Imaging method of spatial structure quality spectrum division of fruits

Abstract

An imaging method for spatial structure quality of fruit comprises; obtaining fruit samples from different light conditions in a canopy of a fruit tree; taking tissue sections from different light par

Description

Illumination, orientation and other environmental factors have a great impact on that quality of the fruits. For existing fruit nondestructive quantitative detection, factors of illumination, growth position and other factors have not been taken into account. Because the fruits have various sizes of particle micro-clusters, the refractive index distribution and spatial arrangement of different micelles are various, and the local and overall characteristics have similar fractal characteristics. The light entering the detector is multiple scattered and diffracted by different parts of the fruit tissues, which has a great influence on the detection results. Currently, for different distance remote sensing detection, the multi-points average method is often used to obtain the quality spectrum, and the spectral density change of spatial structure in spectral transmission has not been considered. In addition, the refractive index and pore size distribution of fruit components change greatly at different ripening stages, resulting in great differences in microstructure and characterization functions of different characteristic scales. In the existing technology, the influence of microstructure changes on the spectrum has not been considered when detecting the components of fruit epidermis. In particular, it is more difficult to detect the fruit quality of different growth directions of fruit trees, and its detection accuracy is low. Therefore, it is an urgent problem for those skilled in the art to provide a fractal quantitative detection method for fruit based on spatial spectral imaging which can consistently detect fruit surface features and internal quality with different scales and has high detection accuracy. SUMMARY In view of the above, an object of the present disclosure is to provide an imaging method of spatial structure quality spectrum division of fruits. The consistent expression of the structure image and the function quality imaging of different scale of fruits is thus realized, and at the same tine the detection precision of fruit is improved. It can also function as reference for agricultural machinery agriculture. Technical solutions of the present disclosure are specifically described as follows. An imaging method of spatial structure quality spectrum division of fruits is disclosed. The method includes obtaining fruit samples under different light conditions in a canopy of fruit tree in the order from an outside to an inside and from top to bottom according to a growth orientation of the fruits; taking tissue sections from different light parts of each fruit sample separately; marking the tissue sections of the fruit samples directionally according to a maturity stage, a growth position and a section position of the sample; obtaining multi-angle polarized spectrum images of tissue sections of each fruit sample and constructing a spatial spectrum database; simulating the tissue sections of fruit samples of different ripening stages, and obtaining fractal dimension of spectrum of different fruit samples at different spatial scales; classifying the spatial spectral database with spectral fractal dimension and generating a spectral library of spatial fractal characteristics; and determing a corresponding relationship between fruit quality and growth position and illumination position based on the spatial fractal feature spectral library. Preferably, the spectral fractal dimension in the spatial fractal feature spectral library corresponds to the multi-angle polarized spectral images with different feature scales and different illumination conditions. Preferably, the fruit samples also include fruits at different times, different temperatures, and different topographical conditions. In comparison with that prior art, an imaging method of spatial structure quality spectrum division of fruits is disclosed, which comprehensively considers the influence of factors such as different ripening stages, different growth position and different illumination conditions. In the method, fruit under different light condition are sampled, a sample database is enriched, the relationship between local position characteristics and environmental factor characteristic of a plurality of fruits and the overall quality change is correlated, and the fruit quality of different regions is finely distinguished. In order to establish a corresponding relationship between fruit quality and environmental factors such as growth direction, illumination, temperature, etc., not only the orchard quality distribution can be retrieved, but also the management reference can be provided for agricultural machinery agriculture. It can also greatly improve the detection accuracy of fruit quality, and is of great significance in the areas, time-sharing, multilevel picking, and accurate classification of fruit evaluation. In addition, the disclosure considers the spectral density change of the spectrum in the process of the spatial transmission of the internal tissue of the fruit in different ripening stages, and carries out the coherent scale differentiation position verification on the polarized spectral image of a fruit samples. In the method, the polarized spectrum image is finely classified, the scale effect is solved, a spatial fractal characteristic spectrum library is obtained, and the consistent expression of the imaging of different scale structures and the functional quality image is realized through the spatial fractal characteristics spectrum library. That is, the consistent expression of fruit surface characteristics and internal