Intelligent identification method for mycelial growth characteristics in recovery period of factory-like bottle cultivation produced edible mushrooms

Abstract

The invention discloses an intelligent identification method for hypha growth characteristics in a recovery period of factory-like bottle cultivation produced edible mushrooms, and relates to the technical field of modern facility agriculture. Comprising the following steps: shooting on-site growth RGB images of hyphae at t moment in a bottle cultivation production edible fungus recovery period; performing water drop interference and fuzzy detection and processing on the RGB image to obtain a clear RGB image; carrying out blue light interference processing on the clear RGB image; an improved top cap self-adaptive binarization method is adopted, and a background-free hypha binarization image is obtained through accurate segmentation and extraction; and generating a gray histogram of the background-free hypha binary image, reading the distribution frequency number Gt of white pixel points with the pixel value of 255, and calculating hypha growth characteristics in the recovery period according to the Gt. According to the method, the growth characteristics of the bottle cultivation production edible fungi in the recovery period can be accurately calculated, and a phenotypic characteristic basis is provided for fruiting intelligent cultivation and mushroom house precise management and control of factory-like bottle cultivation production edible fungi.

Description

Technical Field

[0001] The present invention relates to the technical field of modern facility agriculture, and in particular to an intelligent identification method for mycelium growth characteristics in a recovery period of factory-based bottle-grown edible fungi. Background Art

[0002] In recent years, the factory-based production model of bottle-bag-cultivated edible fungi has been further developed, and the automation level of the entire production is relatively high. However, as the key to crop production, the intelligent degree of mushroom cultivation technology in mushroom houses is relatively low, and it is urgently needed to realize intelligent and precise regulation to improve the yield and quality of edible fungi. The mycelium of edible fungi provides nutrients for the subsequent kink points, buds, and mushroom buds of the fruiting body period of edible fungi, so the recovery of mycelium has a critical impact on the yield and quality of edible fungi. However, at present, it mainly relies on technicians to enter the mushroom house every day for inspection and rely on manual experience to observe the growth characteristics of mycelium in the recovery period. This method has a low degree of standardization, a large amount of labor, and is prone to misjudgment or missed judgment, and cannot achieve timely and precise regulation of mushroom environment parameters.

[0003] In recent years, scholars have begun to explore the digital extraction of hyphae characteristics, but most of them are to put the sclerotia into the center of the culture dish or sterile plate to make samples, and then inject nutrient solution for culture, and then detect and analyze the grown mycelium samples. This method is only suitable for the ideal laboratory environment, and is used for the evaluation and identification of edible fungus germplasm resources, and cannot meet the complex interference of the factory-based on-site mushroom production environment. The mycelium in the recovery period is not like the mycelium in the culture dish of the strain research, which is complete and integrated. The mycelium in the recovery period has been destroyed by the fungus and is not integrated. Moreover, the culture medium and the mycelium are deeply integrated, which makes it difficult to better segment and extract the mycelium from the culture medium. Because the light in the mushroom house is unstable, the traditional fixed threshold binarization method is easy to misjudge the background as mycelium. In addition, in recent years, many edible fungi varieties have begun to use blue light instead of the original white light, but the opening of blue light will cause the mycelium to be unable to be accurately identified in the collected image, affecting the extraction of mycelium and the calculation of growth characteristics. In the existing technology, there is still no effective method for processing blue light sources in the method of edible fungus mycelium detection. In addition, the mycelium characteristics that are of concern in fungus culture and the mycelium growth characteristics that are of concern in the recovery period of the fruiting stage and serve as the basis for environmental regulation are also different. For example, existing studies have divided the image area into multiple concentric circles with the center of the culture dish as the center, and calculated the average pixel density in each concentric circle to evaluate the uniformity of the distribution. However, this method only considers the radial density distribution, that is, the gradient from the center to the outside, thereby ignoring the directional information, and is less applicable to non-circular or irregular distributions.

