A method for synergic control of fine processing and grading screening of rice

Abstract

The application discloses a rice fine processing grading screening synergic control method, relates to the rice fine processing grading technical field, and comprises appearance quality screening, optical characteristic screening and chemical residual screening. The first stage obtains the rice grain chroma coordinate, the chalky area proportion and the grain type length-width ratio through multi-spectrum imaging, generates the appearance sorting threshold through statistical analysis and iterative adjustment, and drives the color selection unit to execute rejection. The second stage collects the apparent refractive index of the intermediate material and the absorbance at multiple characteristic wavelengths, combines the feedback parameters of the color selection link to compensate and correct the sorting boundary, and generates the light transmission detection control parameter group. The third stage collects the near-infrared diffuse reflection data of the secondary intermediate material, analyzes the characteristic response of cellulose, lignin, silicon element and fatty acid, and maps the joint control parameters of the frequency of the variable frequency fan, the phase angle of the eccentric block of the vibrating screen and the longitudinal inclination angle of the stone removal screen surface.

Description

Technical Field

[0001] This invention relates to the field of rice fine processing and grading technology, specifically to a collaborative control method for rice fine processing, grading, and screening. Background Technology

[0002] Rice is one of the major food crops, and its processing quality directly affects its edible value and market price. As consumers' demands for rice quality continue to rise, traditional extensive processing methods can no longer meet market needs. Fine processing, grading, and screening have become the core links in the rice processing industry. Precise grading of rice based on its appearance, optical properties, and chemical residues can effectively improve the consistency and safety of rice products, enhancing market competitiveness. Therefore, a collaborative control method for fine processing, grading, and screening of rice is needed.

[0003] In existing technologies, rice grading and screening mainly employs a single-stage, independently controlled approach: appearance screening primarily relies on color sorters to remove discolored and chalky grains; optical screening mainly utilizes visible light transmission to detect internal cracks in rice grains; and chemical residue detection often employs offline laboratory methods, such as high-performance liquid chromatography (HPLC). While some advanced equipment has begun to incorporate near-infrared detection technology, it remains limited to the detection of a single indicator and has not yet achieved integrated control with subsequent physical screening stages.

[0004] The existing technology has the following drawbacks: 1. Each screening stage operates independently, and there is a lack of data communication and collaborative control mechanisms between them. The sorting error generated in the previous stage cannot be corrected or compensated by the subsequent stages. It will continue to accumulate and amplify as the processing progresses, eventually resulting in low overall grading accuracy of rice and difficulty in ensuring product quality consistency.

[0005] 2. The sorting threshold in the appearance screening process is mostly a fixed value preset by humans. It cannot be dynamically adapted and adjusted according to the differences in variety, origin and maturity of different batches of raw materials. When the quality of raw materials fluctuates naturally, it is very easy to missort or miss sorting phenomena such as mis-colored grains being collected, chalky grains being missed, and normal grains being mistakenly rejected.

[0006] 3. The detection results in the optical sorting process are easily affected by environmental factors such as gray scale changes caused by dust accumulation on the background plate and light source aging and decay during equipment operation. Furthermore, the lack of an effective real-time feedback compensation mechanism can lead to drift in the detection benchmark, resulting in poor sorting stability and difficulty in maintaining the accuracy requirements for long-term continuous production.

[0007] 4. Currently, chemical residue detection mostly adopts offline laboratory testing methods, which require manual sampling and sending to professional laboratories for analysis. This has the problems of long testing cycles and delayed result feedback, making it impossible to guide the production process in real time and difficult to accurately remove rice with excessive chemical residues online, thus posing certain food safety risks. Summary of the Invention

[0008] To address the aforementioned technical shortcomings, the present invention aims to provide a collaborative control method for grading and screening in the fine processing of rice.

[0009] To solve the above technical problems, the present invention adopts the following technical solution: The present invention provides a collaborative control method for grading and screening of finely processed rice, including the following steps: Step 1, appearance quality screening: collect appearance quality data of rice samples to be processed, perform statistical analysis on the appearance quality data of rice samples to be processed, determine the initial appearance sorting boundary, iteratively adjust the initial appearance sorting boundary to obtain the appearance sorting threshold, generate a set of appearance quality detection control parameters, complete the appearance quality screening, and obtain intermediate materials.

[0010] Step 2, Optical Property Screening: Collect the optical properties of the intermediate materials, perform statistical analysis on the optical properties of the intermediate materials to obtain the preliminary sorting boundary, then collect the actual execution parameter feedback data, and based on the actual execution parameter feedback data, compensate and correct the preliminary sorting boundary to generate a light transmission detection control parameter set, complete the optical property screening, and obtain the secondary intermediate materials.

[0011] Step 3: Chemical Residue Screening: Collect near-infrared diffuse reflectance data of the secondary intermediate material, analyze the near-infrared diffuse reflectance data of the secondary intermediate material to obtain chemical state feedback data, analyze the chemical state feedback data, set physical screening control parameter groups, and thus complete the chemical residue screening.

[0012] Preferably, the iterative adjustment of the initial appearance sorting boundary is as follows: the initial appearance sorting boundary includes the initial color sorting threshold boundary, the initial chalky white rejection threshold, the initial particle size sorting lower limit, and the initial particle size sorting upper limit.

[0013] The chromaticity threshold boundary is adjusted to obtain the updated chromaticity threshold boundary.

[0014] Rice grains with a chalky white content exceeding the chalky white rejection threshold are recorded as chalky grains. The number of chalky grains falling into the acceptance region is counted and divided by the total number of rice grains to obtain the chalky white excess rate. Rice grains with an aspect ratio exceeding the upper and lower limits of grain shape sorting are recorded as grain shape abnormalities and divided by the total number of rice grains to obtain the grain shape abnormality rate.

[0015] The misidentification rate of discolored grains, the chalkiness rate, and the abnormal grain shape rate are multiplied by preset weights and then summed to obtain the comprehensive misidentification cost index for this adjustment. This is used to obtain the comprehensive misidentification cost index for each adjustment.

