Meal replacement powder grinding quality detection method

Abstract

The invention relates to the technical field of image processing, in particular to a meal replacement powder grinding quality detection method. The method comprises the following steps: acquiring an image of produced meal replacement powder, and uniformly dividing the image to obtain a plurality of sub-blocks; obtaining a spectrogram of each sub-block, and extracting a low-frequency region where a low-frequency component is located from the spectrogram; marking frequency points exceeding a set range in the low-frequency region as over-limit frequency points; obtaining an abnormal sub-block; obtaining dispersion of an intermediate frequency region where an intermediate frequency component of the spectrogram corresponding to the abnormal sub-block is located, and determining the abnormal sub-block when the dispersion is greater than a set value as a mildewed region; determining the abnormal sub-block when the dispersion is less than or equal to a set value as a caking area; and judging the quality of the meal replacement powder according to all mildewed areas or caking areas. According to the scheme, the quality of the meal replacement powder can be efficiently and accurately detected.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing. More specifically, the present invention relates to a method for detecting the grinding quality of meal replacement powder. Background Art

[0002] The detection of the grinding quality of meal replacement powder mainly serves the food processing and nutritional supplement industries, covering the production fields of functional foods such as protein powder, dietary fiber powder, and vitamin powder.

[0003] Since the quality of meal replacement powder is not only related to the safety, nutritional value, and market competitiveness of the product, but also directly affects the health and usage experience of consumers, the quality detection of meal replacement powder is crucial.

[0004] The quality problems of meal replacement powder usually manifest as mildew and caking. Among them, the occurrence of mildew is usually related to problems such as raw material contamination, unqualified production environmental hygiene, packaging process contamination, or storage and transportation links. The occurrence of caking is usually related to problems such as raw material characteristics, humidity influence (humidity in the production, storage, or transportation environment), grinding process, and packaging tightness. Therefore, the wide range of raw material sources, complex processing technology, and uncontrollable factors such as the grinding link and storage and transportation environment all affect the quality of meal replacement powder.

[0005] Among them, for the quality detection of meal replacement powder during production, the existing technology usually relies on a collaborative method of multiple detection technologies and manual sampling inspection. Due to the high investment in high-precision detection equipment, the need for detailed analysis of a large number of samples, and the experience judgment of professional quality inspection personnel in this collaborative method, the above collaborative method generally has problems such as low detection efficiency, high false detection rate, and insufficient sensitivity to agglomerates or mildew, etc., and it is difficult to meet the refined quality control requirements in scenarios of large-scale production and high-quality requirements, which directly affects the quality and safety of meal replacement powder and the consumption experience of consumers, and increases the food safety risk.

[0006] Therefore, how to accurately detect the quality of meal replacement powder, effectively identify problems such as agglomerates and mildew, improve the accuracy of the detection algorithm, and reduce the risk of missed detection, so as to avoid frequent downtime verification on the production line, cause waste of raw materials, lead to batch quality fluctuations in the finished product, and increase the pressure on the enterprise's quality control cost, is a particularly important issue. Summary of the Invention

[0007] The purpose of the present invention is to propose a method for detecting the grinding quality of meal replacement powder to solve the problem that the quality of meal replacement powder cannot be detected efficiently and accurately in the existing technology; for this purpose, the present invention provides a solution in the following aspect.

[0008] A method for detecting the grinding quality of meal replacement powder provided by the present invention includes: Obtain an image of the meal replacement powder produced, and evenly divide the image to obtain a plurality of sub-blocks; Obtain the spectrogram of each sub-block, and extract the low-frequency region where the low-frequency components are located from the spectrogram; Mark the frequency points exceeding the set range in the low-frequency region as over-limit frequency points; Calculate the first ratio of the number of all over-limit frequency points in the low-frequency region to the low-frequency components in the low-frequency region; Obtain the second ratio of the number of low-frequency components in the low-frequency region to all components of the spectrogram of the corresponding sub-block; If the first ratio is greater than or equal to the first threshold and the second ratio is greater than or equal to the second threshold, mark the corresponding sub-block as an abnormal sub-block; Obtain the dispersion degree of the intermediate-frequency region where the intermediate-frequency components of the spectrogram corresponding to the abnormal sub-block are located, and determine the abnormal sub-block with a dispersion degree greater than the set value as mildew; Determine the abnormal sub-block with a dispersion degree less than or equal to the set value as caking; Judge the quality of the meal replacement powder according to all mildew regions or caking regions.