[0004] Therefore, it is an urgent problem for those skilled in the art to propose an intelligent identification method for mycelium growth characteristics in the recovery period of factory-based bottle-grown edible fungi to solve the difficulties existing in the prior art. Summary of the invention

[0005] In view of this, the present invention provides an intelligent identification method for the growth characteristics of mycelium in the recovery period of factory-based bottle-cultivated edible fungi, which can more accurately calculate the growth characteristics of the recovery period of bottle-cultivated edible fungi, and provide a phenotypic characteristic basis for the intelligent cultivation of mushrooms and precise control of mushroom houses in factory-based bottle-cultivated edible fungi, which helps to improve the production output and quality of factory-based bottle-cultivated edible fungi and effectively promote the intelligent production of edible fungi facility agriculture.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] An intelligent identification method for mycelium growth characteristics of edible fungi in the recovery period of factory-scale bottle cultivation, comprising:

[0008] S1, photographing and obtaining the on-site growth RGB image of mycelium at time t during the recovery period of bottle-grown edible fungi;

[0009] S2, performing water drop interference and blur detection and processing on the acquired RGB image to obtain a clear RGB image;

[0010] S3, performing blue light interference processing on the clear RGB image;

[0011] S4, converting the image processed by blue light interference into a binary image with enhanced foreground objects by using a top-hat adaptive binarization method;

[0012] S5, searching for all contours in the binary image of the enhanced foreground object, selecting the contour with the largest area, and calculating the circumscribed rectangle and the center point of the rectangle based on the contour with the largest area;

[0013] S6, draw a circle with the longest distance from the center point of the rectangle to the rectangular border as the radius, retain the white pixels in the circular area, and replace all the pixels outside the circular area with black pixels to obtain a hyphae binary image without background;

[0014] S7, generating a grayscale histogram of the background-free mycelium binary image, and reading the distribution frequency number G of the white pixel points with a pixel value of 255 in the histogram t , according to G t The mycelial growth characteristics during the recovery period were calculated.

[0015] In the above method, optionally, in S2, the water drop interference and blur detection is specifically as follows: converting the RGB image of the mycelium on-site growth in the recovery period into a grayscale image, processing the grayscale image using the Laplace operator, and calculating the Laplace variance Q of the image;

[0016] If Q max When it is determined that there are small water droplets in the image, an alarm is prompted and the lens is scraped, and then the image is re-taken to obtain the RGB image of the on-site growth of mycelium in the recovery period;

[0017] If Q>Q max , determine that the image is clear.

[0018] In the above method, optionally, in S3, an improved IAWB-GWPR-CLAHE algorithm is used to perform blue light interference processing on the clear RGB image, specifically:

[0019] S301, converting a clear RGB image from an RGB color space to an HSV color space to obtain an HSV image;

[0020] S302, calculate blue light intensity R B :

[0021]

[0022] Among them, R B Indicates the proportion of blue pixels in the HSV image, N B Indicates the number of blue pixels in the HSV image, N Total Indicates the total number of pixels in the HSV image;

[0023] S303, by brightening factor function L(R B ) Calculate the brightening factor L;

[0024] S304, using the improved IAWB-GWPR-CLAHE algorithm to process the image and calculate IGC(I):

[0025] IGC(I)=f1(L)×IG(I)+f2(L)×GLAHE(I)

[0026]

[0027] Where I represents the input image, IG(I) and CLAHE(I) represent the IAWB-GWPR algorithm and CLAHE algorithm respectively, 0<R B <R Bmin , R Bmin <R B <R Bmid , R Bmid <R B <R Bmax , R Bmax <R B <1 indicates the level of blue light intensity: no blue light, mild blue light, moderate blue light and severe blue light;

[0028] ​If IGC(I) = 0, no blue light processing is required;

[0029] If IGG(I)=IG(I), use the improved automatic white balance algorithm IAWB-GWPR to process blue light;

[0030] If IGC(I)=CLAHE(I), use contrast-limited adaptive histogram equalization algorithm CLAHE for blue light processing;

[0031] If IGC(I)=IG(I)+CLAHE(I), the improved automatic white balance algorithm IAWB-GWPR and the contrast-limited adaptive histogram equalization algorithm CLAHE are used to process blue light.

[0032] The above method, optionally, L(R B ) function is obtained by adjusting different blue light intensities R B , adjust the brightening factor to obtain R B The corresponding ideal brightening factor L is collected through experiments. B ,L) sample data, the blue light brightening factor L is fitted by polynomial as the blue light intensity R B of the changing relationship.