[0016] If the overall misclassification cost index of a certain adjustment is lower than the overall misclassification cost index after the previous adjustment in history, the result of this adjustment is retained and the adjustment continues in the same direction and step size. If the overall misclassification cost index of a certain adjustment is greater than or equal to the overall misclassification cost index after the previous adjustment in history, the step size is halved and the adjustment direction is reversed. When the step size is reduced to the preset minimum step size threshold, the adjustment of the chroma threshold boundary is stopped.

[0017] Based on the adjustment process of the color threshold boundary, the chalk removal threshold, the lower limit of particle shape sorting, and the upper limit of particle shape sorting are adjusted.

[0018] Preferably, the adjustment of the chromaticity threshold boundary is carried out as follows: the current chromaticity threshold closed curve is shifted by a preset step size along the positive direction of the red-green axis, and the center chromaticity coordinates of each chromaticity interval are obtained from the chromaticity distribution histogram. The inside of the closed curve is the acceptance domain, and the outside is the rejection domain. If the center chromaticity coordinates fall into the shifted acceptance domain, all rice grains in the corresponding interval are considered as accepted grains; otherwise, they are considered as rejected grains. Rice grains that fall into the acceptance domain and whose chromaticity value exceeds a preset upper limit are determined as misreceived grains of different colors. The misreceived grains of different colors are obtained by dividing the number of misreceived grains of different colors by the total number of all misreceived grains of different colors.

[0019] The chromaticity threshold closed curve is shifted by a preset step size along the negative direction of the red-green axis, and the heterochromatic particle miscollection rate is recalculated. This yields the heterochromatic particle miscollection rates of the positive, negative, and invariant red-green coordinate boundaries. The boundary corresponding to the minimum heterochromatic particle miscollection rate is taken as the updated chromaticity threshold boundary of the red-green axis.

[0020] Based on the adjustment method of the chromaticity threshold boundary of the red-green axis, the yellow-blue axis is adjusted to obtain the updated chromaticity threshold boundary of the yellow-blue axis.

[0021] The beneficial effects of the present invention are as follows: 1. The appearance quality screening of the present invention, through iterative optimization analysis based on the comprehensive misclassification cost index and combined with the real-time statistical analysis results of raw material quality, can dynamically adjust the acceptance and rejection domains of sorting according to the fluctuation of raw material quality, effectively reducing the misacceptance rate of discolored particles, the chalkiness exceeding the limit rate and the particle shape abnormality rate, and reducing raw material waste while ensuring product quality.

[0022] 2. The optical characteristic screening of this invention introduces dual feedback data of background plate grayscale and light source pulse width modulation duty cycle, which can monitor environmental changes during equipment operation in real time and dynamically compensate and correct the preliminary sorting boundary accordingly. This effectively eliminates the interference of environmental factors on the detection results and significantly improves the stability and reliability of sorting.

[0023] 3. The present invention integrates near-infrared diffuse reflectance online detection technology for chemical residue screening with control parameters of subsequent physical screening. By matching feature vector similarity and calibrating bulk density, the frequency of the variable frequency fan, the phase angle of the eccentric block of the vibrating screen, and the longitudinal tilt angle of the destoning screen are dynamically adjusted, realizing online real-time screening of chemical residues, significantly shortening the detection cycle and effectively improving production efficiency.

[0024] 4. This invention, through multi-level screening and collaborative control and dynamic parameter adjustment mechanism throughout the entire process, effectively balances processing efficiency while strictly ensuring the accuracy of rice grading. It eliminates the need for frequent manual intervention in parameter settings and can adapt to the refined processing needs of different batches, varieties and qualities of rice, thus having broad market application prospects. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention 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 the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 This is a schematic diagram of the implementation steps of the method of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] according to Figure 1 As shown, the present invention provides a collaborative control method for grading and screening of finely processed rice, including the following steps: Step 1, appearance quality screening: collect appearance quality data of rice samples to be processed, perform statistical analysis on the appearance quality data of rice samples to be processed, determine the initial appearance sorting boundary, iteratively adjust the initial appearance sorting boundary to obtain the appearance sorting threshold, generate the appearance quality detection control parameter group, complete the appearance quality screening, and obtain intermediate materials.

[0029] In one specific embodiment, the acquisition process for collecting the appearance quality data of the rice sample to be processed is as follows: the rice to be processed is fed into the detection channel at a set flow rate, and the multispectral line scanning imaging system acquires grayscale images of each rice grain at at least four characteristic wavelengths while the material is in motion.

[0030] It should be noted that the characteristic wavelengths selected cover the bands with the greatest difference in reflectance between chalky and non-chalky areas in the visible to near-infrared range, as well as the characteristic wavelengths of the varieties.

[0031] For each frame of scanned image, dark current correction, flat field correction, and standard whiteboard normalization are performed sequentially.

[0032] The corrected image uses an adaptive threshold segmentation algorithm: the image is traversed with a window of a preset size, a gray-level histogram is calculated in each window, and the valley gray value between the two peaks of the histogram is used as the binarization threshold of the window. If there is no bimodal feature in the window, the interpolation result of the threshold of the adjacent window is used.

[0033] For the effective rice grain area, extract the gray values ​​of all pixels within the area at four wavelengths, calculate the arithmetic mean of the gray values ​​at each wavelength, and obtain the four-dimensional spectral response vector of the rice grain. Transform the four-dimensional spectral response vector to the CIELAB color space through a chromaticity transformation matrix to obtain the chromaticity coordinates of the rice grain (black and white lightness coordinates, red and green coordinates, yellow and blue coordinates).

[0034] It should be noted that the calibration process of the chromaticity transformation matrix is ​​as follows: select several standard color plates with known CIELAB chromaticity values, collect the spectral response vectors of each color plate under a multispectral imaging system, use the spectral response vectors as independent variables and the known chromaticity values ​​as dependent variables, and solve the transformation coefficient matrix through multiple linear regression.

[0035] The band image with the largest difference in reflectance between chalky and non-chalky areas is selected as the chalky identification feature image. The average gray value of the rice grain region on the chalky identification feature image is calculated. A chalky determination coefficient with a preset value greater than 1 is used to check the gray value of each pixel in the region one by one. If the gray value is greater than the product of the chalky determination coefficient and the average gray value, the pixel is marked as a chalky pixel. The number of chalky pixels and the total number of pixels in the rice grain region are counted. The number of chalky pixels is divided by the total number of pixels in the rice grain region to obtain the proportion of the chalky region.