[0009] The above solution divides the image into blocks and analyzes each sub-block after division, which can reduce the calculation amount; When analyzing, by first extracting the low-frequency region in each sub-block and the over-limit frequency points in the low-frequency region, two judgment conditions can be obtained (one is that the energy proportion of the over-limit frequency points in the low-frequency region is higher than the first threshold, and the other is that the low-frequency energy proportion in the low-frequency region is higher than the second threshold), and based on the two judgment conditions, preliminary screening is carried out to screen out abnormal sub-blocks; And through the dispersion degree of the intermediate-frequency region for secondary discrimination, the defect type of the abnormal sub-block is further determined, realizing the efficient and accurate detection of the quality of the meal replacement powder.

[0010] Optionally, the dispersion degree is: ; where is the mean value of all intermediate-frequency components of the intermediate-frequency region corresponding to the i-th sub-block, , are respectively the mean value and standard deviation of all intermediate-frequency components in the spectrogram corresponding to the historical normal image.

[0011] The above solution provides a method for accurately calculating the dispersion degree.

[0012] Optionally, the set range is a confidence interval, where the upper limit value of the confidence interval is the sum of the mean value of the amplitudes of all frequency points in the low-frequency region and the standard deviation multiplied by a set multiple, and the lower limit value of the confidence interval is the difference between the mean value of the amplitudes of all frequency points in the low-frequency region and the standard deviation multiplied by a set multiple.

[0013] The above can accurately obtain over-limit frequency points by setting the confidence interval.

[0014] Optionally, the low-frequency region is obtained by calculating the distance from each frequency point in the spectrogram to the center of the spectrum, marking the frequency points with a distance less than the first distance threshold as low-frequency points, and taking the region composed of all low-frequency points as the low-frequency region; the spectrogram is obtained by performing Fourier transform on each sub-block.

[0015] Optionally, the process of obtaining the intermediate-frequency region is as follows: Mark the frequency points in the spectrogram whose distance to the center of the spectrum is less than the second distance threshold and greater than or equal to the first distance threshold as intermediate-frequency components, and take the region composed of all intermediate-frequency components as the intermediate-frequency region.

[0016] The above solution can accurately extract intermediate-frequency components.

[0017] Optionally, the first distance threshold is , and the second distance threshold is ; where is the first coefficient, is the second coefficient, the second coefficient is greater than the first coefficient, P is the length of the spectrogram corresponding to the sub-block, Q is the width of the spectrogram corresponding to the sub-block, and min( ) is the minimum value function.

[0018] Optionally, it further includes the step of preprocessing the image, specifically: Perform image enhancement on the image to obtain an enhanced image; and use the fourth-order flat-top window method to smooth the enhanced image to obtain a smoothed image.

[0019] Optionally, the historical normal images are multiple normal images of non-bonded blocks and non-mildewed ones collected historically.

[0020] Optionally, the determination of the quality of the meal replacement powder based on all mildewed regions or caked regions includes: extracting the total number of pixel points belonging to mildew or caking in all mildewed regions or caked regions; when the ratio of the total number of pixel points corresponding to caking to the total number of pixel points of the image is greater than the first ratio threshold, it is determined that the caking content in the meal replacement powder exceeds the standard; when the ratio of the total number of pixel points corresponding to mildew to the total number of pixel points of the image is greater than the second ratio threshold, it is determined that the mildew content in the meal replacement powder exceeds the standard.

[0021] The above solution can accurately determine whether the quality of the meal replacement powder is qualified by analyzing the proportion of the caked region or mildewed region.

[0022] Optionally, the extraction of the total number of pixel points belonging to mildew or caking in all mildewed regions or caked regions includes: using the Canny edge detection algorithm to perform edge detection on each mildewed region or caked region to obtain the total number of pixel points with caking abnormalities or mildew abnormalities in the mildewed region or caked region.

