Paddy quality index full-process detection method and device

Abstract

The invention relates to the technical field of polished rice rate detection, in particular to a rice quality index whole-process detection method and device, and the method comprises the following steps: carrying out the image recognition of each detection link according to the rice quality index detection whole process, and carrying out the image recognition of each detection link according to the image collected by each link; and respectively calculating an effective feeding grain index, an imperfect grain ratio index, a rice grain quality classification result, a head rice rate index and a yellow grain ratio. The position and form of each rice grain are recognized through high-precision image processing, effective feeding grain indexes input into a rice hulling chamber are screened, infrared transmission intensity detection of brown rice directly reflects the structural characteristics of the rice grains, the real-time quality evaluation method optimizes the conversion process from the brown rice to polished rice, optimization of the rice quality and yield is ensured, and the quality of the rice grains is improved. The real-time monitoring on the surface brightness and texture change in the milling process is beneficial to adjusting the milling process, the product quality is optimized through dynamic data analysis, and the appearance uniformity is improved.

Description

Technical Field

[0001] The present invention relates to the field of rice polishing rate detection, and in particular to a method and device for detecting rice quality indicators throughout the entire process. Background Art

[0002] The field of rice polishing rate detection technology encompasses technologies for detecting and evaluating various parameters involved in the rice milling process. The core of this technical field revolves around the quality changes of rice at various stages from raw grain to finished product, focusing on the acquisition and calculation of key quality indicators such as rice moisture content determination, impurity ratio detection, rice breakage analysis, and finished rice polishing rate calculation. The entire technical system encompasses the extraction of multiple physical quantities, including sample collection, image recognition, weight and volume measurement, and particle counting. It is widely used in agricultural grain storage and transportation, quality control at processing companies, and grain testing agencies, and is of great significance for ensuring the consistency and standardization of rice processing products.

[0003] Among them, the full-process detection method of rice quality indicators refers to a method for continuously measuring and recording multiple key quality parameters of rice from the time it enters the factory to the time it leaves the factory. The subject of this patent mainly targets the problems existing in rice quality management, such as single-point detection, discontinuous sampling, and incomplete coverage of indicators. By setting fixed points, the rice samples are tested for specific indicators such as grain number statistics, moisture weight ratio determination, whole polished rice rate calculation, and husk residue rate identification. It uses image recognition technology for particle screening, uses a weighing device for precision measurement, and combines with the rice grain shape database for morphological comparison, so as to systematically obtain the required quality information and realize full-link quality detection from factory inspection, pre-processing screening, and finished product evaluation after rice milling.

[0004] Existing technologies still have shortcomings in automation and real-time monitoring, especially in the ability to accurately measure the spacing between rice grains and instantly evaluate structural characteristics. The degree of automation and data processing capabilities are restricted by the limitations of equipment and algorithms, and fine quality sorting cannot be carried out, which hinders the improvement of processing efficiency and product quality. Data processing relies on batch processing and lacks support for dynamic adjustment of production lines. This is particularly evident in large-scale production environments. Processing parameters cannot be adjusted according to real-time data, thereby increasing production costs. Summary of the Invention

[0005] The present invention aims to solve the technical problems existing in the prior art and provides a method and device for detecting the whole process of rice quality indicators.

[0006] The present invention solves the above-mentioned technical problem with the following technical solution: A method for detecting rice quality indicators in a full process, comprising the following steps: On the one hand, S1: collecting feed particle images through an infrared camera, screening rice particles according to the feed particle images, extracting rice texture and contour coordinates according to the surface images of the rice particles, identifying the position and morphology of the rice particles and comparing the sample rice textures, and obtaining the effective feed particle index input into the rice hulling chamber; S2: Perform image recognition on the brown rice kernels output from the hulling chamber, detect the infrared transmission intensity of the brown rice, calculate the average transmission intensity, compare the transmission intensity of the sample brown rice with the average, analyze the structural characteristics of the brown rice, generate a structural integrity rating, and calculate the perfect kernel ratio index and the imperfect kernel ratio index based on the proportion of kernels with qualified transmission intensity to generate a roughness ratio index; S3: Milling the brown rice and performing image recognition on the milled rice grains to identify the contour image of a single rice grain, calculate the contour defect ratio, and classify the rice into broken rice and polished rice to obtain the rice grain quality classification result; S4: Based on the rice grain quality classification result, the rice grains are introduced into the corresponding weighing channel, the cumulative weight of the channel is recorded, the total weight of the rice grains is compared, the weight ratio of the polished rice is calculated, and the whole polished rice rate index is generated; S5: Perform image recognition on the polished rice obtained by milling, identify the proportion of yellow rice based on the image color, calculate the proportion of yellow rice in the total polished rice, calculate the final qualified rate based on the rough rice yield and whole polished rice rate indicators, and output the proportion of yellow rice and comprehensive quality rating.

[0007] On the one hand, the effective feed particle indicators include particle size distribution, arrangement density, and input uniformity; the perfect grain ratio indicators include structural continuity ratio, low transmittance defect rate, and complete grain ratio; the rice grain quality classification results include contour integrity level, surface damage degree, and grain shape consistency index; the whole polished rice rate indicators include weight ratio, polished rice stable output ratio, and polished rice output ratio per unit raw material; the comprehensive quality rating includes color uniformity, polished rice purity, and yellow rice ratio level.

[0008] On the one hand, the step of inputting the effective feed particle index into the husking chamber is specifically as follows: S101: collecting feed particle images through an infrared camera, screening rice grains based on the feed particle images, detecting the edge contours of the rice images based on the surface images of the rice grains, calculating pixel gradient changes in continuous areas of the image, locating the coordinates of the rice edge contours, measuring the texture interval distance of each rice grain, analyzing the texture interval change trend, and obtaining rice edge parameters; S102: Based on the rice edge parameter set, identifying the morphological characteristics of a single rice grain, comparing the length-to-width ratio and surface area of a standard rice grain, determining the contour closure of each rice grain, identifying rice grains with abnormal morphology, analyzing the rice grain morphology distribution range, and generating rice grain morphology parameters; S103: Based on the rice morphological parameters, the number of rice grains that meet the standards is screened to obtain an index of effective feed grains input into the rice hulling chamber.

[0009] On the one hand, the steps of structural integrity rating are specifically as follows: S201: Based on the effective feed particles, using an infrared transmission device to perform transmission detection on the brown rice after husking, collecting transmission data of each brown rice grain in the sample, analyzing the energy change of the transmitted light passing through the brown rice, calculating the transmission intensity value of a single brown rice grain, and generating a brown rice transmission intensity characteristic; S202: Calculating an average transmission intensity of the brown rice based on the brown rice transmission intensity characteristic, comparing the average transmission intensity with the transmission data of a single grain of brown rice, analyzing the degree of deviation of the transmission value, screening brown rice data with abnormal transmission values, determining the numerical range of the abnormal transmission intensity, and generating a transmission intensity deviation value; S203: Based on the brown rice transmission deviation value, analyzing the transmission uniformity of the brown rice, calculating the distribution of the transmission deviation, determining the structural characteristics of the brown rice with abnormal light intensity, screening the brown rice that meets the standard structure, classifying the structural integrity level, and generating a structural integrity rating.

[0010] On the one hand, the transmission intensity value of a single grain of brown rice is calculated using the formula: ; Generate brown rice transmission intensity characteristics; in, Represents the transmittance intensity value of a single grain of brown rice, represents the incident light intensity, represents the optical absorption coefficient of brown rice, Represents the thickness of a single grain of brown rice. represents the intensity of transmitted scattered light, represents the light scattering correction parameter, Represents the base thickness.