[0033] In the above method, optionally, in S4, the top-hat adaptive binarization method is specifically:

[0034] S401, performing a top-hat transformation on the image after the blue light interference processing to remove the uneven illumination, and obtaining I top ;

[0035] S402, performing Gaussian smoothing with a 15×15 Gaussian kernel;

[0036] S403, calculating the local adaptive binarization image threshold T of the image P(x, y) after top-hat smoothing local (x,y), according to T local (x, y) to achieve accurate binarization processing, and convert T local (x, y) is compared with the gray value T(x, y) of the pixel in the binary image to obtain the binary image P'(x, y) of the enhanced foreground object;

[0037]

[0038] Among them, P'(x,y) represents the final output top-hat adaptive enhanced foreground object binary image, T(x,y) represents the gray value corresponding to each pixel in the binary image, B(x,y) represents the offset for adjusting the image brightness or contrast, T localrepresents the local adaptive binary image threshold, P(x,y) represents the image after Gaussian smoothing, I top (x, y) represents the image after top-hat transformation, I(x, y) represents the original image, O(I(x, y)) represents the image after opening the original image, K represents the normalization factor, k represents the radius of the Gaussian kernel, G 15×15 (i, j) represents a Gaussian kernel of size 15×15, and w represents a rectangular local window of size w×w.

[0039] In the above method, optionally, in S7, the mycelium growth characteristics in the recovery period include: mycelium density, mycelium distribution uniformity, mycelium coverage, and mycelium growth rate.

[0040] The above method, optional, mycelium concentration and whiteness are measured by G t It indicates the degree of recovery of mycelium during the recovery period.

[0041] In the above method, the method for calculating the uniformity of mycelium distribution is optional: extract the coordinates of each white pixel representing mycelium in the background-free mycelium binary image, calculate the eigenvalues ​​λ1 and λ2 of the two-dimensional Gaussian matrix of all coordinate points, and calculate the uniformity coefficient H:

[0042]

[0043] When the uniformity coefficient H = 1: it means that the eigenvalues ​​are completely equal and the distribution is completely uniform;

[0044] When the uniformity coefficient H is close to 1: it means that the distribution tends to be uniform;

[0045] When the uniformity coefficient H is greater than 1: it indicates uneven distribution.

[0046] The above method, optionally, the mycelium coverage W F The calculation method is: the number of white pixels N in the binary image of hyphae without background V and the total number of pixels N in the area P The ratio is calculated as:

[0047]

[0048] In the above method, the mycelium growth rate can be calculated by: counting the distribution frequency of white pixels within the time (t2-t1) G t The rate of change is calculated as:

[0049]

[0050] Among them, V GSV represents the mycelial growth rate, G t2 , G t1They represent the distribution frequency of white pixels at time t2 and time t1 respectively, and t2 and t1 represent the time points when automatic photography is taken.

[0051] It can be seen from the above technical solutions that, compared with the prior art, the present invention provides a method for intelligently identifying the growth characteristics of mycelium in the recovery period of edible fungi produced in factory-based bottle cultivation, which has the following beneficial effects: 1) The present invention can intelligently identify and calculate the growth characteristics of mycelium in the recovery period in factory-based bottle cultivation mode and highly humid environment; 2) A water droplet detection and processing method based on Laplace transform variance can obtain high-quality growth images in a high-humidity environment; 3) The top-hat adaptive binarization algorithm is adopted to effectively avoid the deficiency that the traditional fixed threshold binarization method easily misjudges the background as mycelium, and can better handle images with complex culture medium background and uneven lighting, avoid inaccurate mycelium extraction caused by background interference, and achieve more accurate segmentation of mycelium body and complex background; 4) The blue light brightening factor function determination method and the improved IAWB-GWPR-CLAHE algorithm overcome the limitations of traditional methods in handling light and moderate blue light in image color cast detection, and the noise amplification problem during heavy blue light histogram equalization, and significantly reduce The influence of blue light is reduced; 5) For the first time, a calculation method for the digital characteristics of mycelium density, mycelium distribution uniformity, and mycelium coverage based on the growth characteristics of mycelium in the recovery period of bottle-grown edible fungi is proposed, and a calculation method for the mycelium growth rate in the recovery period based on the mycelium density is designed; 6) A calculation method for the uniformity of mycelium distribution based on the two-dimensional Gaussian distribution covariance matrix, by analyzing the eigenvalues ​​of the covariance matrix of the two-dimensional Gaussian distribution of the mycelium binary image, and proposing a uniformity coefficient to quantify the overall density distribution characteristics of the image, by describing the scalability in different directions, it can provide richer spatial distribution information, and has higher accuracy and directional independence in more complex and irregularly shaped distribution scenes; 7) The present invention provides ideas and methods for the intelligent identification and calculation of the growth characteristics of the recovery period of edible fungi in the fruiting stage, which can effectively provide standardized and quantitative evaluation for the growth status in the recovery period, and provide a basis for the intelligent environmental regulation decision-making during this period, which can greatly promote the process of intelligent and precise environmental regulation of edible fungi. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0053] Figure 1 A flow chart of an intelligent identification method for mycelium growth characteristics during the recovery period of factory-scale bottle-grown edible fungi provided by the present invention;

[0054] Figure 2 The present invention provides an image with interference from small water drops.