[0023] The beneficial effects of the present invention are as follows: In the solution of the present invention, the image is analyzed in the frequency domain and judged twice successively. The first judgment is by adopting the "two-dimensional joint judgment strategy" (the proportion of the energy of the over-frequency points in the low-frequency region is higher than the first threshold and the proportion of the low-frequency energy in the low-frequency region is higher than the second threshold), and the second judgment is by adopting the dispersion degree of the intermediate-frequency components. That is, through the two judgments, the quality of the meal replacement powder is efficiently and accurately detected. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 The flowchart of the steps of a method for detecting the grinding quality of a meal replacement powder in this embodiment is schematically shown. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0025] Taking a certain meal replacement powder as an example, a method for detecting the grinding quality of a meal replacement powder of the present invention is introduced.

[0026] Specifically, as Figure 1 shown, a method for detecting the grinding quality of a meal replacement powder in this embodiment includes the following steps: Step S1, obtaining an image of the meal replacement powder and uniformly dividing the image to obtain a plurality of sub-blocks.

[0027] In this embodiment, an image of the meal replacement powder produced by the production line is collected by an industrial camera.

[0028] Among them, by uniformly dividing the image, a plurality of sub-blocks can be obtained.

[0029] In this embodiment, before performing the image division, it further includes the step of preprocessing the image. Specifically: the CLAHE algorithm is used to enhance the image to obtain an enhanced image; and the fourth-order flat-top window method is used to smooth the enhanced image to obtain a smoothed image.

[0030] It should be noted that applying the flat-top window to optimize the image can more accurately retain the true amplitude distribution of the low-frequency energy and at the same time suppress the interference of high-frequency noise on the edge signal. Specifically, the fourth-order flat-top window is applied to each pixel position in the image, and then the image preprocessed by CLAHE is multiplied by the fourth-order flat-top window to make the gray-scale gradient of the edge region of the windowed image smooth.

[0031] Furthermore, before performing the image division, in order to improve the resolution of the subsequent analyzed spectrum, a zero-padding operation is performed on the image, that is, the image size is extended from L×W to M×N, and after that, zero values are filled in the remaining places of the image to obtain a zero-padded image.

[0032] Step S2, screening out abnormal sub-blocks from the sub-blocks.

[0033] Specifically, the steps for obtaining the abnormal sub-blocks include the following steps: Step S21: Obtain the spectrogram of each sub-block, and extract the low-frequency region where the low-frequency components are located from the spectrogram.

[0034] In this embodiment, perform a two-dimensional fast Fourier transform on each sub-block to obtain the corresponding spectrogram.

[0035] The process for obtaining the low-frequency region in this embodiment is as follows: Calculate the distance from each frequency point in the spectrogram to the center of the spectrum, mark the frequency points with a distance less than the first distance threshold as low-frequency points, and use the region formed by all low-frequency points as the low-frequency region.

[0036] Specifically, taking the center (0, 0) of the spectrogram as the origin, calculate the distance from each frequency point to the center of the spectrum. When the distance is less than the first distance threshold, the frequency point belongs to the low-frequency region, and all frequency points form the low-frequency region. Among them, the value of the low-frequency component of the frequency points in the low-frequency region remains unchanged, and the values of other frequency points are set to 0.

[0037] The above low-frequency region is a circular region; it should be noted that the spectrograms obtained in this embodiment are all spectrograms after centering processing (the position on the image represents frequency, and the brightness represents energy), and the spectrogram can be regarded as a "map" of frequency. The center point (0, 0) of the map is the zero frequency, also called the DC component, which is the place with the lowest frequency. Moving outward from the center, the amplitude of the frequency is continuously increasing. Therefore, a preliminary low-frequency region can be delimited by distance.

[0038] Among them, the first distance threshold is , where P is the length of the spectrogram corresponding to the sub-block, Q is the width of the spectrogram corresponding to the sub-block, is the first coefficient, which can be adjusted according to the actual foreign object size, and the empirical value is 0.3.

[0039] Step S22: Mark the frequency points exceeding the set range in the low-frequency region as over-limit frequency points.

[0040] In this embodiment, analyze whether the low-frequency points in the low-frequency region are over-limit frequency points, that is, when the amplitude of the low-frequency points in the low-frequency region exceeds the set range, mark it as an over-limit frequency point in the low-frequency region. Among them, the set range is the confidence interval, where the upper limit value of the confidence interval is the sum of the mean value of the amplitudes of all frequency points in the low-frequency region and the standard deviation multiplied by the set multiple, and the lower limit value of the confidence interval is the difference between the mean value of the amplitudes of all frequency points in the low-frequency region and the standard deviation multiplied by the set multiple.

