centroid in the two-dimensional plane, and use the vertical line to divide all target spectral signal curves in the two-dimensional plane into at least two sets of spectral signal sub-curves. That is, the spectral signal sub-curves belonging to the same wavelength range in each target spectral signal curve are grouped into a set, and one set of spectral signal sub-curves corresponds to one wavelength range.

[0064] In one embodiment, the K-means clustering algorithm is used to cluster all marked inflection points on a two-dimensional plane, resulting in multiple inflection point clusters. Each inflection point cluster represents a fluctuation feature. Then, the centroid within each inflection point cluster is obtained. Based on the position of each centroid in the two-dimensional plane, a vertical line is drawn through each centroid, dividing all target spectral signal curves in the two-dimensional plane into at least two sets of spectral signal sub-curves. (Refer to...) Figure 3 It is a schematic diagram of the division of the target spectral signal curve in a two-dimensional plane. Figure 3 The dashed line in the middle is a vertical line drawn through the centroid of the cluster, corresponding to... Figure 3 The data was divided into 7 sets of spectral signal sub-curves.

[0065] The curve correction module 13 is used to obtain the soil reflection anomaly degree of the corresponding spectral signal sub-curve set based on the difference between each spectral signal sub-curve in each spectral signal sub-curve set, and to correct the reflection intensity of each spectral signal on each target spectral signal curve according to the soil reflection anomaly degree of each spectral signal sub-curve set, so as to obtain the corrected target spectral signal curve.

[0066] After obtaining the set of spectral signal sub-curves, the differences in water loss rate characteristics can be analyzed sequentially by sub-wave segments to obtain the degree of soil reflectance anomaly for each set of spectral signal sub-curves. Therefore, in this embodiment of the invention, for any set of spectral signal sub-curves, the wavelength range corresponding to the set of spectral signal sub-curves is obtained, any wavelength in the wavelength range is taken as the target wavelength, and the absolute value of the difference in reflectance intensity between any two adjacent spectral signal sub-curves in the set of spectral signal sub-curves at the target wavelength is calculated to obtain the sum of the absolute values ​​of the differences.

[0067] The sum of the absolute values ​​of the differences corresponding to each wavelength in the wavelength range is obtained, and the mean of the sum of the absolute values ​​of the differences is used as the degree of signal change difference of any set of spectral signal sub-curves.

[0068] In one embodiment, taking the p-th spectral signal sub-curve set as an example, the degree of difference in signal variation of the p-th spectral signal sub-curve set is calculated:

[0069]

[0070] Among them, C p n represents the degree of difference in signal variation in the p-th spectral signal sub-curve set. p Let represent the number of wavelengths in the wavelength range corresponding to the p-th spectral signal sub-curve set, and m represent the number of spectral signal sub-curves in the p-th spectral signal sub-curve set, which is also the number of target spectral signal curves. i_j Let x represent the emission intensity corresponding to the i-th wavelength on the j-th spectral signal sub-curve in the p-th spectral signal sub-curve set. i_j+1 Let || represent the emission intensity corresponding to the i-th wavelength on the (j+1)-th spectral signal sub-curve in the p-th spectral signal sub-curve set, where || represents the absolute value sign.

[0071] It should be noted that, The overall reflection intensity difference of all spectral signal sub-curves at the same wavelength is characterized. The greater the overall reflection intensity difference, the greater the degree of difference in spectral signal change at that wavelength. Therefore, when the mean of the overall reflection intensity difference of all wavelengths is greater, it indicates that the difference between adjacent spectral signal sub-curves in the p-th spectral signal sub-curve set is greater, which corresponds to the change in the spectral signal of water loss characteristics of the soil under heavy load.

[0072] Following the method described above for obtaining the degree of signal variation difference in the p-th spectral signal sub-curve set, the degree of signal variation difference for each spectral signal sub-curve set is obtained. Based on the degree of signal variation difference for each spectral signal sub-curve set, the degree of soil reflectance anomaly for any spectral signal sub-curve set is obtained. The specific method is as follows:

[0073] Calculate the average value of the differences in signal variation among all spectral signal sub-curve sets. Calculate the absolute value of the difference between the difference in signal variation of any spectral signal sub-curve set and the average value, and record it as the difference feature value. Normalize the product of the difference feature value and the difference in signal variation of any spectral signal sub-curve set to obtain the degree of soil reflectance anomaly of any spectral signal sub-curve set.