[0055] Figure 3 This is a re-photographed image of the mycelium of the seafood mushroom in the recovery stage without blue light.

[0056] Figure 4a for Figure 3 The result of traditional fixed threshold binarization processing;

[0057] Figure 4b for Figure 3 The result image after adaptive binarization processing;

[0058] Figure 4c for Figure 3 The result image after top-hat adaptive binarization processing;

[0059] Figure 5 for Figure 4c The result of extracting the contour with the largest area and then processing it to obtain the background-free hyphae binarization;

[0060] Figure 6a This is the calculation result diagram of the white concentration, uniformity coefficient H, uniformity and coverage of the mycelium of the seafood mushroom in the recovery period without blue light;

[0061] Figure 6b It is a scatter plot showing the uniformity of mycelium of seafood mushroom during the recovery period without blue light;

[0062] Figure 7a For shooting Figure 3 An image of the mycelium of the seafood mushroom in the recovery stage taken 6 hours later without blue light;

[0063] Figure 7b for Figure 7a The result image after top-hat adaptive binarization processing;

[0064] Figure 8a This is an image of the seafood mushroom recovery growth period taken under moderate blue light;

[0065] Figure 8b for Figure 8a The result image after blue light processing;

[0066] Figure 8c For Figure 8a The corresponding image of the seafood mushroom recovery growth period taken without blue light;

[0067] Figure 8d for Figure 8b The result image after top-hat adaptive binarization processing;

[0068] Fig. 9 for Figure 8dThe result of extracting the contour with the largest area and then processing it to obtain the background-free hyphae binarization;

[0069] Fig.10a This is the calculation result diagram of the white concentration, uniformity coefficient H, uniformity and coverage of the mycelium of the seafood mushroom in the recovery period under moderate blue light;

[0070] Fig.10b This is a scatter plot showing the uniformity of mycelium during the recovery period of seafood mushrooms under moderate blue light;

[0071] Fig.11a This is an image of the seafood mushroom's recovery growth period taken under heavy blue light;

[0072] Fig.11b for Fig.11a The result image after blue light processing;

[0073] Fig.11c For Fig.11a The corresponding image of the seafood mushroom recovery growth period taken without blue light;

[0074] Fig.11d This is the result of top-hat adaptive binarization processing of 11b;

[0075] Fig.12 for Fig.11d The result of extracting the contour with the largest area and then processing it to obtain the background-free hyphae binarization;

[0076] Fig.13a This is the calculation result diagram of the whitening degree, uniformity coefficient H, uniformity and coverage of the mycelium of the seafood mushroom in the recovery period under severe blue light;

[0077] Fig.13b It is a scatter plot to visualize the uniformity of mycelium in the recovery period of severe blue light seafood mushroom;

[0078] Fig.14 This is a function curve of the blue light brightening factor of seafood mushroom fitted using 14 sets of data. DETAILED DESCRIPTION

[0079] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0080] In this application, relational terms such as first and second, etc. are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. The terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of more restrictions, the elements defined by the sentence "comprise one..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0081] Reference Figure 1 As shown, the present invention discloses an intelligent identification method for mycelium growth characteristics in the recovery period of factory-based bottle-grown edible fungi, comprising:

[0082] S1, photographing and obtaining the on-site growth RGB image of mycelium at time t during the recovery period of bottle-grown edible fungi;

[0083] S2, performing water drop interference and blur detection and processing on the acquired RGB image to obtain a clear RGB image;

[0084] S3, performing blue light interference processing on the clear RGB image;

[0085] S4, converting the image processed by blue light interference into a binary image with enhanced foreground objects by using a top-hat adaptive binarization method;

[0086] S5, searching for all contours in the binary image of the enhanced foreground object, selecting the contour with the largest area, and calculating the circumscribed rectangle and the center point of the rectangle based on the contour with the largest area;

[0087] S6, draw a circle with the longest distance from the center point of the rectangle to the rectangular border as the radius, retain the white pixels in the circular area, and replace all the pixels outside the circular area with black pixels to obtain a background-free hyphae binary image;

[0088] S7, generating a grayscale histogram of the background-free mycelium binary image, and reading the distribution frequency number G of white pixels with a pixel value of 255 in the histogram t , according to G t The mycelial growth characteristics during the recovery period were calculated.