[0011] On the one hand, the S3 further includes: Based on the structural integrity rating, the polished rice image is denoised, the surface features of the brown rice are analyzed, the brightness fluctuation and texture interval are calculated, the milling sample is called, the data deviating from the milling range is identified, the uniformity of the milling is judged, and the milling quality analysis result is obtained. Based on the milling quality analysis result, the sample is transferred to the image acquisition area.

[0012] On the one hand, the steps of classifying the rice grain quality results are specifically as follows: S301: Based on the structural integrity rating, image data of milled brown rice is retrieved and denoised, pixel information of the rice grain surface area is extracted, the spatial distribution of pixel grayscale values is analyzed, areas with prominent surface pixel brightness are screened, and brightness change values of the areas are calculated to generate rice grain surface brightness fluctuation data; S302: Based on the rice grain surface brightness fluctuation data, analyzing the distribution of pixel areas with prominent brightness changes, measuring the intervals between adjacent brightness extreme value points, calculating the offset amplitude of the brightness mutation interval, and determining the texture difference of the rice grain surface to generate the rice grain texture interval difference; S303: Based on the rice grain surface brightness fluctuation data and the rice grain texture interval difference, calling the rice grain milling sample data, screening the data points that exceed the normal milling range, comparing the surface brightness and texture variation range of each rice grain, analyzing the rice grain surface smoothness, calculating the milling uniformity offset, and obtaining the milling quality analysis result; S304: Based on the milling quality analysis results, the rice sample is transferred to an image acquisition area using a conveyor device, the contour boundary of a single rice grain is detected, the contour edge pixels are selected and connected to form a closed contour, the contour overlap area is calculated for the overlapping area, and the overlapping portion is morphologically segmented to obtain the rice grain contour features; S305: Analyzing the pixel distribution on the rice grain surface based on the rice grain contour features, screening the surface spot area according to the pixel brightness change, detecting the pixel fracture along the main axis, calculating the number and length of cracks, analyzing the structural integrity of the rice grain surface, and obtaining surface damage data; S306: Based on the surface damage data, the surface area of the rice grains is measured and compared with the spots and cracks, the total area ratio of the defective area is calculated, and the rice grains are classified according to the contour defect ratio. Rice grains with smooth and complete surfaces are classified as polished rice, and rice grains with excessive defects are classified as broken rice, thereby obtaining the rice grain quality classification result.

[0013] On the one hand, the calculation of the offset amplitude of the brightness mutation interval adopts the formula: ; and determining the surface texture differences of rice grains to generate the texture interval differences of rice grains; in, Represents the offset amplitude of the brightness mutation interval, Representative The brightness of the maximum brightness point, Representative The brightness of the minimum brightness point, Represents the pixel interval between adjacent brightness extreme points, Represents the total number of detected brightness extreme point pairs.

[0014] On the one hand, the steps of the head rice rate index are specifically as follows: S401: Based on the rice grain quality classification result, the rice grain conveying path is automatically assigned, the broken rice and polished rice are respectively introduced into the corresponding weighing channels, the cumulative weight data of each channel is recorded in real time, and the total amount of each type of rice grain is obtained; S402: Call the total amount of each type of rice grains, call the total weight of the rice grains, compare the cumulative weight of each type of rice grains with the total weight, calculate the weight proportions of broken rice and polished rice respectively, and determine the average proportion of polished rice in the total amount of rice based on the total amount of rice to generate a whole polished rice rate index.

[0015] On the other hand, a rice quality index full-process detection device is provided, which is used to perform the above-mentioned rice quality index full-process detection method, and the device includes: An infrared image acquisition module collects images of feed particles using an infrared camera, screens rice particles based on the images of the feed particles, extracts rice grain texture and contour coordinates based on the surface images of the rice particles, identifies the position and morphology of the rice grains, and compares the sample rice grain texture to obtain an effective feed particle index; a transmission intensity detection module, which detects the infrared transmission intensity of brown rice based on the effective feed particle index, calculates the average transmission intensity, compares the transmission intensity of the sample brown rice with the average value, analyzes the structural characteristics of the brown rice, generates a structural integrity rating, and calculates the perfect grain ratio index and the imperfect grain ratio index based on the proportion of grains with qualified transmission intensity to generate a roughness rate index; The rice grain quality classification module performs image recognition on milled rice grains, identifies the contour image of a single rice grain, calculates the contour defect ratio, and classifies it into broken rice and polished rice to obtain the rice grain quality classification result; a weighing distribution module, which, based on the rice grain quality classification result, directs the rice grains into corresponding weighing channels, records the cumulative weight of the channels, compares the cumulative weight of the rice grains, calculates the weight proportion of polished rice, and generates a polished rice rate indicator; The qualified rate calculation module performs image recognition on the polished rice obtained by milling, identifies the proportion of yellow rice based on the image color, calculates the proportion of yellow rice in the total polished rice, calculates the final qualified rate based on the rough rice yield and whole polished rice rate indicators, and outputs the proportion of yellow rice and the comprehensive quality rating.

[0016] The beneficial effects brought about by the technical solution provided by the embodiment of the present invention include at least: High-precision image processing is used to identify the position and shape of each rice grain, and to screen the effective feed particle indicators input into the rice hulling chamber. The infrared transmittance intensity detection of brown rice directly reflects the structural characteristics of the rice grains. The real-time quality assessment method optimizes the conversion process from brown rice to polished rice, ensuring the optimization of rice quality and yield. Real-time monitoring of surface brightness and texture changes during the milling process helps to adjust the milling process. Dynamic data analysis is used to optimize product quality and improve appearance uniformity. Automated rice grain mass classification and weight ratio calculation not only improves the calculation accuracy of the whole polished rice rate, but also speeds up the processing speed of the production line, significantly improving overall processing efficiency and product quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a flow chart of the main steps of the present invention; Figure 2 This is a diagram of the device of the present invention. DETAILED DESCRIPTION

[0018] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts are within the scope of protection of this application.

[0019] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first" or "second" may explicitly or implicitly include one or more of the described features. In the description of this application, "plurality" means two or more, unless otherwise specifically specified.

[0020] In the description of this application, the term "for example" is used to mean "used as an example, illustration or explanation". Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any person skilled in the art to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art will recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid obscuring the description of the present invention with unnecessary details. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in this application.

[0021] Example 1 In order to make the technical problems, technical solutions and advantages to be solved by the present invention clearer, a detailed description will be given below with reference to the accompanying drawings and specific embodiments.

[0022] The embodiment of the present invention provides a full-process detection method for rice quality indicators, such as Figure 1 As shown, the following steps are included: S1: collecting feed particle images through an infrared camera, screening rice grains based on the feed particle images, extracting rice grain texture and contour coordinates based on the surface images of the rice grains, identifying the position and morphology of the rice grains and comparing the sample rice grain textures, and obtaining effective feed particle indicators for input into the rice hulling chamber; S2: Perform image recognition on the brown rice kernels output from the hulling chamber, detect the infrared transmission intensity of the brown rice, calculate the average transmission intensity, compare the transmission intensity of the sample brown rice with the average, analyze the structural characteristics of the brown rice, generate a structural integrity rating, and calculate the perfect kernel ratio index and the imperfect kernel ratio index based on the proportion of kernels with qualified transmission intensity to generate a roughness ratio index; S3: Milling the brown rice and performing image recognition on the milled rice grains to identify the contour image of a single rice grain, calculate the contour defect ratio, and classify the rice into broken rice and polished rice to obtain the rice grain quality classification result; S4: Based on the rice grain quality classification results, the rice grains are imported into the corresponding weighing channel, the cumulative weight of the channel is recorded, and the total weight of the rice grains is compared to calculate the weight ratio of polished rice to generate the whole polished rice rate index; S5: Perform image recognition on the polished rice obtained by milling, identify the proportion of yellow rice based on the image color, calculate the proportion of yellow rice in the total polished rice, calculate the final qualified rate based on the rough rice yield and whole polished rice yield indicators, and output the proportion of yellow rice and comprehensive quality rating.