[0036] The minimum bounding rectangle of the rice grain region is fitted to obtain the projected length and width of the rice grain in the image coordinate system. The length is divided by the width to obtain the aspect ratio.

[0037] The above process is performed on a preset number of rice grains one by one to obtain the appearance quality data of the rice sample to be processed.

[0038] In one specific embodiment, the statistical analysis of the appearance quality data of the rice sample to be processed is carried out, and the specific data collection process is as follows: the appearance quality data of the rice sample to be processed includes the color coordinates of each rice grain, the proportion of chalky area, and the aspect ratio.

[0039] The color coordinates of a standard rice grain are preset, and the color difference value is calculated for each grain to obtain the color difference value of each rice grain. The mean, standard deviation and skewness of the color difference value distribution are statistically obtained. A color difference distribution histogram is established with the preset color difference value interval as the group interval. At the same time, the mean and quantile of the chalky area proportion are statistically obtained to establish a chalky area proportion distribution histogram. The mean and coefficient of variation of the length-width ratio of the grain are statistically obtained to establish a grain shape distribution histogram.

[0040] Red-green and yellow-blue coordinates are extracted from the chromaticity coordinates and recorded as analysis coordinates. The analysis coordinates of the standard rice grain and each individual rice grain are obtained. Using the analysis coordinates of the standard rice grain as the center, the maximum and minimum values ​​of the red-green and yellow-blue coordinate axes are obtained according to the preset allowable color difference offset. Positive and negative offset limit points are set, and the positive and negative offset limit points are connected in sequence to form a closed curve, thus obtaining the initial chromaticity sorting threshold boundary.

[0041] Starting from the right end of the chalkiness distribution histogram, intervals are set sequentially to the left. The cumulative frequency of each interval is counted. The cumulative frequency of each interval is divided by the total frequency to obtain the cumulative frequency ratio of each interval. When the cumulative frequency ratio of a certain interval is greater than the threshold, the left boundary value of the corresponding interval is the initial chalkiness removal threshold.

[0042] Centered on the average aspect ratio, the initial particle size sorting lower and upper limits are set according to the preset aspect ratio allowable range.

[0043] In one specific embodiment, the iterative adjustment of the initial appearance sorting boundary is as follows: the initial appearance sorting boundary includes the initial color sorting threshold boundary, the initial chalky white rejection threshold, the initial particle size sorting lower limit, and the initial particle size sorting upper limit.

[0044] The chromaticity threshold boundary is adjusted to obtain the updated chromaticity threshold boundary.

[0045] Rice grains with a chalky white content exceeding the chalky white rejection threshold are recorded as chalky grains. The number of chalky grains falling into the acceptance region is counted and divided by the total number of rice grains to obtain the chalky white excess rate. Rice grains with an aspect ratio exceeding the upper and lower limits of grain shape sorting are recorded as grain shape abnormalities and divided by the total number of rice grains to obtain the grain shape abnormality rate.

[0046] The misidentification rate of discolored grains, the chalkiness rate, and the abnormal grain shape rate are multiplied by preset weights and then summed to obtain the comprehensive misidentification cost index for this adjustment. This is used to obtain the comprehensive misidentification cost index for each adjustment.

[0047] If the overall misclassification cost index of a certain adjustment is lower than the overall misclassification cost index after the previous adjustment in history, the result of this adjustment is retained and the adjustment continues in the same direction and step size. If the overall misclassification cost index of a certain adjustment is greater than or equal to the overall misclassification cost index after the previous adjustment in history, the step size is halved and the adjustment direction is reversed. When the step size is reduced to the preset minimum step size threshold, the adjustment of the chroma threshold boundary is stopped.

[0048] Based on the adjustment process of the color threshold boundary, the chalk removal threshold, the lower limit of particle shape sorting, and the upper limit of particle shape sorting are adjusted.

[0049] In one specific embodiment, the adjustment of the chromaticity threshold boundary is carried out as follows: the current chromaticity threshold closed curve is shifted by a preset step size along the positive direction of the red-green axis, and the center chromaticity coordinates of each color difference interval are obtained from the color difference distribution histogram. The inside of the closed curve is the acceptance domain, and the outside is the rejection domain. If the center chromaticity coordinates fall into the shifted acceptance domain, all rice grains in the corresponding interval are considered as accepted grains; otherwise, they are considered as rejected grains. Rice grains that fall into the acceptance domain and whose color difference value exceeds a preset upper limit are determined as misreceived grains of different colors. The misreceived grains of different colors are obtained by dividing the number of misreceived grains of different colors by the total number of all misreceived grains of different colors.

[0050] The chromaticity threshold closed curve is shifted by a preset step size along the negative direction of the red-green axis, and the heterochromatic particle miscollection rate is recalculated. This yields the heterochromatic particle miscollection rates of the positive, negative, and invariant red-green coordinate boundaries. The boundary corresponding to the minimum heterochromatic particle miscollection rate is taken as the updated chromaticity threshold boundary of the red-green axis.

[0051] Based on the adjustment method of the chromaticity threshold boundary of the red-green axis, the yellow-blue axis is adjusted to obtain the updated chromaticity threshold boundary of the yellow-blue axis.

[0052] In one specific embodiment, the adjustment of the chalk removal threshold, the lower limit of particle size sorting, and the upper limit of particle size sorting is carried out as follows: the chalk removal threshold is adjusted by a preset step size in the increasing direction, and the chalk excess rate is recalculated to obtain the chalk excess rate of each chalk adjustment. If the chalk excess rate of a certain chalk adjustment is less than the chalk excess rate of the previous chalk adjustment, the direction is retained and the adjustment continues with the same step size. If the chalk excess rate of a certain chalk adjustment is greater than or equal to the chalk excess rate of the previous chalk adjustment, the direction is reversed and the step size is halved. The chalk adjustment is repeated until the step size is reduced to the preset minimum step size to obtain the corrected chalk removal threshold.