[0089] Furthermore, in S2, water drop interference and blur detection are specifically as follows: converting the RGB image of the mycelium on-site growth in the recovery period into a grayscale image, processing the grayscale image using the Laplace operator, and calculating the Laplace variance Q of the image;

[0090] If Q max When it is determined that there are small water droplets in the image, an alarm is prompted and the lens is scraped, and then the image is re-taken to obtain the RGB image of the on-site growth of mycelium in the recovery period;

[0091] If Q>Q max , determine that the image is clear;

[0092] Specifically, Q max The value of is determined experimentally. First, a large number of mycelium growth images in the recovery period are collected, including images containing water droplets and clear images. Then, the Laplace variance Q of these images is calculated, and the Q value is statistically analyzed. By observing the Q value distribution of water droplet images and clear images, a dividing point Qmax is selected to ensure that the water droplet and clear images can be effectively distinguished on the basis of this value. The specific value of Qmax may be fine-tuned according to factors such as the lighting conditions and resolution of the image to ensure the accuracy of the detection results.

[0093] Furthermore, in S3, the improved IAWB-GWPR-CLAHE algorithm is used to process the blue light interference of the clear RGB image, specifically:

[0094] S301, converting a clear RGB image from an RGB color space to an HSV color space to obtain an HSV image;

[0095] S302, calculate blue light intensity R B :

[0096]

[0097] Among them, R B Indicates the proportion of blue pixels in the HSV image, N B Indicates the number of blue pixels in the HSV image, N Total Indicates the total number of pixels in the HSV image;

[0098] S303, by brightening factor function L(R B ) Calculate the brightening factor L;

[0099] S304, using the improved IAWB-GWPR-CLAHE algorithm to process the image and calculate IGC(I):

[0100] IGC(I)=f1(L)×IG(I)+f2(L)×CLAHE(I)

[0101]

[0102] Where I represents the input image, IG(I) and CLAHE(I) represent the IAWB-GWPR algorithm and CLAHE algorithm respectively, 0<R​B <R Bmin , R Bmin <R B <R Bmid , R Bmid <R B <R Bmax , R Bmax <R B <1 indicates the level of blue light intensity: no blue light, mild blue light, moderate blue light and severe blue light; R Bmax Determination of blue light intensity R B Combined with the blue light on time, the image performs well in CLAHE, but it is difficult to remove the blue light in IAWB-GWPR, and the blue light is aggravated; R Bmax <R B : Corresponding to the heavy blue light image, the blue light intensity R B Very high, and the blue light on time is very long, and as the blue light on time continues to increase, the blue light intensity R B No longer increase; R Bmin , R Bmid Determination of blue light intensity R B Combined with the blue light on time, the image performs well in IAWB-GWPR, but it is difficult to effectively distinguish hyphae from the background in CLAHE; <R B <R Bmin : Corresponding to the image without blue light: that is, the blue light intensity R is calculated for the image when the blue light is not turned on B , through a large amount of data, select a dividing point to distinguish the critical value of blue light on and off; R Bmin <R B <R Bmid : Corresponding to the light blue light image, the blue light intensity R B Usually lower, and the blue light is on for a shorter time; R Bmid <R B <R Bmax : Corresponding to the moderate blue light image, its blue light intensity R B The higher the light intensity, and the longer the blue light is on, the more gradual the impact becomes.

[0103] If IGC(I) = 0, no blue light processing is required;

[0104] If IGC(I)=IG(I), use the improved automatic white balance algorithm IAWB-GWPR to process blue light;

[0105] If IGC(I)=CLAHE(I), use contrast-limited adaptive histogram equalization algorithm CLAHE for blue light processing;

[0106] If IGC(I)=IG(I)+CLAHE(I), the improved automatic white balance algorithm IAWB-GWPR and the contrast-limited adaptive histogram equalization algorithm CLAHE are used to process blue light.