quality can be realized, and the detection accuracy of fruit quality can further be improved. BRIEF DESCRIPTION OF THE DRAWINGS In order to explain the embodiments of the present disclosure or the technical solutions in the prior art more clearly, the following drawings that need to be used in the description of the embodiments or the prior art will be briefly introduced. Obviously, the drawings in the following description are only embodiments of the present disclosure. For those of ordinary skill in the art, other drawings can be obtained based on the drawings disclosed without creative work. FIG. 1 is a schematic diagram of stratified and zonal sampling of fruits of the disclosure; FIG. 2 is a schematic diagram of a partially coherent light source provided by the disclosure, the fruit is equivalent to a random scattering medium, and the spectral densities of different orientations of the detector are different; FIG. 3 is a frame diagram of a fractal method for classifying spatial spectral images of fruit samples provided by the disclosure; FIG. 4 is a schematic diagram of spectral differences of different distances at the same position of the same date provided by the disclosure; FIG. 5 is a polarization parameter image obtained in a multi-angle polarization experiment provided by the disclosure; FIG. 6 is a schematic diagram showing differences in polarization spectrum characteristics of bright spots of red dates in the polarization parameter image provided by the disclosure; FIG. 7 is a schematic diagram showing differences in polarization spectrum characteristics of non-bright spots of red dates in the polarization parameter image provided by the present invention. DETAILED DESCRIPTION OF EMBODIMENTS Technical solutions of the present disclosure will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only part of the embodiments of the present disclosure, not all of them. Other embodiments made by those skilled in the art without sparing any creative effort should fall within the scope of the disclosure. An imaging method of spatial structure quality spectrum division of fruits is disclosed in the disclosure. The method includes: S1. obtaining fruit samples under different light conditions in a canopy of fruit tree in the order from an outside to an inside and from top to bottom according to a growth orientation of the fruits; S2. taking tissue sections from different light parts of each fruit sample separately; S3, marking the tissue sections of the fruit samples directionally according to a maturity stage, a growth position and a section position of the sample; S4. obtaining multi-angle polarized spectrum images of tissue sections of each fruit sample and constructing a spatial spectrum database; S5. simulating the tissue sections of fruit samples of different ripening stages, and obtaining fractal dimension of spectrum of different fruit samples at different spatial scales; S6. classifying the spatial spectral database with spectral fractal dimension and generating a spectral library of spatial fractal characteristics; and S7. determining a corresponding relationship between fruit quality and growth position and illumination position based on the spatial fractal feature spectral library. As shown in FIG. 1, under the influence of the factors such as different periods, different growing positions and different lighting directions, the same fruit tree has different canopy positions (upper, middle and lower) due to the light receiving conditions and nutrient differences. Differences exist between the inner and outer quality of canopy. Different parts of the same fruit have different qualities due to different illumination, and the difference in quality can be expressed by color depth. In the sample process of the fruits, the environment influence factor such as illumination, growth orientation and the like are separated through the spatial fine orientation classification mark sampling of the front end of the fruits. Fruit tissues are classified and labeled with directional slices and the corresponding spectra are collected and analyzed. In the method, the tissue slice of the fruits is detected by a detector through multi-angle and accurate space orientation detection, as shown in Fig. 2, the corresponding relationship between fruit quality and environmental factors is analyzed based on the spectral images of different parts, different illumination conditions and different detection angles. Hereinafter, that above-mentioned steps will be described in details. 1. Sampling the fruits by layers and regions: In the disclosure, the fruits and the fruit tree canopy are divided into a light facing surface and a light backing surface and accurately marked, so that the local characteristic of the fruits correspond to the growth direction of the jujube canopy. In accordance with different part of the inner and out sides of the canopy layer and the upper, middle and lower parts of the canopies, the sample is sequentially sampled from the outside to the inside from the top to the bottom, and the regional marking is carried out by layers to realize the fine marking of different growing parts of fruits. The light direction of canopy growth is judged by the quality difference, and the consistent expression of local and global characteristics was realized. In the method, the local position characteristic and the environmental factor characteristics of multiple fruits are correlated with an overall quality change relationship, and the fruit quality of different regions is finely distinguished. The corresponding relationships are established between fruit quality and growth position, light, temperature and other factors. In addition, the disclosure also carries out tissue slices on different parts of the same fruit, the corresponding relationships between different light directions and different tissue parts of a same fruit are distinguished and the sample data are greatly enriched. Inspection accuracy of fruit quality in the later stage is thus ensured. 