[0053] The upper and lower limits of particle size are expanded outward by a preset step size, and the particle size anomaly rate is recalculated to obtain the particle size anomaly rate for each particle size adjustment. If the particle size anomaly rate of a certain particle size adjustment is less than the particle size anomaly rate of the previous particle size adjustment, the expansion continues. If the particle size anomaly rate of a certain particle size adjustment is greater than or equal to the particle size anomaly rate of the previous particle size adjustment, the particle size is contracted inward and the step size is halved. This process is repeated until the step size is reduced to the preset minimum step size to obtain the corrected particle size sorting boundary.

[0054] In one specific embodiment, the generation process of the appearance quality inspection control parameter set is as follows: Based on the corrected chromaticity threshold boundary, all chromaticity coordinate values ​​contained in the outer region of the closed curve are recorded as the trigger chromaticity coordinate interval; based on the chalk removal threshold, the gray value corresponding to the chalk removal threshold is obtained and recorded as the trigger gray value threshold; based on the particle shape sorting boundary, the lower and upper limits of particle shape sorting are obtained and recorded as the upper and lower limits of the trigger aspect ratio interval of the particle shape sorting spray valve, thus obtaining the trigger particle shape interval threshold; the trigger chromaticity coordinate interval, the trigger gray value threshold, and the trigger particle shape interval threshold are recorded as the appearance quality inspection control parameter set.

[0055] When the chromaticity coordinates of a rice grain fall into the trigger chromaticity coordinate range, or the proportion of the chalky area of ​​the rice grain is greater than the trigger grayscale threshold, or the aspect ratio of the rice grain exceeds the trigger particle size range threshold, the high-pressure spray valve is triggered when any of the conditions are met. The high-pressure spray valve blows the rice grain out of the material flow, and the intermediate material after screening based on appearance quality is obtained.

[0056] Step 2, Optical Property Screening: Collect the optical properties of the intermediate materials, perform statistical analysis on the optical properties of the intermediate materials to obtain the preliminary sorting boundary, then collect the actual execution parameter feedback data, and based on the actual execution parameter feedback data, compensate and correct the preliminary sorting boundary to generate a light transmission detection control parameter set, complete the optical property screening, and obtain the secondary intermediate materials.

[0057] In one specific embodiment, the optical properties of the intermediate material are collected, and the specific collection process is as follows: the optical properties of the intermediate material include the apparent refractive index of each intermediate rice grain and the absorbance at each characteristic wavelength. The characteristic wavelengths are arranged in order to obtain the absorbance vector of each intermediate rice grain.

[0058] An array of total internal reflection critical angle refractive index sensors is installed on the material conveying channel. This array consists of multiple sapphire right-angle prisms arranged sequentially along the material flow direction to collect the apparent refractive index of rice grains.

[0059] It should be noted that the specific process for collecting the apparent refractive index of rice grains is as follows: Under the action of gravity, rice grains slide one by one along an inclined slide over the upper surface of the prism, forming local optical contact with the prism's measurement surface; an LED light source illuminates the prism-rice grain interface from the back of the prism at a preset divergence angle, and the reflected light is received by a linear CCD detector after exiting through the other side of the prism. The position of each pixel on the linear CCD corresponds to a specific incident angle; for the gray value sequence of each pixel output by the linear CCD, the first-order difference of the gray value of adjacent pixels is calculated, and the pixel position corresponding to the maximum absolute value of the difference is located. This position is converted into a critical angle measurement value through a pre-calibrated pixel number-angle relationship curve; according to Snell's law, the apparent refractive index of the rice grain is calculated one grain at a time from the known refractive index of the prism and the measured critical angle.

[0060] A near-infrared transmission spectroscopy detection module is set up immediately downstream of the refractive index sensor array. The near-infrared LED array provides near-infrared light with multiple characteristic wavelengths, including the absorption peak of OH bonds in water and the overtone absorption peak of NH bonds in proteins. After collimation, the light is irradiated through a single grain of rice. The transmitted light is received by a photodetector and converted into voltage. The absorbance of each wavelength is calculated, and each grain of rice obtains an apparent refractive index value and a vector containing the absorbance of several characteristic wavelengths. A preset number of grains of rice data are collected, and the refractive index values ​​are arranged in the order of passage to form a refractive index sequence. The absorbance vectors are arranged row by row to form an absorbance matrix.

[0061] In one specific embodiment, the statistical analysis of the optical properties of the intermediate material is carried out as follows: the optical properties of the intermediate material include the apparent refractive index of each intermediate rice grain and the absorbance at each characteristic wavelength. The absorbance vector of each intermediate rice grain is obtained by arranging the characteristic wavelengths in order.

[0062] The apparent refractive indices are arranged in order of passage to form a refractive index sequence, and the absorbance vectors are arranged row by row to form an absorbance matrix. The refractive index sequence is binned with a fixed group interval, and the frequency of grains falling into each refractive index interval is counted to establish a refractive index frequency distribution histogram. The histogram is smoothed with Gaussian kernel density to obtain a smooth density curve. The mode of the maximum density refractive index and the half-peak width are collected. With the mode of the maximum density refractive index as the center, the half-peak width is extended to both sides by a preset multiple to obtain the preliminary lower and upper limits of refractive index sorting.

[0063] Principal component analysis was performed on the absorbance matrix to obtain the absorbance ratio sequence and the first principal component score of each intermediate rice grain. Then, upper and lower limits for the first principal component score and the upper and lower limits for the absorbance ratio were set.

[0064] It should be noted that the mean and standard deviation of the first principal component scores of the target product grade standard samples are read, and a preset multiple of the mean minus the standard deviation is used as the lower limit of acceptance for the first principal component scores, and a preset multiple of the mean plus the standard deviation is used as the upper limit of acceptance for the first principal component scores; the mean and standard deviation of the absorbance ratios of the standard samples are read, and a preset multiple of the mean of the absorbance ratio minus the standard deviation is used as the lower limit of acceptance for the absorbance ratio, and a preset multiple of the mean of the absorbance ratio plus the standard deviation is used as the upper limit of acceptance for the absorbance ratio.

[0065] The upper and lower limits of refractive index sorting, the upper and lower limits of first principal component score acceptance, and the upper and lower limits of absorbance ratio acceptance together constitute the preliminary sorting boundary.