[0023] The effective feed particle indicators include particle size distribution, arrangement density, and input uniformity. The complete particle ratio indicators are specifically the structural continuity ratio, low transmittance defect rate, and the proportion of complete particles. The rice grain quality classification results include the contour integrity level, surface damage degree, and grain shape consistency index. The whole polished rice rate indicators are specifically the weight ratio, stable polished rice output ratio, and unit raw material polished rice output ratio. The comprehensive quality rating includes color uniformity, polished rice purity, and yellow rice ratio.

[0024] On the one hand, the steps for inputting the effective feed particle index of the husking chamber are as follows: S101: collecting feed particle images through an infrared camera, screening rice grains according to the feed particle images, detecting the edge contours of the rice images according to the surface images of the rice grains, calculating pixel gradient changes in continuous areas of the image, locating the coordinates of the rice edge contours, analyzing the trend of texture interval changes, and obtaining rice edge parameters; The surface image of rice is collected based on an infrared camera. Specifically, an infrared camera with a resolution of 2048x1536 pixels is set up, which is vertically aligned with a conveyor belt moving at a uniform speed (for example, 5 cm / s) below. The rice is spread in a single layer on the conveyor belt. The camera is 30 cm away from the surface of the conveyor belt. An 850nm infrared light source is used to uniformly illuminate the collection area, and the surface image data of the rice is captured frame by frame. An edge detection operator is applied to each frame of image data obtained, such as the Canny operator, with a low threshold of 50 and a high threshold of 150, to perform preliminary identification of edge pixels, and connect these edge pixels to form an initial contour. Then, the difference in infrared grayscale value between each pixel in the image and its adjacent pixels (for example, eight neighborhoods) is calculated to obtain a pixel gradient amplitude map, the pixel points on the initial contour are traversed, their gradient amplitudes are analyzed, and the pixel sequence with the most significant gradient amplitude change is located as the final rice edge contour coordinates. For example, for a grain of rice, its contour coordinate sequence is obtained. Subsequently, in the internal area of the located rice grain contour, along the long axis of the rice grain, the gray level co-occurrence matrix analysis method is used to extract texture features, identify and locate the periodic changes of surface texture (such as the gullies of rice husk), calculate the pixel distance between adjacent significant texture features (such as the local minimum point of grayscale), and convert it into physical spacing distance according to the pre-calibrated pixel to physical size ratio (such as 0.05 mm per pixel). For example, the texture spacing distance sequence on a certain rice grain is measured as {0.82 mm, 0.85 mm, 0.81 mm, 0.88 mm}. Further, these measured texture spacing distance data are analyzed. , calculate its mean (for example, (0.82+0.85+0.81+0.88) / 4=0.84 mm) and standard deviation (for example, 0.029 mm is calculated), and compare the standard deviation with a preset stability threshold (for example, 0.05 mm, which is set based on the analysis of the statistical distribution of the standard deviation of texture intervals of a large number of standard rice samples, and the 80% quantile of the distribution is taken). If the standard deviation is less than 0.05 mm, it is judged that the trend of texture interval change is stable, otherwise it is unstable. The edge contour coordinate sequence, average texture interval distance, texture interval standard deviation and other information of the rice grain are summarized to obtain the edge parameters of the rice grain.

[0025] S102: Based on the rice edge parameter set, identify the morphological characteristics of a single rice grain, compare the length-to-width ratio and surface area of a standard rice grain, determine the closedness of the contour of each rice grain, identify rice grains with abnormal morphology, analyze the distribution range of rice grain morphology, and generate rice grain morphological parameters; Based on the rice edge parameter set obtained in S101, the parameter set includes the edge contour coordinate sequence of multiple rice grains, the average texture interval distance, the texture interval standard deviation, etc. First, the edge contour coordinate sequence of a single rice grain is , calculate its minimum enclosing rectangle and obtain the length of the rice (e.g. 10.5 mm) and width (e.g. 3.2 mm), calculate the aspect ratio (For example, 10.5 / 3.2≈3.28), and use the contour coordinates to calculate the contour perimeter of the rice and the area enclosed (For example, the area calculated by Green's formula or pixel counting method is 25.8 square millimeters), and the calculated aspect ratio is and area Compare with the reference values in the pre-stored standard rice morphology database. For example, the length-to-width ratio range of the standard "medium-grain indica rice" is set to [3.0, 3.5], and the area range is set to [24.0, 28.0] square millimeters (these ranges are obtained by measuring and counting more than 1,000 "medium-grain indica rice" samples that meet the national standards and taking their 95% confidence intervals). If a certain rice grain has and square millimeters, then its morphological characteristics initially meet the standards. Then, the closure of the rice contour is evaluated, specifically by checking the edge contour coordinate sequence The first and last coordinates of and Euclidean distance between , set a closing threshold (For example Set to 2 pixels, based on image resolution and acceptable contour breakage). , the contour is considered to be well closed, otherwise it is considered to be incomplete. Then, according to the aspect ratio ,area The three conditions of whether it is within the standard range and whether the contour closure is good are comprehensively used to identify rice with abnormal morphology. For example, if the aspect ratio of a rice grain is 2.5 (less than 3.0) or the area is 30.0 square millimeters (greater than 28.0) or the contour is not closed, it will be marked as morphologically abnormal. The morphological parameters (aspect ratio, area) of all rice grains in the batch are counted and their distribution is analyzed. For example, a histogram of the aspect ratio and area is drawn to determine the main distribution range. For example, it is found that 85% of the rice grains have aspect ratios concentrated between 3.1 and 3.4, and areas concentrated between 25.0 and 27.5 square millimeters. The parameters of rice grains that meet the standard morphology (aspect ratio and area within the standard range and closed contour) are combined to generate rice morphological parameters.

[0026] S103: Based on the rice morphological parameters, the number of rice grains that meet the standards is screened to obtain an index of effective feed particles input into the rice hulling chamber.