[0107] Furthermore, a linear regression analysis was performed on the measured values ​​of the improved IAWB-GWPR-CLAHE algorithm and the other three blue light processing algorithms (when the blue light was turned on: no blue light treatment, only using the IAWB-GWPR algorithm, and only using the CLAHE algorithm) under the blue light on and the blue light off. The improved algorithm had a higher correlation with the blue light off, and its linear regression fitting value reached 0.94, and the RMSE was smaller than that of the other three methods, which minimized the impact of the blue light source on the mycelium density and whiteness detection.

[0108] Furthermore, L(R B ) function is obtained by adjusting different blue light intensities R B , adjust the brightening factor to obtain R B The corresponding ideal brightening factor L is collected through experiments. B ,L) sample data, the blue light brightening factor L is fitted by polynomial as the blue light intensity R B of the changing relationship.

[0109] Furthermore, in S4, the top-hat adaptive binarization method is specifically:

[0110] S401, performing a top-hat transformation on the image after the blue light interference processing to remove the uneven illumination, and obtaining I top ;

[0111] S402, performing Gaussian smoothing with a 15×15 Gaussian kernel to reduce noise interference;

[0112] S403, calculating the local adaptive binarization image threshold T of the image P(x, y) after top-hat smoothing local (x,y), according to T local (x, y) to achieve accurate binarization processing, and convert T local (x, y) is compared with the gray value T(x, y) of the pixel in the binary image to obtain the binary image P'(x, y) of the enhanced foreground object;

[0113]

[0114] Among them, P'(x,y) represents the final output top-hat adaptive enhanced foreground object binary image, T(x,y) represents the gray value corresponding to each pixel in the binary image, B(x,y) represents the offset for adjusting the image brightness or contrast, T localrepresents the local adaptive binary image threshold, P(x,y) represents the image after Gaussian smoothing, I top (x, y) represents the image after top-hat transformation, I(x, y) represents the original image, O(I(x, y)) represents the image after opening the original image, K represents the normalization factor, k represents the radius of the Gaussian kernel, G 15×15 (i, j) represents a Gaussian kernel of size 15×15, and w represents a rectangular local window of size w×w;

[0115] Compared with the traditional fixed threshold binarization method, the problem of misjudging the background as the target area in the traditional method is effectively avoided, and the image with complex culture medium background and uneven illumination can be better processed, and the accuracy of mycelium extraction is further improved; the measured values ​​of the present invention and the manual measurement interval are compared and analyzed, and the in-range proportion (In-range Proportion, IRP) represents the proportion of all calculated values ​​in the manual measurement interval, and the relative deviation (Relative Deviation, RD) represents the ratio of the deviation between the measured values ​​of the present invention and the manual measurement interval relative to the standard interval. The reliability of the present invention and the manual measurement value is relatively high, and the in-range proportion IRP of the concentration and coverage is 89.21% and 89.52% respectively, and the relative error RD is 11.87% and 10.84% ​​respectively, which is within the allowable range, indicating that the present invention has good practicality.

[0116] Furthermore, in S7, the mycelium growth characteristics during the recovery period include: mycelium density, mycelium distribution uniformity, mycelium coverage, and mycelium growth rate.

[0117] Furthermore, the degree of mycelium whiteness is measured by G t It indicates the degree of recovery of mycelium during the recovery period. When the degree of mycelium whitening is within the specified range during the specified growth period, it indicates that the recovery state is normal. The degree of whitening can be divided into four different ranges, namely 0<G t <G tmin , G tmin <G<G tmid , G tmid <G<G tmax , G tmax <G corresponds to four levels of mycelium whitening: no whitening, mild whitening, moderate whitening, and severe whitening.

[0118] Furthermore, the method for calculating the uniformity of mycelium distribution is as follows: extract the coordinates of each white pixel representing mycelium in the binary image of mycelium without background, calculate the eigenvalues ​​λ1 and λ2 of the two-dimensional Gaussian matrix of all coordinate points, and calculate the uniformity coefficient H:

[0119]

[0120] When the uniformity coefficient H = 1: it means that the eigenvalues ​​are completely equal and the distribution is completely uniform;

[0121] When the uniformity coefficient H is close to 1: it means that the distribution tends to be uniform;

[0122] When the uniformity coefficient H is greater than 1: it indicates uneven distribution.