[0066] In one specific embodiment, the compensation and correction of the initial sorting boundary is carried out as follows: the actual execution parameter feedback data includes the background grayscale feedback value and the pulse width modulation duty cycle feedback value of each band of light source.

[0067] Based on the pre-calibrated background grayscale-refractive index bias lookup table curve, the refractive index bias is obtained by looking up the background grayscale feedback value. Based on the pre-calibrated light source duty cycle-absorbance baseline drift coefficient, the absorbance drift is obtained by looking up the light source pulse width modulation duty cycle feedback value for each band.

[0068] It should be noted that the pre-calibrated background grayscale-refractive index bias lookup table curve is constructed as follows: under different background grayscale settings, the mode of refractive index of the same batch of standard rice grains is measured, and the one-to-one correspondence between the background grayscale value and the offset of the mode of refractive index is recorded; the difference ΔG between the current background grayscale feedback value and the standard grayscale value is used to perform linear interpolation between two adjacent calibration points of the lookup table curve to obtain the refractive index bias.

[0069] Method for constructing the pre-calibrated light source duty cycle-absorbance baseline drift coefficient: Under different light source duty cycles, measure the absorbance of the standard reflector, divide the absorbance offset by the duty cycle change, and take the average of multiple measurements; multiply the difference between the current duty cycle feedback value and the standard duty cycle by this coefficient to obtain the absorbance drift.

[0070] Add the refractive index bias to the lower and upper limits of the initial sorting boundary to obtain the updated lower and upper limits of the refractive index sorting. Add the absorbance drift to the lower and upper limits of the initial first principal component score to obtain the updated upper and lower limits of the first principal component score. Add the absorbance drift to the lower and upper limits of the absorbance ratio to obtain the updated upper and lower limits of the absorbance ratio. This gives us the updated sorting boundary.

[0071] In one specific embodiment, the generation process of the light transmission detection control parameter group is as follows: the updated sorting boundaries are arranged in a preset order to obtain the sorting boundary vector, and the similarity between the vector and the voltage of each refractive index comparator is calculated. The voltage of the refractive index comparator with the highest similarity is recorded as the voltage of the current refractive index comparator.

[0072] Based on the current voltage of the refractive index comparator, as each intermediate grain passes through, the first principal component score and absorbance ratio are collected in real time by the refractive index comparator. If the first principal component score of a certain intermediate grain does not fall within the range between the upper and lower limits of the updated first principal component score, or if the absorbance ratio collected in real time does not fall within the range between the upper and lower limits of the updated absorbance ratio, the intermediate grain is removed by the high-pressure spray valve.

[0073] Step 3: Chemical Residue Screening: Collect near-infrared diffuse reflectance data of the secondary intermediate material, analyze the near-infrared diffuse reflectance data of the secondary intermediate material to obtain chemical state feedback data, analyze the chemical state feedback data, set physical screening control parameter groups, and thus complete the chemical residue screening.

[0074] In one specific embodiment, the near-infrared diffuse reflectance data of the secondary intermediate materials is collected in the following specific process: the near-infrared diffuse reflectance data of the secondary intermediate materials includes the area ratio of cellulose characteristic peaks, the area ratio of lignin characteristic peaks, the relative intensity of silicon absorption bands, and the characteristic absorption values ​​of fatty acids for each secondary intermediate rice.

[0075] A near-infrared diffuse reflectance spectroscopy acquisition probe is installed directly above the material conveyor belt before the shell separation unit. The probe maintains a constant distance from the surface of the conveyor belt. A halogen tungsten lamp light source is installed inside the probe, with a spectral range covering visible light to near-infrared. The light outlet illuminates the secondary intermediate material on the conveyor belt at a preset angle. The diffuse reflected light from the material surface is focused by a collecting lens and transmitted to the spectrometer via optical fiber. The wavelength acquisition range of the spectrometer is set to the first harmonic and combination absorption regions of hydrogen-containing groups. The spectrometer performs a preset number of scans in each acquisition cycle and takes the average to obtain the original diffuse reflectance spectrum. Each raw spectrum is sequentially subjected to standard normal transformation and Savitzky-Golay first-order derivative filtering to obtain the preprocessed spectrum; On the preprocessed spectrum, the characteristic absorption peaks of cellulose, lignin, silicon-oxygen bonds, and fatty acid ester carbonyl groups were located in their respective bands. For each characteristic band, the spectral values ​​of each wavelength point within the band were multiplied by the wavelength interval, and the sum was obtained to obtain the area of ​​each characteristic band. The area of ​​each characteristic band was then divided by the total area under the full-band spectral curve to obtain the area ratio of cellulose characteristic peaks, the area ratio of lignin characteristic peaks, the relative intensity of the silicon absorption band, and the characteristic absorption value of fatty acids.

[0076] It should be noted that the center wavelength and band width of each band are predetermined based on the known characteristic peak positions of the standard sample.

[0077] In one specific embodiment, the analysis of the near-infrared diffuse reflectance data of the secondary intermediate material is carried out as follows: after calculating the mean value of the near-infrared diffuse reflectance data of the secondary intermediate material, the data are arranged in a preset order to obtain the near-infrared diffuse reflectance feature vector of the secondary intermediate material. The similarity between the vector and the near-infrared diffuse reflectance feature vector of each physical screening control parameter is calculated, and the physical screening control parameter with the highest similarity is recorded as the primary physical screening control parameter.

[0078] In one specific embodiment, the setting of the physical screening control parameter group is specifically set as follows: the primary physical screening control parameters include the initial value of the frequency of the variable frequency fan, the initial value of the phase angle of the eccentric block of the vibrating screen, and the initial value of the longitudinal inclination angle of the destoning screen surface.

[0079] The bulk density of the material in a fixed container is measured and calibrated under different shell residue levels. An axis fitting curve is made to obtain the shell residue-bulk density relationship curve. The current bulk density of the material is obtained by looking up the sum of the characteristic peak area ratios of cellulose and lignin. The standard bulk density is subtracted to obtain the bulk density difference. The difference is multiplied by the preset wind speed compensation coefficient to obtain the frequency correction amount of the variable frequency fan. The frequency correction amount of the variable frequency fan is added to the initial value of the variable frequency fan frequency to obtain the frequency of the variable frequency fan.