[0027] On the one hand, the steps of structural integrity rating are as follows: S201: Based on the effective feed particles, a transmission test is performed on the brown rice after husking using an infrared transmission device, transmission data of each brown rice grain in the sample is collected, energy changes of the transmitted light passing through the brown rice are analyzed, and the transmission intensity value of each brown rice grain is calculated to generate a brown rice transmission intensity characteristic; Based on the effective feed particles obtained in S103, these particles are rice grains with normal shape and appropriate spacing during the feeding process. After the rice hulling process, they become brown rice. These brown rice grains are detected one by one using an infrared transmission device. The device includes an infrared light source (for example, a wavelength of 940nm, a power adjustable range of 1-10mW, set to 5mW) and an infrared detector (for example, an InGaAs photodiode). The brown rice passes through a fixed optical path between the light source and the detector grain by grain. The infrared light beam emitted by the light source penetrates the brown rice and is received by the detector. The detector signal intensity data is collected when each brown rice sample passes through the optical path. The data reflects the energy of the transmitted light, that is, the transmitted light intensity. The energy change of the transmitted light when passing through a single grain of brown rice is analyzed, mainly by recording the lowest signal value measured by the detector when the brown rice completely blocks the optical path, and converting it into the transmitted light intensity. To eliminate the influence of background light and equipment noise, the background light intensity is measured when no brown rice passes through. , and measure the dark current intensity in the area where the light source does not illuminate the detector when brown rice passes through , for the original measurement signal To make a correction: , and then calculate the transmission intensity value of a single grain of brown rice , using the formula: , this formula is used to calculate the theoretical transmission intensity value of a single grain of brown rice, where Represents the transmittance intensity value of a single grain of brown rice (unit: mW / cm²). Represents the incident light intensity (unit: mW / cm²), which is the measured intensity when the infrared light source directly illuminates the detector. It needs to be calibrated in advance. For example, mW / cm², Represents the optical absorption coefficient of brown rice to infrared light of this wavelength (unit: cm -1 ), which is related to the internal components of brown rice (such as moisture and starch structure), can be obtained by fitting the transmission experiment of a standard brown rice sample with known components. For example, for a specific variety of brown rice, cm -1 , Represents the actual thickness of a single grain of brown rice (unit: cm). This thickness can be measured by a laser thickness gauge or image analysis (based on the side view) before or during transmission testing. For example, the thickness of a grain of brown rice is mm=0.28cm, Represents the intensity of scattered light generated during transmission (unit: mW / cm²). This part of light passes through the brown rice but deviates from the direct transmission path. Its intensity is affected by the scattering characteristics of the surface and internal structure of the brown rice. It can be measured and estimated by setting an integrating sphere around the detector to collect the scattered light, for example, mW / cm², Represents the light scattering correction parameter (dimensionless), which is used to correct the influence of different thicknesses of brown rice on the scattering effect. This parameter needs to be obtained by experimental calibration of samples of different thicknesses. For example, , Represents the benchmark thickness (unit: cm), usually set to the average thickness of the batch or variety of brown rice, for example, by measuring a large number of samples in advance to obtain the average thickness mm=0.25cm, substitute these parameters into the formula for calculation: , calculate the exponential term: , calculate the correction term in the denominator: , calculate the square root term: , Substitute the formula: mW / cm². This measurement and calculation process is performed for each grain of effective raw brown rice that passes the test, and the transmission intensity value of each grain of brown rice is recorded. , forming a list of transmission intensity values, such as {7.52, 7.61, 7.48, 7.55, 7.40, 7.65, ...}, which constitutes the transmission intensity characteristic of brown rice. The logic and purpose of the formula: This formula aims to estimate the intensity of infrared light transmitted through a single grain of brown rice by comprehensively considering the two main attenuation mechanisms of light absorption and scattering. Derived from the Beer-Lambert law, which describes the exponential attenuation of light due to absorption in a homogeneous medium, where the absorption coefficient and thickness However, brown rice is not a uniform and transparent medium and has significant scattering phenomenon, so the introduction of The term is used to compensate for the scattered light contribution outside the direct transmission path (the square root form is a simplified model based on the scattered light intensity, incident light and medium properties). At the same time, considering the thickness of the brown rice It is not constant, and thickness changes affect the relative contributions of absorption and scattering, introducing a correction term in the denominator , which is expressed by the dimensionless parameter and relative thickness To adjust the total transmission intensity, the greater the thickness, the greater the correction term, the transmission intensity is relatively reduced, and vice versa. The absolute value ensures that the final intensity is non-negative. The benefit of the formula is that by taking into account the absorption ( ),scattering( ) and the combined effect of thickness on these two effects ( ), compared to using only the Beer-Lambert law, it can more accurately simulate the actual transmission intensity of infrared light when it passes through non-uniform, scattering brown rice particles, providing a more reliable physical quantity basis for subsequent judgment of the integrity of the internal structure of brown rice based on transmission intensity. mW / cm² indicates that for this grain of brown rice with a thickness of 0.28 cm, its theoretical transmission intensity value is approximately 7.52 mW / cm² under given incident light intensity, absorption coefficient, scattered light intensity, and correction parameters. This value is a specific example of the transmission intensity characteristic of brown rice and will be used in subsequent deviation analysis and structural integrity rating.

[0028] S202: Calculating an average transmission intensity of the brown rice based on the brown rice transmission intensity characteristic, comparing the average transmission intensity with the transmission data of a single grain of brown rice, analyzing the degree of deviation of the transmission value, screening brown rice data with abnormal transmission values, determining the numerical range of the abnormal transmission intensity, and generating a transmission intensity deviation value; Based on the brown rice transmission intensity characteristics generated in S201, that is, the transmission intensity value calculated for each grain of brown rice The set of {7.52,7.61,7.48,7.55,7.40,7.65,…}, first calculate the average transmittance intensity of this batch of brown rice samples For example, if the sample contains 100 grains of brown rice and the sum of their transmitted intensity values is 753.5 mW / cm², the average value is mW / cm², this average As the benchmark of the transmittance intensity of the batch of brown rice, the transmittance intensity value of each grain of brown rice in the data set is then With the average Compare and calculate the transmission intensity deviation of each grain of brown rice , for example, for mW / cm², with a deviation of mW / cm², for mW / cm², with a deviation of mW / cm². Next, analyze the degree of deviation of these deviation values and set a deviation threshold. To filter out brown rice data with abnormal transmittance values, the threshold The setting can be based on a percentage of the mean, or based on the standard deviation of the sample deviation values, for example, calculating all deviation values Standard deviation , assuming that the calculated mW / cm², set the abnormality criterion to the absolute value of the deviation exceeding times the standard deviation, for example , then the abnormal threshold is mW / cm² (the The choice of means that it covers about 95% of the normal fluctuation range. It is based on statistical principles and the experience of the uniformity of the internal structure of brown rice. The absolute value of the deviation of each grain of brown rice is determined. Is it greater than mW / cm², if , then the transmission intensity of the brown rice is judged to be abnormal, for example, , , judged as normal, , , is also judged to be normal, assuming there is another grain of brown rice mW / cm², its deviation mW / cm², , then determine The transmission intensity is abnormal (low), and we assume that a grain of brown rice mW / cm², its deviation mW / cm², , then determine Also for abnormal transmittance (high), record the transmittance values of all brown rice that are judged to be abnormal, and determine the value range of abnormal transmittance. For example, in this batch, the range of abnormal low is mW / cm², with abnormally high ranges being mW / cm², each grain of brown rice is compared to its corresponding deviation value The values of the transmission intensity deviation values are correlated to generate a set of transmission intensity deviation values.

[0029] Table 1 Example of brown rice transmission intensity and deviation

[0030] Note: Table 1 assumes incident light intensity , absorption coefficient , scattered light intensity , correction coefficient , reference thickness .Average transmission intensity . As shown in Table 1, the measured thickness of 6 brown rice samples and the transmission intensity values calculated according to the formula in S201 are listed. , relative to the batch average Deviation value calculated from mW / cm² , and whether the absolute value of the deviation exceeds the threshold mW / cm² is used to determine whether the transmittance intensity of the brown rice grain is abnormal. Table 1 shows the specific process of abnormal screening based on transmittance intensity characteristics.

[0031] S203: Based on the brown rice transmission deviation value, the transmission uniformity of the brown rice is analyzed, the distribution of the transmission deviation is calculated, the structural characteristics of the brown rice with abnormal light intensity are determined, the brown rice meeting the standard structure is selected, the structural integrity level is classified, and a structural integrity rating is generated.

[0032] On the one hand, the transmission intensity value of a single grain of brown rice is calculated using the formula: ; Generate brown rice transmission intensity characteristics; in, Represents the transmittance intensity value of a single grain of brown rice, represents the incident light intensity, represents the optical absorption coefficient of brown rice, Represents the thickness of a single grain of brown rice. represents the intensity of transmitted scattered light, represents the light scattering correction parameter, Represents the base thickness.