[0041] Specifically, the set range is , where k is the set multiple, which is also a coefficient adjusted according to the false detection rate tolerance (for example, when k = 3, it corresponds to a 99.7% confidence interval).

[0042] Step S23: Calculate the first ratio of the number of all over-limit frequency points in the low-frequency region to the low-frequency components in the low-frequency region; obtain the second ratio of the number of low-frequency components in the low-frequency region to all components of the frequency spectrum diagram of the corresponding sub-block; if the first ratio is greater than or equal to the first threshold and the second ratio is greater than or equal to the second threshold, then mark the corresponding sub-block as an abnormal sub-block.

[0043] In this embodiment, calculate the ratio of the number of over-limit frequency points in each low-frequency region to the total number of frequency points in the low-frequency region, and use this ratio as the first ratio; use the ratio of the total number of frequency points in the low-frequency region to the total number of all frequency points in the frequency spectrum diagram of the corresponding sub-block as the second ratio.

[0044] Among them, the first threshold is 0.8; the second threshold is 0.6; of course, as other implementation manners, it can also be determined according to the actual situation.

[0045] Step S3: Classify the abnormal sub-blocks to obtain the sub-blocks belonging to the caking category and the sub-blocks belonging to the mildew category, and finally determine the quality of the meal replacement powder.

[0046] It should be noted that in this embodiment, it is considered that: when the energy concentration degree of local abnormal frequency points in the low-frequency region is significantly high, if there are local defects such as agglomeration blocks in the image, the corresponding low-frequency region will present an energy spike characteristic, and at the same time, the low-frequency energy accounts for a relatively high proportion, reflecting the lack of high-frequency components. This may be because the continuity of the material structure is damaged, that is, the typical signal characteristic of the agglomeration block type defect.

[0047] For the mildew area, the local density abnormality caused by microbial activities or metabolite aggregation will also cause a significant increase in the proportion of low-frequency energy (it should be noted that due to the synergistic effect of local structural damage and metabolite aggregation caused by microbial metabolic activities, some mildew characteristics may also be manifested as a high proportion of high-frequency energy. In this embodiment, this situation is not considered, and only the situation where the mildew area may cause an increase in the proportion of low-frequency energy if accompanied by local density abnormality is considered). However, the destruction of texture continuity by the mycelial network or local cracks will be manifested as a significant increase in the standard deviation of the intermediate frequency amplitude spectrum; while the agglomeration defect has a smooth edge characteristic formed by the overall density change, and its corresponding intermediate frequency standard deviation is usually low, reflecting the uniformity of the texture distribution.

[0048] Based on the above differential characteristics, for the abnormal sub-blocks, it is necessary to further distinguish the defect types through the threshold determination of the dispersion degree of the intermediate frequency energy: when the dispersion degree exceeds the threshold, it is determined as mildew, otherwise, it is determined as caking.

[0049] Specifically, obtain the dispersion of the intermediate frequency region where the intermediate frequency components of the spectrogram corresponding to the abnormal sub-block are located. Determine that the abnormal sub-block is mildewed when the dispersion is greater than the set value; determine that the abnormal sub-block is caked when the dispersion is less than or equal to the set value.

[0050] In one embodiment, the frequency points less than the second distance threshold and greater than or equal to the first distance threshold are used as intermediate frequency points, and all the intermediate frequency points form an intermediate frequency region.

[0051] Among them, the second distance threshold is ; among them is the second coefficient, which can be determined according to experience specifically. The value in this embodiment can be 0.7, P is the length of the spectrogram corresponding to the sub-block, and Q is the width of the spectrogram corresponding to the sub-block.

[0052] In another embodiment, the extraction method of the intermediate frequency region can retain the intermediate frequency components through a set band-pass filter.

[0053] Among them, after obtaining the intermediate frequency region, statistical analysis is performed on the amplitude spectrum of the intermediate frequency region, that is, the mean value of the amplitude values of all intermediate frequency points in this region is calculated, and this mean value is used as the key feature quantity of the uniformity of the intermediate frequency signal distribution of the sub-block.