[0080] The shell residue-air resistance coefficient relationship curve was calibrated by shell residue gradient test, and the sum of the characteristic peak area ratios of cellulose and lignin was mapped to the air resistance coefficient of the material in the airflow. The additional correction amount of the phase angle of the vibrating screen eccentric block was obtained by multiplying the difference between the air resistance coefficient and the standard value by the preset phase angle compensation coefficient. The phase angle of the vibrating screen eccentric block was obtained by adding the additional correction amount of the phase angle of the vibrating screen eccentric block to the initial value of the phase angle of the vibrating screen eccentric block. By measuring the critical sliding angle of rice grains on an adjustable inclined plane under different fatty acid contents and calculating the friction coefficient calibration, the relationship curve between the characteristic absorption value of fatty acids and the friction coefficient is obtained. The friction coefficient is obtained by interpolation based on the current characteristic absorption value of fatty acids. It is compared with the standard friction coefficient, and the difference is multiplied by the preset inclination angle compensation coefficient to obtain the additional correction amount of the longitudinal inclination angle of the destoning screen. The longitudinal inclination angle of the destoning screen is obtained by adding the initial value of the longitudinal inclination angle of the destoning screen to the additional correction amount of the longitudinal inclination angle of the destoning screen. Thus, the physical screening control parameter set is obtained.

[0081] It should be noted that, by operating with the physical screening control parameter group, the secondary intermediate materials are subjected to air separation, vibration screening and destoning to obtain finished rice. The finished materials discharged after hull separation are sampled and the near-infrared diffuse reflectance data acquisition process is repeated to obtain the characteristic peak area ratio of cellulose, characteristic peak area ratio of lignin and characteristic absorption value of fatty acids for re-examination.

[0082] If the area ratio of cellulose characteristic peaks exceeds the threshold or the area ratio of lignin characteristic peaks exceeds the threshold, the frequency is increased by a preset step size within the adjustment stroke of the variable frequency fan frequency. After each increase, the frequency is rechecked until the residue meets the standard or the frequency reaches the upper limit allowed by the frequency converter. If the residue still exceeds the standard after the frequency reaches the upper limit, the phase angle is increased by a preset step size within the adjustment stroke of the phase angle of the eccentric block of the vibrating screen. After each increase, the frequency is rechecked until the residue meets the standard or the phase angle reaches the upper limit.

[0083] If the characteristic absorption value of fatty acids exceeds the threshold, the tilt angle is adjusted within the adjustment stroke of the longitudinal tilt angle of the destoning screen by a preset step size. The adjustment direction is to increase the tilt angle to accelerate the material sliding speed. After each adjustment, the screen is re-inspected until the standard is met or the tilt angle reaches the upper limit.

[0084] The multispectral line scanning imaging technology, dark current correction and flat field correction methods, adaptive threshold segmentation algorithm, colorimetric transformation matrix calibration method based on multiple linear regression, total internal reflection critical angle refractive sensing technology, near-infrared transmission spectral detection technology, Gaussian kernel density smoothing method, principal component analysis method, standard normal variable transformation and Savitzky-Golay first derivative filtering method, connected component labeling algorithm, minimum bounding rectangle fitting method, first-order difference peak finding algorithm, engineering application method of Snell's law, Euclidean distance or cosine similarity calculation method, and parameter mapping table establishment method based on orthogonal experiment described in this invention are all existing technologies and can be found on the Internet and related public literature, so they will not be described in detail here.

[0085] The chalkiness determination coefficient, preset color difference value interval, allowable color difference offset, aspect ratio tolerance range, iterative adjustment step size, minimum step size threshold, weight of each item in the comprehensive misclassification cost index, preset multiplier of half-peak width, wind speed compensation coefficient, phase angle compensation coefficient, tilt angle compensation coefficient, cellulose characteristic peak area ratio threshold, lignin characteristic peak area ratio threshold, fatty acid characteristic absorbance value threshold, preset upper limit of color difference value, upper limit of allowable chalky particle proportion for target product grade, preset multiple of the first principal component score receiving upper and lower limits, preset multiple of the absorbance ratio receiving upper and lower limits, and preset ratio of refractive index measurement signal change, etc., mentioned in this invention are all preset values, which are set by the staff according to the actual production requirements. For example: the chalkiness judgment coefficient is set to 1.2, the preset color difference value interval is set to 0.5, the allowable color difference offset is set to 2.5, the aspect ratio tolerance range is set to [2.5, 3.5], the initial value of the iterative adjustment step size is set to 0.1, the minimum step size threshold is set to 0.01, the weight of the miscollection rate of discolored particles is set to 0.4, the weight of the chalkiness exceedance rate is set to 0.3, the weight of the particle shape anomaly rate is set to 0.3, the preset multiplier of the half-peak width is set to 2, the wind speed compensation coefficient is set to 0.5, and the phase angle compensation coefficient is set to... The following parameters are set: tilt angle compensation coefficient (0.1), cellulose characteristic peak area ratio threshold (0.05), lignin characteristic peak area ratio threshold (0.05), fatty acid characteristic absorbance value threshold (0.1), color difference value preset upper limit (2.5), allowable chalky grain ratio for target product grade (5%), first principal component score receiving upper and lower limit preset multiple (2), absorbance ratio receiving upper and lower limit preset multiple (2), and refractive index measurement signal change preset ratio (30%).

[0086] The examples described in this invention are not limited to the specific embodiments listed above. The examples are merely illustrative to facilitate understanding of the invention and do not constitute a limitation on the scope of protection of this invention. Any modifications, equivalent substitutions, etc., made within the spirit and principles of this invention should be included within the scope of protection.

[0087] The above description is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in this specification, they should all fall within the protection scope of the present invention.