[0074] In one embodiment, the expression for calculating the degree of soil reflectance anomaly in the p-th spectral signal sub-curve set is:

[0075]

[0076] Among them, W pThe soil reflectance anomaly is represented by the set of the p-th spectral signal sub-curves, norm() represents the normalization function, and C p This indicates the degree of difference in signal variation within the set of the p-th spectral signal sub-curves. This represents the average of the differences in signal variation among all spectral signal sub-curves, where || denotes the absolute value sign.

[0077] It should be noted that, The difference between the soil reflectance anomaly degree of the p-th spectral signal sub-curve set and the mean soil reflectance anomaly degree of all spectral signal sub-curve sets is used to characterize the difference. The fluctuation differences among normal spectral signal curves of sweet potato tubers are small and similar, while the fluctuation differences of spectral signal curves caused by soil reflectance are greater, and the differences from spectral signal curves in other wavelength ranges are more obvious. Therefore... The larger the value, the more likely the p-th spectral signal sub-curve set is to be disturbed by soil reflection, and the greater the degree of soil reflection anomaly. At the same time, the greater the difference in signal variation of the p-th spectral signal sub-curve set, the larger its own variation difference, and the more likely it is to be disturbed by soil reflection, and the greater the degree of soil reflection anomaly.

[0078] Similarly, the degree of soil reflectance anomaly is obtained for each set of spectral signal sub-curves. Then, based on the degree of soil reflectance anomaly for each set of spectral signal sub-curves, the reflectance intensity of each spectral signal on each target spectral signal curve is corrected to eliminate interference from soil reflectance and improve the realism of the spectral signal data of sweet potato tubers. The correction method is as follows:

[0079] The degree of soil reflection anomaly in each set of spectral signal sub-curves is compared with a preset threshold for soil reflection anomaly. If the degree of soil reflection anomaly in any set of spectral signal sub-curves is less than the threshold for soil reflection anomaly, the reflection intensity of each spectral signal on each spectral signal sub-curve in each set of spectral signal sub-curves remains unchanged.

[0080] If the soil reflectance anomaly degree of any set of spectral signal sub-curves is greater than or equal to the soil reflectance anomaly degree threshold, then the reflectance intensity of each spectral signal on each spectral signal sub-curve in the set of any set of spectral signal sub-curves is corrected to obtain the corrected set of spectral signal sub-curves.

[0081] Using the spectral signal sub-curves in each of the modified spectral signal sub-curve sets, the corresponding spectral signal sub-curves in each of the target spectral signal curves are replaced to obtain each modified target spectral signal curve.

[0082] In one embodiment, a soil reflection anomaly threshold of 0.7 is set. If the soil reflection anomaly of any spectral signal sub-curve set is less than 0.7, it indicates that the spectral signal data in the wavelength range corresponding to that spectral signal sub-curve set is not affected by soil reflection, thus keeping each spectral signal sub-curve in that set unchanged. Conversely, if the soil reflection anomaly of any spectral signal sub-curve set is greater than or equal to 0.7, it is considered that the spectral signal data in the wavelength range corresponding to that spectral signal sub-curve set is affected by soil reflection, thus requiring correction of the spectral signal data of each spectral signal sub-curve in that set. In this case, each spectral signal sub-curve set is traversed, and based on the traversal results, each target spectral signal curve is corrected and adjusted to obtain the corrected target spectral signal curve, which is the true spectral signal curve that has eliminated soil reflection interference.

[0083] Preferably, the reflection intensity of each spectral signal on each spectral signal sub-curve in the arbitrary spectral signal sub-curve set is corrected to obtain a corrected spectral signal sub-curve set, including:

[0084] For any spectral signal on each spectral signal sub-curve in the arbitrary spectral signal sub-curve set, obtain the product between a preset multiple and the degree of soil reflectance anomaly in the arbitrary spectral signal sub-curve set. Use the sum of the constant 1 and the product as the correction coefficient of the arbitrary spectral signal. Based on the product of the correction coefficient and the reflectance intensity of the arbitrary spectral signal, obtain the corrected reflectance intensity of the arbitrary spectral signal.