[0033] Based on the brown rice transmittance deviation value set {(-0.015), (0.075), (-0.135), (0.115), ...} generated in S202, the transmittance uniformity of the whole batch of brown rice is first analyzed. This can be achieved by calculating the distribution of the deviation values, for example, calculating the standard deviation of the deviation values. (Calculated in S202, 0.04 mW / cm²) or calculate the coefficient of variation (For example , i.e. 0.53%), a smaller standard deviation or coefficient of variation indicates that the transmission intensity of the whole batch of brown rice is relatively uniform. Next, the abnormal transmission intensity (i.e. ) of the structural characteristics of brown rice, it is generally believed that the transmission intensity is abnormally low (such as brown rice No. 5, , deviation -0.135) may mean that there are cracks, gaps or impurities inside the brown rice. These defects increase the scattering or absorption of light, resulting in a decrease in the transmitted intensity, while the transmitted intensity is abnormally high (such as brown rice No. 6, , deviation +0.115) may mean that the brown rice is thinner than the average thickness, or the internal structure is abnormally dense, resulting in reduced scattering (the latter is less common). Then, the brown rice that meets the standard structure is screened according to the transmission intensity deviation value, and a structural integrity standard is set, for example, the absolute value of the transmission intensity deviation is exist (For example, brown rice with a value less than 0.08mW / cm²) is considered to have a complete structure, and the absolute value of the deviation is greater than The brown rice is judged as having incomplete structure or suspected defects. Furthermore, the structural integrity can be graded according to the size of the deviation and multiple thresholds can be set. For example, Grade A (excellent): (Right now Grade B (good): (Right now ); Level C (Medium): (Right now ); Grade D (poor): (Right now ), these thresholds are set based on a large amount of experimental data, correlating the performance of brown rice with different transmission deviation ranges in subsequent milling and finished product quality testing (such as broken rice rate, crack rate). For example, for brown rice No. 1 (deviation -0.015), it belongs to grade A, for brown rice No. 2 (deviation +0.075), it belongs to grade B, for brown rice No. 5 (deviation -0.135), due to , belongs to grade D, for brown rice number 6 (deviation +0.115), due to , belonging to grade C, each grain of brown rice is assigned a structural integrity grade (A, B, C, D), and this set of grade information is the final generated structural integrity rating.

[0034] On the one hand, S3 also includes: Based on the structural integrity rating, the polished rice image is denoised, the surface characteristics of the rough rice are analyzed, the brightness fluctuation and texture interval are calculated, the milling sample is called, the data deviating from the milling range is identified, the uniformity of the milling is judged, and the milling quality analysis results are obtained. Based on the milling quality analysis results, the sample is transferred to the image acquisition area.

[0035] On the one hand, the steps for classifying the rice grain quality results are as follows: S301: Based on the structural integrity rating, image data of milled brown rice is retrieved and denoised. Pixel information of the rice grain surface is extracted, the spatial distribution of pixel grayscale values is analyzed, and areas with prominent surface pixel brightness are screened. The brightness variation value of the area is calculated to generate rice grain surface brightness fluctuation data. Based on the structural integrity rating generated by S203 (each grain of brown rice has an A, B, C, or D grade), optical image data of polished rice formed after the brown rice has been milled is retrieved. These images are usually taken by a CCD or CMOS camera under visible light, for example, with a resolution of 1600x1200 pixels. First, the acquired polished rice image is denoised by applying, for example, a Gaussian filter and setting a standard deviation. , to smooth out sensor noise and slight surface irregularities. Then, use image segmentation technology (such as Otsu thresholding or color space analysis) to extract the pixel information of the surface area of each grain of polished rice, excluding background pixels. For all pixels in the extracted surface area of each grain of polished rice, analyze their spatial distribution of grayscale values and calculate the average grayscale value of the pixels in this area. and standard deviation ,Next, the area with prominent brightness of the surface pixels is filtered, which can be ,achieved by setting a brightness threshold, for example, ,setting a brightness value greater than , Identify the pixel area with a brightness value less than The pixel area is identified as a dark area (the threshold coefficient 1.5 is set based on the statistical analysis of the brightness distribution of the normal polished rice surface, aiming to capture significant local brightness anomalies). For example, the average grayscale of a polished rice , standard deviation , then the screening brightness is greater than or less than The pixel area of these prominent areas is counted, and the ratio of the total area of these prominent areas to the total area of the rice grain is calculated. The average brightness and The difference is used as the brightness change value of the area. For example, if the average brightness of a highlight area is 205, the brightness change value is , the average brightness of a dark area is 150, then the brightness change value is , all identified prominent areas and their corresponding brightness change values are summarized to form the surface brightness fluctuation data of the rice grain.

[0036] S302: Based on the brightness fluctuation data of the rice grain surface, the distribution of pixel areas with prominent brightness changes is analyzed, the intervals between adjacent brightness extreme points are measured, the offset amplitude of the brightness mutation interval is calculated, and the texture difference of the rice grain surface is determined to generate the rice grain texture interval difference; Based on the rice grain surface brightness fluctuation data generated in S301, which contains the position, area and brightness change value information of the brightness prominent area on the surface of each rice grain, the distribution of these pixel areas with prominent brightness changes on the rice grain surface is first analyzed, such as calculating the centroid coordinates of these areas, analyzing whether they are concentrated or dispersed, and whether they are arranged along a specific direction (such as the long axis of the rice grain). Then, in order to quantify the roughness or irregularity of the surface texture, it is necessary to measure the intervals between adjacent brightness extreme points. This is usually done along the main axis of the rice grain, and one or more brightness profile lines are extracted on the main axis of the rice grain. The profile line data is smoothed (such as moving average), and then a peak detection algorithm is used to find the brightness maximum point. and brightness minimum point The position and brightness value of , is the total number of extreme point pairs detected), calculate the pixel interval between adjacent extreme points (a maximum and a minimum) , for example, the first extreme point pair: maximum At pixel position 50, the minimum At pixel position 68, then Pixel, second extreme point pair: maximum At pixel position 110, the minimum At pixel position 135, then pixels, assuming that a total of For extreme points, 5 groups were obtained Then, calculate the offset amplitude of the brightness mutation interval , using the formula: , this formula calculates the weighted average offset amplitude of the brightness mutation interval, where Represents the offset amplitude of the brightness mutation interval (the unit depends on the brightness and distance unit, here is grayscale pixel), Representative The brightness (gray level) of the maximum brightness point, Representative The brightness of the minimum brightness point (gray level), Representative For the pixel interval between adjacent brightness extreme points (unit: pixel), Represents the total number of detected brightness extreme point pairs, summation symbol Indicates that all Accumulate the extreme points, and calculate the average brightness of each pair of extreme points (Ignore the division by 2 because it is a relative comparison) multiply by the interval between them , and then summing, which is equivalent to weighted summing of the average brightness by interval, reflecting the product of the fluctuation amplitude of the overall brightness and the spatial frequency. The denominator is is the root mean square (RMS) of all interval distances and is used to normalize the weighted sum of the numerator so that the result It mainly reflects the relative fluctuation amplitude of brightness, reduces the dependence on the absolute length of rice grains, and the absolute value ensures that the result is non-negative. Assume , using the example data above to calculate: , calculate the numerator: , calculate the denominator: ,calculate , according to the calculated The value is used to judge the difference in rice grain surface texture and set the texture difference threshold, for example, It is judged that the texture is very different (the surface is rough or unevenly ground). The texture difference is judged to be medium. The texture difference is judged to be small (smooth surface). In this case, , it is determined that the rice grain texture is very different, and the calculated The value or its corresponding level (large / medium / small) generates the rice grain texture interval difference as a result.