2. Classifying the spatial spectrum images of the fruit samples into the following fractal types. Firstly, the spatial spectrum images of the fruit canopy from outside to inside and from top to bottom are obtained in a directional sequence, and a spatial spectrum database of different fruit directions is established. In accordance with that diffuse diffusion method of the point light source, a single fruit sample is sequentially obtained from the outside to the inside, and tissue sections of different part of the fruit are finely marked from the top to the bottom. From collecting an upper, middle and bottom (i.e., facing and backing sunlight) spectrum, the correspondence relationship between the spectrum and the collected direction is established, and the spatial spectrum database of different part of the fruit tissue section is respectively established. Secondly, the scatter kernel function, the scattering medium structure and the scatter potential correlation function, different scatterers, and the scattering kernel function are construct by simulating the different quality fruit samples, the refractive index and the scale as the main parameters. Driving the experiment with different structure simulation, different angles of the standard plate sub-regional coherence, to verify the coherent scale differentiation location, to solve the scale effect. Finally, the scale effect of different spatial spectra of fruit not only have the direction of characteristic angle, but also have the characteristic scale. The structure function of variogram function, scatter potential correlation function, 2dcos two-dimensional spectrum correlation software are utilized to analyze and sort the correlation coefficients of different azimuth spectra in order according to the size of correlation coefficients. By setting the threshold value as the spatial resolution accuracy, for example, 0.01 corresponds to a spatial resolution distance of 0.1 meter, and fine classification of the spatial spectral data, the spectral fractal feature spectral library corresponding to the spatial scale and environmental factors such as light and temperature is obtained. 3. From spatial division of image formation to quality imaging. In the disclosure, the structure characteristic of the fruits are obtained through fine classification and analogy analysis of the spatial spectrum. Different growth position and environmental differences and quality differences are separated, so that the detection accuracy of the quality distribution of the fruits is improved. At the same time, the correspondence and correlation between the environmental factors and the fruit quality by collecting the fruit samples under various environmental factors is established, and by marking the fruits in different geographical positions and growing in different regions, a regional quality marking is achieved. It is obvious that the growth position is related to the quality distribution, and quality distribution is labeled with geographical orientation, so as to realize the purpose of imaging different orientation structures to different areas. By finely separating the quality factor and the environment factor, the corresponding relationship between the quality of the fruit and environment factors such as growth direction, light, temperature and the like is established, so as to provide management reference for the agricultural machinery agronomy of fruit. It is of great significance in fruit division, time-sharing and precise evaluation of fruit classification. The fractal classification is carried out on the spatial spectral images of the fruit sample from two aspects (micro and macro) respectively, so as to solve the scale effect and realize the fractal characteristic dimension of the fruits under different spatial scale. The process of obtaining the spectral fractal dimension in S5 is described in detail in the following, and the specific framework is shown in FIG. 3. S51: The fruits of different ripening stages are equivalent to quasi-uniform scattering media with different coherent scales, decomposed into isotropic particle clusters of different scales, and combined into anisotropic scatterers of different dimensions through different orientations; S52: the change rules of spectral scattering characteristics from near field to far field of different scatterers at different angles are obtained based on the scaling theorem of spectral transmission. S53, the spatial coherence property of the light source and the spatial scale of the illumination distribution are analyzed by using the cross spectral density function of different parts based on the spectral scattering change law and the radiation transmission change law, so as to obtain the fruit spatial characteristic spectrum; S54: calculate a scattering factor by simulation, and use the scattering factor to describe the unevenness of the spatial characteristic spectrum of the fruit; S5 5: determine the position of the coherent scale discrimination of the multi-angle polarization spectrum image of the fruit sample as the characteristic scale based on the scattering factor, and determine the fractal dimension of different mature fruit samples through the characteristic scales. The light propagates in a straight line in a homogeneous medium and scatters in a inhomogeneous medium, and the degree of scattering is determined by the non uniformity. The structure is uniform in a certain scale range, and this scale is the characteristic scale. In different characteristic scale, the quality characteristic of the optical spectrum are different, the scattering factor is use for scale classification, the conversion and reconstruction of the different characteristic scales are carried