[0085] In one embodiment, because soil moisture loss is faster, the spectral signal curve affected by soil reflection will have lower fluctuations in the 1400nm to 1900nm range compared to the spectral signal curve of normal sweet potato tubers. Therefore, correction requires correspondingly increasing the reflection intensity of the spectral signal data in the abnormal wavelength range. Taking the k-th spectral signal on any spectral signal sub-curve in the p-th spectral signal sub-curve set as an example, the calculation expression for the corrected reflection intensity of the k-th spectral signal is as follows:

[0086] x′ k =x k ×(1+0.2×W p )

[0087] Where, x′ k x represents the corrected reflection intensity of the k-th spectral signal, which is also the corrected reflection intensity. k W represents the reflection intensity of the k-th spectral signal, which is also the reflection intensity before correction. 1 indicates a constant, and 0.2 indicates a preset multiple. p This represents the degree of soil reflectance anomaly in the p-th spectral signal sub-curve set.

[0088] At this point, the spectral signal curve of each target has been corrected, resulting in multiple corrected target spectral signal curves.

[0089] The status monitoring module 14 is used to monitor the growth status of the target sweet potato tuber based on each of the corrected target spectral signal curves.

[0090] Each corrected target spectral signal curve is compared with the standard spectral signal curve during normal sweet potato growth to obtain the peak significance of each corrected target spectral signal curve. When the peak significance of any corrected target spectral signal curve exceeds the allowable peak significance error range, it indicates that the growth status of the target sweet potato tuber after irrigation is abnormal. It is necessary to check the target sweet potato tuber in time to see if there is uneven watering or poor irrigation flow, and make timely adjustments.

[0091] It should be noted that the focus of this invention is on how to correct the spectral signal curve of sweet potato tubers. Monitoring the growth status of target sweet potato tubers based on the corrected target spectral signal curve is existing technology and will not be elaborated here.

[0092] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A sweet potato tuber growth status monitoring system based on spectral imaging analysis, characterized in that, The system includes: The data acquisition module is used to acquire the infrared spectral signal of the target sweet potato tuber within a preset period and construct the corresponding spectral signal curve. The horizontal axis of the spectral signal curve is the wavelength, and the vertical axis is the reflection intensity. Based on the peaks and troughs in the spectral signal curve, the spectral signal curve is smoothed to obtain the target spectral signal curve. The curve division module is used to acquire target spectral signal curves within at least two consecutive preset periods, map all target spectral signal curves onto the same two-dimensional plane, where the horizontal axis of the two-dimensional plane is wavelength and the vertical axis is reflection intensity, and divide all target spectral signal curves in the two-dimensional plane into at least two sets of spectral signal sub-curves based on the inflection point of each target spectral signal curve in the two-dimensional plane. The curve correction module is used to obtain the degree of soil reflectance anomaly of the corresponding spectral signal sub-curve set based on the difference between each spectral signal sub-curve in each spectral signal sub-curve set, and to correct the reflection intensity of each spectral signal on each target spectral signal curve according to the degree of soil reflectance anomaly of each spectral signal sub-curve set, so as to obtain the corrected target spectral signal curve. The status monitoring module is used to monitor the growth status of the target sweet potato tuber based on each of the corrected target spectral signal curves. The data acquisition module smooths the spectral signal curve based on the peaks and troughs in the spectral signal curve to obtain the target spectral signal curve, including: Obtain the peak and trough points in the spectral signal curve. For any one of the peak and trough points, obtain the scattering anomaly degree of the point based on the adjacent data points on the spectral signal curve. If the scattering anomaly at any point is greater than or equal to a preset anomaly threshold, then the average reflection intensity between the left and right adjacent data points of any point is obtained, and the average reflection intensity is used to replace the reflection intensity of any point; if the scattering anomaly at any point is less than the preset anomaly threshold, then the reflection intensity of any point remains unchanged. By traversing each of the stated peaks and troughs, the target spectral signal curve is obtained; The curve segmentation module divides all target spectral signal curves in the two-dimensional plane into at least two sets of spectral signal sub-curves based on the inflection points of each target spectral signal curve in the two-dimensional plane, including: Cluster the inflection points of all target spectral signal curves in the two-dimensional plane to obtain at least two inflection point clusters. Obtain the centroid of each inflection point cluster. Based on the position of each centroid in the two-dimensional plane, draw a vertical line through each centroid. Use the vertical line to divide all target spectral signal curves in the two-dimensional plane into at least two sets of spectral signal sub-curves. The curve correction module obtains the degree of soil reflectance anomaly for each spectral signal sub-curve set based on the differences between the various spectral signal sub-curves in each spectral signal sub-curve set, including: For any set of spectral signal sub-curves, obtain the wavelength range corresponding to the set of spectral signal sub-curves, take any wavelength in the wavelength range as the target wavelength, calculate the absolute value of the difference in reflection intensity between every two adjacent spectral signal sub-curves in the set of spectral signal sub-curves at the target wavelength, and obtain the sum of the absolute values ​​of the differences. The sum of the absolute values ​​of the differences corresponding to each wavelength in the wavelength range is obtained, and the mean of the sum of the absolute values ​​of the differences is used as the degree of signal change difference of any set of spectral signal sub-curves. The degree of signal variation difference for each of the spectral signal sub-curve sets is obtained, and the degree of soil reflectance anomaly for any spectral signal sub-curve set is obtained based on the degree of signal variation difference for each of the spectral signal sub-curve sets. The curve correction module corrects the reflection intensity of each spectral signal on each target spectral signal curve according to the degree of soil reflectance anomaly in each set of spectral signal sub-curves, to obtain the corrected target spectral signal curve, including: The degree of soil reflection anomaly in each set of spectral signal sub-curves is compared with a preset threshold for soil reflection anomaly. If the degree of soil reflection anomaly in any set of spectral signal sub-curves is less than the threshold for soil reflection anomaly, the reflection intensity of each spectral signal on each spectral signal sub-curve in each set of spectral signal sub-curves remains unchanged. If the soil reflectance anomaly degree of any set of spectral signal sub-curves is greater than or equal to the soil reflectance anomaly degree threshold, then the reflectance intensity of each spectral signal on each spectral signal sub-curve in the set of any set of spectral signal sub-curves is corrected to obtain the corrected set of spectral signal sub-curves. Using the spectral signal sub-curves in each of the modified spectral signal sub-curve sets, the corresponding spectral signal sub-curves in each of the target spectral signal curves are replaced to obtain each modified target spectral signal curve.