[0037] S303: Based on the rice grain surface brightness fluctuation data and the rice grain texture interval difference, the rice grain milling sample data is retrieved, and data points outside the normal milling range are screened. The surface brightness and texture variation range of each rice grain are compared, the rice grain surface smoothness is analyzed, and the milling uniformity offset is calculated to obtain the milling quality analysis result. Based on the rice grain surface brightness fluctuation data generated by S301 (including brightness standard deviation and brightness highlight area information) and the rice grain texture interval difference generated by S302 (quantized as The first step is to call the pre-established rice milling sample database, which stores the parameters of a large number of polished rice samples processed by standard milling procedures. These samples are classified by factors such as variety, original rough rice quality (which may contain S203 structure rating information), and milling time, and their corresponding surface brightness standard deviation is recorded. and texture interval difference For example, the database shows that for "First-class long-grain indica rice" with grade A structural integrity, after standard milling, The normal range is [5,12], The normal range is [200,350], then the current rice grain to be analyzed is calculated (For example, the rice grain 1 in Table 2 , rice grain 3 )and (For example, the rice grain 1 in Table 2 , rice grain 3 ) is compared with the corresponding conventional milling interval in the database, and the data points that exceed the conventional milling interval are screened out. In this example, assuming that rice grains 1 and 3 are both milled from Grade A brown rice, the In the range [5,12], but Out of the range [200,350]; rice grain 3 Out of the range [5,12], The surface parameters of rice grains 1 and 3 are also outside the range of [200,350]. Therefore, the surface parameters of rice grains 1 and 3 are judged to deviate from the conventional milling results. Then, the surface brightness of each rice grain is comprehensively compared (in terms of as a representative) and the texture variation range (represented by As a representative), the smoothness of the rice grain surface is analyzed. Smooth rice grains should have lower and lower The value is (such as rice grains 2 and 4 in Table 2), otherwise it means the surface is rough or the grinding is uneven (such as rice grains 1 and 3). Then, the grinding uniformity offset is calculated, which can be defined as the parameter point of the rice grain. The normalized distance to the center point of the normal range of its corresponding category in the database, or simply count and The number of out-of-range items (0, 1, or 2) in the two indicators is used as the level of the offset. For example, rice grain 1 has one indicator ( ) is out of range, with an offset level of 1; both indicators of rice grain 3 are out of range, with an offset level of 2; rice grains 2 and 4 are both 0. By analyzing the offset distribution of the milling uniformity of a batch of rice grains, for example, calculating the proportion of rice grains with average offset levels or high offset levels (such as levels 1 and 2), the milling quality analysis results of the entire batch are obtained. This result summarizes the degree and proportion of rice grain surface parameters deviating from the normal range, reflecting the uniformity and effect of the milling process.

[0038] S304: Based on the milling quality analysis results, the rice sample is transferred to the image acquisition area using a conveyor device, the contour boundary of the single rice grain is detected, the contour edge pixels are selected and connected to form a closed contour, the contour overlap area is calculated for the overlapping area, and the overlapping part is morphologically segmented to obtain the rice grain contour features; Based on the milling quality analysis result obtained in S303, the result indicates which rice grains may have abnormal surface characteristics (smoothness, uniformity), and all rice samples (regardless of the milling quality analysis result) are transferred to the next high-resolution image acquisition area using a conveyor device (for example, a vibrating feeder with a narrow belt conveyor to ensure that the rice grains move in a single column and separated state). In this area, another camera (for example, a linear array CCD camera, with a conveyor belt of appropriate speed, can obtain higher resolution rice grain images) is used to image the rice grains passing through one by one to detect the contour boundary of the single rice grain. The edge detection algorithm (such as Sobel or Canny operator, the threshold is adjusted according to the current imaging conditions) can be used again, or a method based on region growing or active contour model can be used to obtain more accurate contour pixel points, and these contour edge pixel points are screened out and connected through a connection algorithm (such as eight-neighborhood connection or These pixel points are connected based on the minimum perimeter principle to form a closed rice grain outline. During the transmission process, there may be contact or slight overlap between rice grains. In this case, processing is required after contour detection to identify potential overlapping areas (usually manifested as contour depressions or irregular connection points) and calculate the contour overlapping area of these suspected overlapping areas. The overlapping points can be identified by comparing the expected area of a single rice grain (based on the average size) with the detected connected domain area, or by analyzing the curvature change of the contour. The confirmed overlapping parts are morphologically segmented, for example, using a watershed algorithm or a concave point-based segmentation method to separate the contours of the adhered rice grains, ensuring that each object corresponds to a complete rice grain. Finally, accurate contour feature data is generated for each separated rice grain, including a contour pixel coordinate sequence and basic geometric parameters calculated therefrom (such as area, perimeter, major axis, and minor axis).

[0039] S305: Analyze the pixel distribution on the rice grain surface based on the rice grain contour features, screen the surface spot areas based on pixel brightness changes, detect pixel fractures along the main axis, calculate the number and length of cracks, analyze the structural integrity of the rice grain surface, and obtain surface damage data; Based on the single rice grain contour features (accurate contour coordinate sequence) obtained in S304, the rice grain surface area inside the contour is first analyzed by pixel analysis to check the distribution of pixel grayscale values. Using a similar method as in S301, the surface spot area is screened based on the difference in pixel brightness with its neighboring pixels or the global / local average brightness. For example, a pixel is defined as a pixel whose grayscale value is lower than the average of its 8 neighboring pixels by a threshold. (For example Grayscale, the threshold is set based on the analysis of the contrast between typical rice grain spots (such as diseased spots, black spots) and normal rice grain surfaces), then they are marked as spot pixels, and connected spot pixels constitute a spot area. The area of each spot area is calculated, and at the same time, pixel breaks along the main axis of the rice grain, i.e., cracks, are detected. This can be achieved by applying a line segment detection algorithm (such as Hough transform to detect straight line segments) or a specific crack detection filter (such as based on morphological operations or tensor voting) to the rice grain image, identifying dark or low-contrast line segments representing cracks in the image, calculating the number of each detected crack and its length (for example, by measuring the pixel distance between the two end points of the crack and converting it to millimeters), and then combining the spot and crack information to analyze the structural integrity of the rice grain surface. For example, the total area of the spots and the total length of the cracks on each grain of rice are counted, and these measurement results are summarized to obtain the surface damage data of the rice grain, which may be a vector or record containing the total area of the spots (for example, 0.5 square millimeters) and the total length of the cracks (for example, 1.2 millimeters).

[0040] S306: Based on the surface damage data, the surface area of the rice grains is measured and compared with the spots and cracks. The total area ratio of the defective area is calculated, and the rice grains are classified according to the contour defect ratio. Rice grains with smooth and intact surfaces are classified as polished rice, and rice grains with excessive defects are classified as broken rice. The rice grain quality classification results are obtained.

[0041] Based on the rice grain surface damage data obtained by S305 (including the total area of spots and total crack length , and the surface area of the rice grain profile obtained in S304 ), first measure or call the total surface area of the rice grain (For example square millimeters), then compare the surface damage data with the total area to calculate the total area ratio of the defect area. For cracks, the length of the crack can be Multiply by an assumed average width (For example mm, based on microscopic observation settings) to estimate its area (For example square millimeters), the total defect area (For example square millimeters), calculate the defect area ratio (For example , i.e. 2.34%), based on this contour defect ratio Perform final classification of rice grains and set classification thresholds (For example , which is set according to the definition of polished rice and broken rice in national or commercial standards, usually based on the allowable content of imperfect kernels), if (e.g. 2.34% 5%), and the rice grain itself is intact (not broken, which can be judged by the integrity of the outline and the aspect ratio), then the rice grain is judged as "polished rice". , or if the rice grain itself is broken (for example, the length is less than half of a normal rice grain), it is classified as "broken rice". This classification process is performed on all tested rice grains to obtain the final rice grain quality classification result, that is, each rice grain is marked as "polished rice" or "broken rice".