out, and the zoom lens can realize the spectral conversion and quality analysis at different distance, the classification accuracy and efficiency can be improved. In S6, the process of obtaining the spatial fractal characteristic spectrum library is as follows: S61: quantitatively and finely classify the multi-angle polarization spectral image of the fruit sample by using the characteristic scale to obtain a spatial fractal characteristic spectral image; S62: a spatial fractal feature spectral library is constructed based on the spatial fractal characteristic spectral image of the fruit sample, in which the spectral fractal dimension corresponds to the spatial scale and the illumination. The spatial fractal feature spectrum database in S6 includes a fruit spectral polarization fractal feature database and a fruit tissue microstructure fractal database. The construction process of the fruit spectral polarization fractal feature database is as follows: high-spectrum polarization experiment of fruit quality is carried out according to the angle distribution range of scatter light intensity of fruit with different mature period typical characteristic scales, and a multi-angle polarization experimental model is established; the quantitative power function model is used to generate the fruit quality distribution under different polarization conditions based on the multi-angle polarization experimental model. The quality distribution of fruit is described quantitatively by Lorentz diffusion curve characteristic parameters. the isotropic distribution and anisotropic distribution in different characteristic scales of epidermal tissue slice structure of fruit were determined, and the imaging intensity and distribution law of polarization spectrum are obtained based on the results of quantitative description. the fractal feature database of fruit spectral polarization is obtained based on the intensity and distribution of spectral imaging. In S6, the process of building the fractal database of fruit tissue microstructure is as follows: the mathematical morphology spectrum of the skin tissue section with different thickness and different orientation of the fruit is measured, and the scattering characteristics of different particles is simulated by means of the numerical morphology spectrum. the fractal theory is used to measure the shape of the digital image of the mathematical shape spectrum and obtain the distribution spectrum of the particle shape. the microstructure of the epidermal tissue section of the fruit is quantitatively characterize by using the polarization image parameters, and the particle number density distribution in the particle shape distribution spectrum is extracted by combining the mathematical morphology. The characteristic scale of fruit tissue structure at different maturity and the polarization fractal spectrum image at this scale is obtained. Fractal database of fruit tissue microstructure in different ripening stages is established based on polarized fractal spectral images. The outdoor fruit quality monitor can be carried out base on the spatial spectrum fractal process of the darkroom, and the method has the advantages of high monitoring efficiency and high precision, and is suitable for the quality distribution detection of fruit such as red dates, apples and fragrant pears. The jujube is taken as an example for relevant experiments with following. As shown in Table 1, a list of the consistency calibration tests for the external quality and internal quality of the dates in Nanjiang is described. Jujube samples are used in different periods, white ripe period, crisp ripe period and mature period. Physical and chemical indexes of jujube and its leaves: Appearance of red jujube: Color, texture, density, refractive index, refraction index, glossiness, electric conductivity, pulp structure or particle distribution; Jujube interior: Water, sugar, acidity; Leaf: Chlorophyll. Physical and chemical indexes of the test sample Test device method Test parameter model External quality Appearance fullness, color, texture roughness, Hardness, gloss High spectral camera of multi- spectral camera Multi-angle shooting quantitative measurement, machine vision image processing, Multi-angle statistical pore size distribution; Structural Light Illumination as Spatial Orientation Marking position detection and saturation detection Internal quality Moisture, sugar content, Acidity, determination Near infrared spectrum (600-2500 nm) The model of fruit quality inspection is established Fruit skin Tissue sections Ultrasoun d Measure structure scale (1mm, 5mm, 15mm) statistics Scale model of pore size distribution Tissue sections Multi-angle Polarization Microscopy 4 polarization directions (0, 90, 45, 135) Angle of detection (0, 90, 45, 135) Tissue sections Mimic model Photoelasticity measurement The Structure Characteristic of Hardness Refractive Index Distribution Imitation fruit fortissue section Different Refractive Index (n = 1.1 ~ 1.33), Size and Thickness (L = 20-2000um) Random combination fractal characteristic model Different quality fruits Polarizati on Calibration Experiment of Roughness and surface texture, the statistical difference of roughness on different scales, show Indoor calibration Space Structure of Stereo Camera the difference of scattering polarization. Single fruit High light spectrum diffuse scatter polarization experiment Diffusion of different qualities Diffusion curve a, b, c, d parameters, Describes slope, peak height, peak width 4 polarization directions (0, 90, 45, 135) Histological section of fruit at different maturity stages Backscatt er Polarizati on experiment Describe the characteristics of polarization specificity; In the Muller matrix array element, four parameters corresponding to the polarization conversion parameters, namely amplitude intercept angle, amplitude