[0054] To avoid interference from misdetection sources such as the vibration frequency band of the device, multiple normal images of non-bonded blocks and mildewed ones are pre-acquired, and the standard deviation and mean value of the normal intermediate frequency region are calculated based on these images. After the mean value of the intermediate frequency components of the intermediate frequency region of the current sub-block is subjected to Z-score normalization processing through the normal mean value and amplitude spectrum standard deviation, the dispersion of the intermediate frequency is obtained.

[0055] Among them, the dispersion is: ; among them, is the mean value of all intermediate frequency components of the intermediate frequency region corresponding to the i-th sub-block, 、 are the mean value and standard deviation of all intermediate frequency components in the spectrogram corresponding to the historical normal images respectively.

[0056] The above historical normal images are multiple normal images of non-bonded blocks without mildew collected historically.

[0057] In this embodiment, a set value is set. When the dispersion is greater than the set value, it is determined that there is mildew at this time. Otherwise, it is determined that there is caking in the abnormal sub-block.

[0058] The above set value can be obtained through historical data statistics; specifically, multiple historical normal samples (without defects) and known mildew / caking samples can be collected, and the mean value of the historical dispersions of the intermediate frequency components of all samples is used as the set value.

[0059] Further, based on all the mildewed areas or caked areas, the quality of the meal replacement powder is determined, including: extracting the total number of pixel points belonging to mildew or caking in all the mildewed areas or caked areas; when the ratio of the total number of pixel points corresponding to caking to the total number of pixel points of the image is greater than the first ratio threshold, it is determined that the caking content in the meal replacement powder exceeds the standard; when the ratio of the total number of pixel points corresponding to mildew to the total number of pixel points of the image is greater than the second ratio threshold, it is determined that the mildew content in the meal replacement powder exceeds the standard.

[0060] Both the above-mentioned first ratio threshold and second ratio threshold can be obtained based on experience; exemplarily, both the first ratio threshold and second ratio threshold can be 0.03. Of course, the setting of the first ratio threshold and second ratio threshold can also be determined according to the actual situation.

[0061] In one embodiment, the Canny edge detection algorithm is used to perform edge detection on each mildewed area or caked area to obtain the total number of pixel points with abnormal caking or abnormal mildew in the mildewed area or caked area.

[0062] In another embodiment, image segmentation can also be performed on each mildewed area or caked area to obtain the target area of each mildewed area or caked area, and obtain the total number of pixel points in the target area.

[0063] In this embodiment, it is considered that the characteristics of the bonded block are high low-frequency energy ratio and low dispersion of intermediate-frequency energy. The former reveals the continuity of the overall structure, while the latter reflects the uniformity of the intermediate-frequency distribution. For the mildewed area, due to the aggregation of metabolites, the low-frequency energy ratio is high, but the dispersion of the intermediate-frequency energy is also high. Therefore, the caked area or mildewed area is distinguished by calculating the dispersion of the intermediate-frequency area.

[0064] The quality detection method of the solution of the present invention adopts a "two-dimensional joint determination strategy": first, preliminary screening is carried out through two dimensions of "the ratio of the energy of the over-limit frequency points in the low-frequency area is higher than the first threshold" and "the ratio of the low-frequency energy in the low-frequency area is higher than the second threshold". Among them, a high low-frequency energy ratio indicates a continuous change in the overall structure of the detection area (such as a large-scale bonded block), and a high ratio of the energy of abnormal frequency points in the low-frequency area further determines the existence of an abnormal area with local energy concentration (such as small particle caking or the initial lesion of mildew), so as to accurately delineate the area of interest suspected of being a bonded block or mildew. Subsequently, secondary discrimination is carried out through the dispersion of the intermediate-frequency area: if the dispersion is lower than the threshold corresponding to the mildew characteristics, it is confirmed that the area is a bonded block; otherwise, it points to defects such as mildew and texture damage, realizing the efficient and accurate detection of the quality of the meal replacement powder.

[0065] In the description of this specification, the meaning of "a plurality" is at least two, such as two, three or more, etc., unless otherwise specifically defined.

[0066] Although this specification has shown and described several embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many variations, changes and alternative methods will occur to those skilled in the art without departing from the spirit and scope of the present invention. It should be understood that various alternatives to the embodiments of the present invention described herein may be employed in practicing the present invention.