Claims

1. A method for coordinated control of grading and screening in the fine processing of rice, characterized in that, Includes the following steps: Step 1: Appearance Quality Screening: Collect appearance quality data of rice samples to be processed, perform statistical analysis on the appearance quality data of rice samples to be processed, determine the initial appearance sorting boundary, iteratively adjust the initial appearance sorting boundary to obtain the appearance sorting threshold, generate the appearance quality detection control parameter group, complete the appearance quality screening, and obtain the intermediate material. Step 2, Optical property screening: Collect the optical properties of intermediate materials, perform statistical analysis on the optical properties of intermediate materials to obtain the preliminary sorting boundary, then collect the actual execution parameter feedback data, and based on the actual execution parameter feedback data, compensate and correct the preliminary sorting boundary, generate a light transmission detection control parameter group, complete the optical property screening, and obtain the secondary intermediate material; Step 3: Chemical Residue Screening: Collect near-infrared diffuse reflectance data of the secondary intermediate material, analyze the near-infrared diffuse reflectance data of the secondary intermediate material to obtain chemical state feedback data, analyze the chemical state feedback data, set physical screening control parameter groups, and thus complete the chemical residue screening.

2. The method for coordinated control of rice fine processing grading and screening according to claim 1, characterized in that, The statistical analysis of the appearance quality data of the rice samples to be processed was carried out, and the specific data collection process is as follows: The appearance quality data of the rice samples to be processed include the color coordinates of each grain, the proportion of chalky area, and the aspect ratio. The color coordinates of a pre-set standard rice grain are used to calculate the color difference value for each grain. The mean, standard deviation, and skewness of the color difference value distribution are statistically analyzed. A color difference distribution histogram is established with the pre-set color difference value interval as the group interval. At the same time, the mean and quantile of the chalky area proportion are statistically analyzed to establish a chalky area proportion distribution histogram. The mean and coefficient of variation of the length-width ratio of the grain are statistically analyzed to establish a grain shape distribution histogram. Red-green and yellow-blue coordinates are extracted from the chromaticity coordinates and recorded as analysis coordinates. The analysis coordinates of the standard rice grain and each rice grain are obtained in this way. With the analysis coordinates of the standard rice grain as the center, the maximum and minimum values ​​of the red-green and yellow-blue coordinate axes are obtained according to the preset allowable color difference offset. Positive offset limit points and negative offset limit points are set. The positive offset limit points and negative offset limit points are connected in sequence to form a closed curve, and the initial chromaticity sorting threshold boundary is obtained. Starting from the right end of the chalkiness distribution histogram, intervals are set sequentially from left to right. The cumulative frequency of each interval is counted. The cumulative frequency of each interval is divided by the total frequency to obtain the cumulative frequency ratio of each interval. When the cumulative frequency ratio of a certain interval is greater than the threshold, the left boundary value of the corresponding interval is the initial chalkiness removal threshold. Centered on the average aspect ratio, the initial particle size sorting lower and upper limits are set according to the preset aspect ratio allowable range.

3. The method for coordinated control of rice fine processing grading and screening according to claim 2, characterized in that, The initial appearance sorting boundary is iteratively adjusted, and the specific adjustment process is as follows: The initial appearance sorting boundaries include the initial color sorting threshold boundary, the initial chalky white rejection threshold, the initial particle shape sorting lower limit, and the initial particle shape sorting upper limit; The chromaticity threshold boundary is adjusted to obtain the updated chromaticity threshold boundary; Rice grains with a chalky white content exceeding the chalky white rejection threshold are recorded as chalky grains. The number of chalky grains that fall into the acceptance region is counted and divided by the total number of rice grains to obtain the chalky white excess rate. Rice grains with an aspect ratio exceeding the upper and lower limits of grain shape sorting are recorded as grain shape abnormalities and divided by the total number of rice grains to obtain the grain shape abnormality rate. The misidentification rate of discolored grains, the chalkiness rate, and the abnormal grain shape rate are multiplied by preset weights and then summed to obtain the comprehensive misidentification cost index for this adjustment. This is used to obtain the comprehensive misidentification cost index for each adjustment. If the overall misclassification cost index of a certain adjustment is lower than the overall misclassification cost index after the previous adjustment in history, the result of this adjustment is retained and the adjustment continues in the same direction and step size. If the overall misclassification cost index of a certain adjustment is greater than or equal to the overall misclassification cost index after the previous adjustment in history, the step size is halved and the adjustment direction is reversed. When the step size is reduced to the preset minimum step size threshold, the adjustment of the chroma threshold boundary is stopped. Based on the adjustment process of the color threshold boundary, the chalk removal threshold, the lower limit of particle shape sorting, and the upper limit of particle shape sorting are adjusted.

4. The method for coordinated control of rice fine processing grading and screening according to claim 3, characterized in that, The adjustment process for the chromaticity threshold boundary is as follows: The current chromaticity threshold closed curve is shifted along the positive direction of the red-green axis by a preset step size. The center chromaticity coordinates of each chromaticity interval are obtained from the chromaticity distribution histogram. The inside of the closed curve is the acceptance region, and the outside is the rejection region. If the center chromaticity coordinates fall within the shifted acceptance region, all rice grains in the corresponding interval are considered as accepted grains; otherwise, they are considered as rejected grains. Rice grains that fall within the acceptance region and whose chromaticity value exceeds the preset upper limit are judged as misreceived grains of different colors. The misreceived grain rate of different colors is obtained by dividing the number of misreceived grains of different colors by the total number of all misreceived grains. The chromaticity threshold closed curve is shifted by a preset step size along the negative direction of the red-green axis, and the heterochromatic particle miscollection rate is recalculated. This yields the heterochromatic particle miscollection rates of the red-green coordinate positive direction boundary, negative direction boundary, and invariant boundary. The boundary corresponding to the minimum heterochromatic particle miscollection rate is taken as the updated chromaticity threshold boundary of the red-green axis. Based on the adjustment method of the chromaticity threshold boundary of the red-green axis, the yellow-blue axis is adjusted to obtain the updated chromaticity threshold boundary of the yellow-blue axis.