tl and t2 parameters, intercept b parameter and angle a parameter, are extracted quantitatively Different quality fruits Indoor and outdoor calibration field PBRDF Stacking characteristics Analog calibration In the Roujean model, the structure parameter kO is isotropic, KI geometric structure parameter, k2 body scattering; Separating the scattering depolarization matrix, the phase delay matrix, Dichroic matrix Table 1 Part of experimental results of spectral polarization transmission: In the transmission process of the partially coherent light source, the receiving spectrum density difference of the detectors at different distances are influenced by the spatial correlation of different point. The difference of different distance spectra of the same jujube has certain influences on the result of quantitative detection. It can be seen from FIG. 4 that the spectra of different distances at the same part of the selected fruit red dates are quite different at the positions of 233 band and 1676.01 band. With the increasing of distance, the interval between different characteristic reflection peaks and absorption peaks is not uniform. The reason is that the spectral density and spectral coherence of different distances are different, and the corresponding absorption characteristics are also different. By adding the multi-angle polarization experiment, the polarization parameter image is obtained. In FIG. 5, A is the original image, B is the degree of polarization (dolp), and C is the polarization phase angle (orient). The spatial characteristics of the image change, and the polarization has the effect of strong light weakening and weak light strengthening, which improves the image resolution effect. FIG. 6 and FIG. 7 respectively show the spectra of different parts of the same red date, and A, B and C respectively correspond to the spectrum of the red date in the original image, the degree of polarization dolp and the angle of polarization orient. As the quality of different parts of the same jujube is very small, but the difference of polarization spectrum characteristics is obvious, it can be seen that the recognition effect can be improved through the fractal of spectrum polarization characteristics. Due to the influence of angle and transmission distance on the spatial spectrum of fruit, it is necessary to carry out further radiometric calibration, polarization calibration and spectral calibration of the measuring instrument to improve the accuracy of quantitative detection. In the disclosure, base on the correlation theory of the spectral scattering and the radiation transmission in the medium, the diffuse diffusion propagation law in the scatter medium, a method for expressing the quantitative polarization parameters, The change law between the spectral structure and the spatial structure such as the size and distribution of particles in the medium, and the change of spatial structure represented by the cross spectral density. In order to study the application of multi-scale biological tissue detection methods, it is necessary to establish the changing rules and corresponding relations of spectrum shape characteristics by the factors such as the size and position of light source, and to extract the abundant angle information from polarimetric muller muller matrix. Each embodiment in this specification is described in a progressive manner, and each embodiment is focused on the difference from the other embodiments. Same parts similar to each other in each embodiment can be referred to respectively. As for the device disclosed in the embodiments, the description is relatively simple because it corresponds to the method disclosed in embodiments, and reference may be made to the description of the method section for the relevant points. The above description of the disclosed embodiments enables those skilled in the art to make or use the present invention. Numerous modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented without departing from the spirit or scope of the invention, In other embodiment implementations that. Accordingly, the invention will not be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein. 10 04 25 1. An imaging method of spatial structure quality spectrum characterisation of fruits, comprising: obtaining fruit samples under different light conditions in a canopy of fruit tree in the order from an outside to an inside and from top to bottom according to a growth orientation of the fruits; taking tissue sections from different light parts of each fruit sample separately; marking the tissue sections of the fruit samples according to a maturity stage, a growth position and a section position of the sample; obtaining multi-angle polarized spectrum images of tissue sections of each fruit sample and constructing a spatial spectrum database; simulating the tissue sections of fruit samples of different ripening stages, and obtaining fractal dimension of spectrum of different fruit samples at different spatial scales; classifying the spatial spectral database with spectral fractal dimension and generating a spectral library of spatial fractal characteristics; and determining a corresponding relationship between fruit quality and growth position and illumination position based on the spatial fractal feature spectral library. 2. The imaging method of spatial structure quality spectrum characterisation of fruits of claim 1, wherein the spectral fractal dimension in the spatial fractal feature spectral library- corresponds to the multi-angle polarized spectral images with different feature scales and different illumination conditions.

Claims

3. The imaging method of spatial structure quality spectrum characterisation of fruits of claim 1, wherein the fruit samples also include fruits at different times, different temperatures, and different topographical conditions.10 04 25

Images