Claims

1. A method for detecting the grinding quality of a meal replacement powder, characterized in that Including: Obtain an image of the produced meal replacement powder, and evenly divide the image to obtain a plurality of sub-blocks; Obtain the spectrogram of each sub-block, and extract the low-frequency region where the low-frequency components are located from the spectrogram; Mark the frequency points exceeding the set range in the low-frequency region as over-limit frequency points; Calculate the first ratio of the number of all over-limit frequency points in the low-frequency region to the low-frequency components in the low-frequency region; Obtain the second ratio of the number of low-frequency components in the low-frequency region to all components of the spectrogram of the corresponding sub-block; If the first ratio is greater than or equal to the first threshold and the second ratio is greater than or equal to the second threshold, mark the corresponding sub-block as an abnormal sub-block; Obtain the dispersion of the intermediate-frequency region where the intermediate-frequency components of the spectrogram corresponding to the abnormal sub-block are located, and determine the abnormal sub-block with the dispersion greater than the set value as the mildew region; Determine the abnormal sub-block with the dispersion less than or equal to the set value as the caking region; Judge the quality of the meal replacement powder according to all mildew regions or caking regions.

2. The quality inspection method for grinding a meal replacement powder according to claim 1, characterized in that The dispersion degree is as follows: ; wherein, is the mean value of all intermediate frequency components in the intermediate frequency region corresponding to the i-th sub-block, , are respectively the mean value and standard deviation of all intermediate frequency components in the spectrogram corresponding to the historical normal image.

3. A method for detecting the grinding quality of a meal replacement powder according to claim 1, characterized in that, The set range is a confidence interval, where the upper limit value of the confidence interval is the sum of the mean value of the amplitudes of all frequency points in the low-frequency region and the standard deviation of the set multiple, and the lower limit value of the confidence interval is the difference between the mean value of the amplitudes of all frequency points in the low-frequency region and the standard deviation of the set multiple.

4. A method for detecting the grinding quality of a meal replacement powder according to claim 2, characterized in that, The low-frequency region is: calculate the distance from each frequency point in the spectrogram to the center of the spectrum, mark the frequency points with the distance less than the first distance threshold as low-frequency points, and use the region composed of all low-frequency points as the low-frequency region; The spectrogram is obtained by performing a Fourier transform on each sub-block.

5. A method for detecting the grinding quality of a meal replacement powder according to claim 4, characterized in that, The process of obtaining the intermediate-frequency region is: Mark the frequency points with the distance from each frequency point in the spectrogram to the center of the spectrum less than the second distance threshold and greater than or equal to the first distance threshold as intermediate-frequency components, and use the region composed of all intermediate-frequency components as the intermediate-frequency region.

6. A method for detecting the grinding quality of a meal replacement powder according to claim 5, characterized in that, The first distance threshold is , and the second distance threshold is ; where is the first coefficient, is the second coefficient, the second coefficient is greater than the first coefficient, P is the length of the spectrogram corresponding to the sub-block, Q is the width of the spectrogram corresponding to the sub-block, and min( ) is the function for taking the minimum value.

7. A method for detecting the grinding quality of a meal replacement powder according to claim 1, characterized in that, It also includes the step of preprocessing the image, specifically: Perform image enhancement on the image to obtain an enhanced image; and use the fourth-order flat-top window method to perform smoothing processing on the enhanced image to obtain a smoothed image.

8. A method for detecting the grinding quality of a meal replacement powder according to claim 2, characterized in that The historical normal images are multiple normal images without caking and mildew collected historically.

9. A method for detecting the grinding quality of a meal replacement powder according to claim 1, characterized in that, The judging of the quality of the meal replacement powder according to all mildew regions or caking regions includes: extracting the total number of pixel points belonging to mildew or caking in all mildew regions or caking regions; When the ratio of the total number of pixel points corresponding to caking to the total number of pixel points of the image is greater than the first ratio threshold, it is determined that the caking content in the meal replacement powder exceeds the standard; When the ratio of the total number of pixel points corresponding to mildew to the total number of pixel points of the image is greater than the second ratio threshold, it is determined that the mildew content in the meal replacement powder exceeds the standard.

10. A method for detecting the grinding quality of a meal replacement powder according to claim 9, characterized in that, The extracting the total number of pixel points belonging to mildew or caking in all mildew regions or caking regions includes: using the Canny edge detection algorithm to perform edge detection on each mildew region or caking region to obtain the total number of pixel points with caking abnormality or mildew abnormality in the mildew region or caking region.

Images