5. The method for coordinated control of grading and screening in fine rice processing according to claim 3, characterized in that, The specific process for generating the appearance quality inspection control parameter set is as follows: Based on the corrected chromaticity threshold boundary, all chromaticity coordinate values ​​contained in the outer region of the closed curve are recorded as the trigger chromaticity coordinate interval. Based on the chalk removal threshold, the gray value corresponding to the chalk removal threshold is obtained and recorded as the trigger gray value threshold. Based on the particle shape sorting boundary, the lower and upper limits of particle shape sorting are obtained and recorded as the upper and lower limits of the trigger aspect ratio interval of the particle shape sorting spray valve. The trigger particle shape interval threshold is obtained. The trigger chromaticity coordinate interval, the trigger gray value threshold, and the trigger particle shape interval threshold are recorded as the appearance quality inspection control parameter group. When the chromaticity coordinates of a rice grain fall into the trigger chromaticity coordinate range, or the proportion of the chalky area of ​​the rice grain is greater than the trigger grayscale threshold, or the aspect ratio of the rice grain exceeds the trigger particle size range threshold, the high-pressure spray valve is triggered when any of the conditions are met. The high-pressure spray valve blows the rice grain out of the material flow, and the intermediate material after screening based on appearance quality is obtained.

6. The method for coordinated control of grading and screening in fine rice processing according to claim 1, characterized in that, The statistical analysis of the optical properties of the intermediate materials is described in the following process: The optical properties of the intermediate material include the apparent refractive index of each intermediate rice grain and the absorbance at each characteristic wavelength. By arranging the characteristic wavelengths in order, the absorbance vector of each intermediate rice grain is obtained. The apparent refractive indices are arranged in order of passage to form a refractive index sequence, and the absorbance vectors are arranged in rows to form an absorbance matrix. The refractive index sequence is binned with a fixed group interval, and the frequency of grains falling into each refractive index interval is counted to establish a refractive index frequency distribution histogram. The histogram is smoothed by Gaussian kernel density smoothing to obtain a smooth density curve. The mode of the maximum density refractive index and the half-peak width are collected. With the mode of the maximum density refractive index as the center, the half-peak width is extended to both sides by a preset multiple to obtain the preliminary lower and upper limits of refractive index sorting. Principal component analysis was performed on the absorbance matrix to obtain the absorbance ratio sequence and the first principal component score of each intermediate rice grain. Then, the upper and lower limits of acceptance for the first principal component score and the upper and lower limits of acceptance for the absorbance ratio were set. The upper and lower limits of refractive index sorting, the upper and lower limits of first principal component score acceptance, and the upper and lower limits of absorbance ratio acceptance together constitute the preliminary sorting boundary.

7. The method for coordinated control of rice fine processing grading and screening according to claim 6, characterized in that, The compensation and correction process for the initial sorting boundary is as follows: The actual execution parameter feedback data includes the background grayscale feedback value and the pulse width modulation duty cycle feedback value of each band of light source; Based on the pre-calibrated background grayscale-refractive index bias lookup table curve, the refractive index bias is obtained by looking up the table using the background grayscale feedback value. Based on the pre-calibrated light source duty cycle-absorbance baseline drift coefficient, the absorbance drift is obtained by looking up the table using the light source pulse width modulation duty cycle feedback value for each band. Add the refractive index bias to the lower and upper limits of the initial sorting boundary to obtain the updated lower and upper limits of the refractive index sorting. Add the absorbance drift to the lower and upper limits of the initial first principal component score to obtain the updated upper and lower limits of the first principal component score. Add the absorbance drift to the lower and upper limits of the absorbance ratio to obtain the updated upper and lower limits of the absorbance ratio. This gives us the updated sorting boundary.

8. The method for coordinated control of grading and screening in fine rice processing according to claim 7, characterized in that, The specific process for generating the light transmittance detection control parameter set is as follows: The updated sorting boundaries are arranged in a preset order to obtain the sorting boundary vector. The similarity between the vector and the voltage of each refractive index comparator is calculated, and the voltage of the refractive index comparator with the highest similarity is recorded as the voltage of the current refractive index comparator. Based on the current voltage of the refractive index comparator, as each intermediate grain passes through, the first principal component score and absorbance ratio are collected in real time by the refractive index comparator. If the first principal component score of a certain intermediate grain does not fall within the range between the upper and lower limits of the updated first principal component score, or if the absorbance ratio collected in real time does not fall within the range between the upper and lower limits of the updated absorbance ratio, the intermediate grain is removed by the high-pressure spray valve.

9. The method for coordinated control of rice fine processing grading and screening according to claim 1, characterized in that, The near-infrared diffuse reflectance data of the secondary intermediate material were analyzed, and the specific analysis process is as follows: Near-infrared diffuse reflectance data of secondary intermediate materials include the area ratio of cellulose characteristic peaks, the area ratio of lignin characteristic peaks, the relative intensity of silicon absorption bands, and the characteristic absorbance values ​​of fatty acids for each secondary intermediate rice. After calculating the mean value of the near-infrared diffuse reflectance data of the secondary intermediate materials, they are arranged in a preset order to obtain the near-infrared diffuse reflectance feature vector of the secondary intermediate materials. The similarity between this vector and the near-infrared diffuse reflectance feature vector of each physical screening control parameter is calculated, and the physical screening control parameter with the highest similarity is recorded as the primary physical screening control parameter.

10. The method for coordinated control of grading and screening in fine rice processing according to claim 9, characterized in that, The specific process for setting the physical screening control parameter group is as follows: The initial physical screening control parameters include the initial value of the frequency of the variable frequency fan, the initial value of the phase angle of the eccentric block of the vibrating screen, and the initial value of the longitudinal inclination angle of the destoning screen surface. The bulk density of the material in a fixed container is measured and calibrated under different shell residue levels. An axis fitting curve is made to obtain the shell residue-bulk density relationship curve. The current bulk density of the material is obtained by looking up the table using the sum of the characteristic peak area ratios of cellulose and lignin. The standard bulk density is subtracted to obtain the bulk density difference. The difference is multiplied by the preset wind speed compensation coefficient to obtain the frequency correction amount of the variable frequency fan. The frequency correction amount of the variable frequency fan is added to the initial value of the variable frequency fan frequency to obtain the frequency of the variable frequency fan. Based on the method for obtaining the frequency of the variable frequency fan, the phase angle of the eccentric block of the vibrating screen and the longitudinal inclination angle of the destoning screen surface are obtained respectively, thereby obtaining the physical screening control parameter set.

Images