[0042] On the one hand, the offset amplitude of the brightness mutation interval is calculated using the formula: ; and determining the surface texture differences of rice grains to generate the texture interval differences of rice grains; in, Represents the offset amplitude of the brightness mutation interval, Representative The brightness of the maximum brightness point, Representative The brightness of the minimum brightness point, Represents the pixel interval between adjacent brightness extreme points, Represents the total number of detected brightness extreme point pairs.

[0043] On the one hand, the steps for the head rice rate index are as follows: S401: Based on the rice grain quality classification results, the rice grain conveying path is automatically assigned, broken rice and polished rice are respectively introduced into corresponding weighing channels, and the cumulative weight data of each channel is recorded in real time to obtain the total amount of each type of rice grain; Based on the rice grain quality classification result obtained in S306 (each grain of rice is labeled as "polished rice" or "broken rice," as shown in Table 3), the control device automatically assigns the subsequent conveying path for each grain of rice according to this classification label. For example, a fast-response sorting device (such as a pneumatic air blow valve or a shifting fork mechanism) is set at the end of the conveyor belt. When the detection device identifies a grain of rice as "polished rice" (such as rice grains 1, 2, and 5 in Table 3), the sorting device does not operate or guides it to path A. When it is identified as "broken rice" (such as rice grains 3 and 4 in Table 3), the sorting device operates (such as blowing or shifting) to guide it to path B. Path A leads to a polished rice collection container, and path B leads to a broken rice collection container. A weighing sensor (such as a high-precision electronic scale or a dynamic weighing module) is installed under each collection container or on the conveying path to monitor and record the cumulative weight data of the rice grains flowing into the corresponding channel in real time. For example, the weighing sensor reading of the polished rice channel is accumulated starting from 0 and recorded as The readings of the weighing sensor in the broken rice channel also start to accumulate from 0 and are recorded as ,As the rice grains are continuously classified and conveyed, these two weight data will be continuously updated, and the total amount (expressed in weight) of each type of rice grains (polished rice and broken rice) in the current processing batch (or over a period of time) can be obtained.

[0044] S402: Call the total amount of each type of rice grains, call the total weight of the rice grains, compare the cumulative weight of each type of rice grains with the total weight, calculate the weight proportion of broken rice and polished rice respectively, combine the total amount of rice, determine the average proportion of polished rice in the total amount of rice, and generate a whole polished rice rate index.

[0045] The total amount of each type of rice grains obtained by calling S401, i.e. the cumulative weight of polished rice (For example kg) and the cumulative weight of broken rice (For example kg), calculate the total weight of the processed rice grains (For example kg), compare the cumulative weight of each type of rice grains with the total weight, calculate the weight ratio of broken rice and polished rice, and the broken rice rate (For example ), polished rice rate (For example ), these two percentages reflect the composition of the product after milling and sorting, which then needs to be combined with the total amount of rice originally processed To calculate the polished rice rate, assume that the total amount of rice processed in this batch is Kilograms (this data needs to be obtained from the feed weighing or batch records), determine the average proportion of polished rice (represented by weight) in the total amount of rice, this proportion is the polished rice rate (For example ), this indicator reflects the total yield from the original rice to the final complete and qualified polished rice. The calculated polished rice rate (e.g. 65.38%) as the final output to generate the head rice rate indicator.

[0046] See also Figure 2 The present invention provides a full-process detection device for rice quality indicators, which is used to perform the above-mentioned full-process detection method for rice quality indicators. The device includes: An infrared image acquisition module collects images of feed particles using an infrared camera, screens rice particles based on the images of the feed particles, extracts rice grain texture and contour coordinates based on the surface images of the rice particles, identifies the position and morphology of the rice grains, and compares the sample rice grain texture to obtain an effective feed particle index; a transmission intensity detection module, which detects the infrared transmission intensity of brown rice based on the effective feed particle index, calculates the average transmission intensity, compares the transmission intensity of the sample brown rice with the average value, analyzes the structural characteristics of the brown rice, generates a structural integrity rating, and calculates the perfect grain ratio index and the imperfect grain ratio index based on the proportion of grains with qualified transmission intensity to generate a roughness rate index; The rice grain quality classification module performs image recognition on milled rice grains, identifies the contour image of a single rice grain, calculates the contour defect ratio, and classifies it into broken rice and polished rice to obtain the rice grain quality classification result; a weighing distribution module, which, based on the rice grain quality classification result, directs the rice grains into corresponding weighing channels, records the cumulative weight of the channels, compares the cumulative weight of the rice grains, calculates the weight proportion of polished rice, and generates a polished rice rate indicator; The qualified rate calculation module performs image recognition on the polished rice obtained by milling, identifies the proportion of yellow rice based on the image color, calculates the proportion of yellow rice in the total polished rice, calculates the final qualified rate based on the rough rice yield and whole polished rice rate indicators, and outputs the proportion of yellow rice and the comprehensive quality rating.

[0047] It should be noted that, in the above embodiments, the description of each embodiment has its own focus. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant description of other embodiments.

[0048] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, apparatus, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable devices (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0049] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (apparatus), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0050] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0051] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0052] Although the preferred embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0053] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. A full-process detection method for rice quality indicators, characterized in that: The method comprises: S1: collecting feed particle images through an infrared camera, screening rice grains based on the feed particle images, extracting rice grain texture and contour coordinates based on the surface images of the rice grains, identifying the position and morphology of the rice grains and comparing the sample rice grain textures, and obtaining effective feed particle indicators for input into the rice hulling chamber; S2: Perform image recognition on the brown rice kernels output from the hulling chamber, detect the infrared transmission intensity of the brown rice, calculate the average transmission intensity, compare the transmission intensity of the sample brown rice with the average, analyze the structural characteristics of the brown rice, generate a structural integrity rating, and calculate the perfect kernel ratio index and the imperfect kernel ratio index based on the proportion of kernels with qualified transmission intensity to generate a roughness ratio index; S3: Milling the brown rice and performing image recognition on the milled rice grains to identify the contour image of a single rice grain, calculate the contour defect ratio, and classify the rice into broken rice and polished rice to obtain the rice grain quality classification result; S4: Based on the rice grain quality classification result, the rice grains are introduced into the corresponding weighing channel, the cumulative weight of the channel is recorded, the total weight of the rice grains is compared, the weight ratio of the polished rice is calculated, and the whole polished rice rate index is generated; S5: Perform image recognition on the polished rice obtained by milling, identify the proportion of yellow rice based on the image color, calculate the proportion of yellow rice in the total polished rice, calculate the final qualified rate based on the rough rice yield and whole polished rice rate indicators, and output the proportion of yellow rice and comprehensive quality rating.

2. The method for detecting rice quality indicators according to claim 1, wherein: The effective feed particle indicators include particle size distribution, arrangement density, and input uniformity; the perfect grain ratio indicators include structural continuity ratio, low transmittance defect rate, and complete grain ratio; the rice grain quality classification results include contour integrity level, surface damage degree, and grain shape consistency index; the whole polished rice rate indicators include weight ratio, polished rice stable output ratio, and polished rice output ratio per unit raw material; the comprehensive quality rating includes color uniformity, polished rice purity, and yellow rice ratio level.