2. The sweet potato tuber growth status monitoring system based on spectral imaging analysis according to claim 1, characterized in that, The step of obtaining the scattering anomaly degree of any point based on adjacent data points on the spectral signal curve includes: Calculate the absolute value of the difference in reflection intensity between any point and its left and right adjacent data points respectively, obtain the average absolute value of the difference, and normalize the average absolute value of the difference to obtain the local difference degree of any point. On the spectral signal curve, obtain the left adjacent peak or left adjacent trough of any point, count the number of left interval data points between any point and the left adjacent peak or left adjacent trough, obtain the number of right interval data points between any point and the right adjacent peak or right adjacent trough, and use a preset exponential function to inversely normalize the sum of the number of left interval data points and the number of right interval data points to obtain the fluctuation continuity of any point; The degree of scattering anomaly at any given point is obtained based on the degree of local difference at that point and the degree of fluctuation continuation.

3. The sweet potato tuber growth status monitoring system based on spectral imaging analysis according to claim 1, characterized in that, The step of obtaining the soil reflectance anomaly degree of any spectral signal sub-curve set based on the degree of signal change difference of each spectral signal sub-curve set includes: Calculate the average value of the signal variation difference of all spectral signal sub-curve sets, calculate the absolute value of the difference between the signal variation difference of any spectral signal sub-curve set and the average value, and record it as the difference feature value. Normalize the product of the difference feature value and the signal variation difference of any spectral signal sub-curve set to obtain the soil reflectance anomaly degree of any spectral signal sub-curve set.

4. The sweet potato tuber growth status monitoring system based on spectral imaging analysis according to claim 1, characterized in that, The step of correcting the reflection intensity of each spectral signal on each spectral signal sub-curve in the arbitrary spectral signal sub-curve set to obtain the corrected spectral signal sub-curve set includes: For any spectral signal on each spectral signal sub-curve in the arbitrary spectral signal sub-curve set, obtain the product between a preset multiple and the degree of soil reflectance anomaly in the arbitrary spectral signal sub-curve set. Use the sum of the constant 1 and the product as the correction coefficient of the arbitrary spectral signal. Based on the product of the correction coefficient and the reflectance intensity of the arbitrary spectral signal, obtain the corrected reflectance intensity of the arbitrary spectral signal.

Images