3. The method for detecting rice quality indicators in a whole process according to claim 1, wherein: The steps of inputting the effective feed particle index into the hulling chamber are specifically as follows: S101: collecting feed particle images through an infrared camera, screening rice grains according to the feed particle images, detecting the edge contours of the rice images according to the surface images of the rice grains, calculating pixel gradient changes in continuous areas of the image, locating the coordinates of the rice edge contours, analyzing the trend of texture interval changes, and obtaining rice edge parameters; S102: Based on the rice edge parameter set, identifying the morphological characteristics of a single rice grain, comparing the length-to-width ratio and surface area of a standard rice grain, determining the contour closure of each rice grain, identifying rice grains with abnormal morphology, analyzing the rice grain morphology distribution range, and generating rice grain morphology parameters; S103: Based on the rice morphological parameters, the number of rice grains that meet the standards is screened to obtain an index of effective feed grains input into the rice hulling chamber.

4. The method for detecting rice quality indicators in a whole process according to claim 1, wherein: The steps of structural integrity rating are specifically as follows: S201: Based on the effective feed particles, using an infrared transmission device to perform transmission detection on the brown rice after husking, collecting transmission data of each brown rice grain in the sample, analyzing the energy change of the transmitted light passing through the brown rice, calculating the transmission intensity value of a single brown rice grain, and generating a brown rice transmission intensity characteristic; S202: Calculating an average transmission intensity of the brown rice based on the brown rice transmission intensity characteristic, comparing the average transmission intensity with the transmission data of a single grain of brown rice, analyzing the degree of deviation of the transmission value, screening brown rice data with abnormal transmission values, determining the numerical range of the abnormal transmission intensity, and generating a transmission intensity deviation value; S203: Based on the brown rice transmission deviation value, analyzing the transmission uniformity of the brown rice, calculating the distribution of the transmission deviation, determining the structural characteristics of the brown rice with abnormal light intensity, screening the brown rice that meets the standard structure, classifying the structural integrity level, and generating a structural integrity rating.

5. The method for detecting rice quality indicators in a whole process according to claim 4, wherein: The transmission intensity value of a single grain of brown rice is calculated using the formula: ； Generate brown rice transmission intensity characteristics; in, Represents the transmittance intensity value of a single grain of brown rice, represents the incident light intensity, represents the optical absorption coefficient of brown rice, Represents the thickness of a single grain of brown rice. represents the intensity of transmitted scattered light, represents the light scattering correction parameter, Represents the base thickness.

6. The method for detecting rice quality indicators in a whole process according to claim 1, wherein: Said S3 further comprises: Based on the structural integrity rating, the polished rice image is denoised, the surface features of the brown rice are analyzed, the brightness fluctuation and texture interval are calculated, the milling sample is called, the data deviating from the milling range is identified, the uniformity of the milling is judged, and the milling quality analysis result is obtained. Based on the milling quality analysis result, the sample is transferred to the image acquisition area.

7. The method for detecting rice quality indicators in the whole process according to claim 6, wherein: The steps of classifying the rice grain quality results are specifically as follows: S301: Based on the structural integrity rating, image data of milled brown rice is retrieved and denoised, pixel information of the rice grain surface area is extracted, the spatial distribution of pixel grayscale values is analyzed, areas with prominent surface pixel brightness are screened, and brightness change values of the areas are calculated to generate rice grain surface brightness fluctuation data; S302: Based on the rice grain surface brightness fluctuation data, analyzing the distribution of pixel areas with prominent brightness changes, measuring the intervals between adjacent brightness extreme value points, calculating the offset amplitude of the brightness mutation interval, and determining the texture difference of the rice grain surface to generate the rice grain texture interval difference; S303: Based on the rice grain surface brightness fluctuation data and the rice grain texture interval difference, calling the rice grain milling sample data, screening the data points that exceed the normal milling range, comparing the surface brightness and texture variation range of each rice grain, analyzing the rice grain surface smoothness, calculating the milling uniformity offset, and obtaining the milling quality analysis result; S304: Based on the milling quality analysis results, the rice sample is transferred to an image acquisition area using a conveyor device, the contour boundary of a single rice grain is detected, the contour edge pixels are selected and connected to form a closed contour, the contour overlap area is calculated for the overlapping area, and the overlapping portion is morphologically segmented to obtain the rice grain contour features; S305: Analyzing the pixel distribution on the rice grain surface based on the rice grain contour features, screening the surface spot area according to the pixel brightness change, detecting the pixel fracture along the main axis, calculating the number and length of cracks, analyzing the structural integrity of the rice grain surface, and obtaining surface damage data; S306: Based on the surface damage data, the surface area of the rice grains is measured and compared with the spots and cracks, the total area ratio of the defective area is calculated, and the rice grains are classified according to the contour defect ratio. Rice grains with smooth and complete surfaces are classified as polished rice, and rice grains with excessive defects are classified as broken rice, thereby obtaining the rice grain quality classification result.

8. The method for detecting rice quality indicators in the whole process according to claim 7, wherein: The calculation of the offset amplitude of the brightness mutation interval adopts the formula: ； and determining the surface texture differences of rice grains to generate the texture interval differences of rice grains; in, Represents the offset amplitude of the brightness mutation interval, Representative The brightness of the maximum brightness point, Representative The brightness of the minimum brightness point, Represents the pixel interval between adjacent brightness extreme points, Represents the total number of detected brightness extreme point pairs.

9. The method for detecting rice quality indicators in a whole process according to claim 1, wherein: The steps of the head rice rate index are specifically as follows: S401: Based on the rice grain quality classification result, the rice grain conveying path is automatically assigned, the broken rice and polished rice are respectively introduced into the corresponding weighing channels, the cumulative weight data of each channel is recorded in real time, and the total amount of each type of rice grain is obtained; S402: Call the total amount of each type of rice grains, call the total weight of the rice grains, compare the cumulative weight of each type of rice grains with the total weight, calculate the weight proportions of broken rice and polished rice respectively, and determine the average proportion of polished rice in the total amount of rice based on the total amount of rice to generate a whole polished rice rate index.

10. A full-process detection device for rice quality indicators, characterized in that: The device stores a computer software program, and when the computer software program is executed by a processor, the method for detecting the whole process of rice quality indicators according to any one of claims 1 to 9 is implemented. The device comprises: An infrared image acquisition module collects images of feed particles using an infrared camera, screens rice particles based on the images of the feed particles, extracts rice grain texture and contour coordinates based on the surface images of the rice particles, identifies the position and morphology of the rice grains, and compares the sample rice grain texture to obtain an effective feed particle index; a transmission intensity detection module, which detects the infrared transmission intensity of brown rice based on the effective feed particle index, calculates the average transmission intensity, compares the transmission intensity of the sample brown rice with the average value, analyzes the structural characteristics of the brown rice, generates a structural integrity rating, and calculates the perfect grain ratio index and the imperfect grain ratio index based on the proportion of grains with qualified transmission intensity to generate a roughness rate index; The rice grain quality classification module performs image recognition on milled rice grains, identifies the contour image of a single rice grain, calculates the contour defect ratio, and classifies it into broken rice and polished rice to obtain the rice grain quality classification result; a weighing distribution module, which, based on the rice grain quality classification result, directs the rice grains into corresponding weighing channels, records the cumulative weight of the channels, compares the cumulative weight of the rice grains, calculates the weight proportion of polished rice, and generates a polished rice rate indicator; The qualified rate calculation module performs image recognition on the polished rice obtained by milling, identifies the proportion of yellow rice based on the image color, calculates the proportion of yellow rice in the total polished rice, calculates the final qualified rate based on the rough rice yield and whole polished rice rate indicators, and outputs the proportion of yellow rice and the comprehensive quality rating.

Images