[0123] Furthermore, the mycelium coverage W F The calculation method is: the number of white pixels N in the binary image of hyphae without background V and the total number of pixels N in the area P The ratio is calculated as:

[0124]

[0125] Furthermore, the calculation method of mycelium growth rate is: count the distribution frequency of white pixels in the time (t2-t1) G t The rate of change is calculated as:

[0126]

[0127] Among them, V GSV represents the mycelial growth rate, G t2 , G t1 They represent the distribution frequency of white pixels at time t2 and time t1 respectively, and t2 and t1 represent the time points when automatic photography is taken.

[0128] In a specific embodiment, if the seafood mushroom recovery growth image has no blue light, the specific steps of intelligently identifying its growth characteristics are:

[0129] Reference Figure 2 As shown, the on-site growth images of the mycelium of the seafood mushroom during the recovery period are photographed and obtained;

[0130] Convert the mycelium growth image of seafood mushroom in the recovery period from RGB image to grayscale image;

[0131] Then use the Laplace operator to process the grayscale image and calculate the Laplace variance of the image Q = 171.44; in seafood mushrooms, take Q max =200, because Q<200, it is determined that there are small water droplets in the image, an alarm is prompted and the lens is scraped, and then the image of mycelium growth in the recovery period is taken again. Figure 3 As shown;

[0132] To deal with blue light interference, Figure 3 The RGB image is converted to an HSV image and the number of blue pixels N is calculated as follows: B =139, blue light intensity R B=0.24%,

[0133]

[0134] Among them, N B Indicates the number of blue pixels in the image, N Total Indicates the total number of pixels in the image;

[0135] According to the different durations of blue light on-site in the seafood mushroom growing room, the blue light intensity can be divided into four levels: no blue light, mild blue light, moderate blue light and severe blue light, corresponding to the four intervals of 0%-1%, 1%-50%, 50%-99%, and 99%-100% respectively. Figure 3 The blue light intensity R B =0.24%, then Figure 3 The blue light intensity level was determined to be no blue light;

[0136] Through the blue light brightening factor function L(R B ), calculate R B The corresponding brightening factor L = 1.24, then f1(L) = 0, f2(L) = 0 in the improved IAWB-GWPR-CLAHE algorithm, that is, IGC(I) = 0. The calculation results show that blue light processing is not required;

[0137] Using the top-hat adaptive binarization method Figure 3 Convert to binary image: First, perform top-hat transformation on the image to remove uneven illumination, then perform Gaussian smoothing with a 15×15 Gaussian kernel to reduce noise interference, and then calculate the local adaptive threshold of the image after top-hat smoothing to achieve accurate binarization. The top-hat adaptive binarization processing result is as follows: Figure 4a As shown;

[0138] The results of the traditional fixed threshold binarization method are as follows Figure 4b As shown in Figure 1, the uneven light intensity has a significant impact on the extraction of nutrients as hyphae. The results after processing by the adaptive binarization method are shown in Figure 1. Figure 4c As shown in the figure, although the problem of uneven illumination is solved, the complex background is misjudged as hyphae, which interferes with the calculation of subsequent hyphae growth characteristics. Compared with these two methods, the method proposed in the present invention effectively avoids the problem of misjudging nutrients as hyphae due to uneven illumination in the traditional method, and the problem of misjudging complex background as hyphae in the adaptive binarization method.

[0139] Extract the contour with the largest area to obtain a background-free mycelium binary image, such as Figure 5 As shown;

[0140] Count the frequency G of white pixels in the grayscale histogram of the binary image of hyphae without background t, further analyzed the growth characteristics of seafood mushrooms during the recovery period: the degree of mycelium density, mycelium distribution uniformity and mycelium coverage. The results are as follows Figure 6a As shown;

[0141] The calculation method of the whiteness of the mycelium of seafood mushroom is as follows: calculate the distribution frequency of white pixel points G t , through four G t The values ​​divide the degree of thick whiteness of the mycelium growth of the seafood mushroom into four different intervals, namely 0-500, 500-5500, 5500-15000 and 15000-25000, which correspond to the four levels of thick whiteness of the mycelium: no thick whiteness, slight thick whiteness, moderate thick whiteness and severe thick whiteness; the calculation result of the thick whiteness is G t =12953, indicating medium-rich white;

[0142] The method for calculating the uniformity of the mycelium distribution of oyster mushrooms is as follows: extract the coordinates of each white pixel representing the mycelium in the binary mycelium image, calculate the two-dimensional Gaussian matrix eigenvalues ​​λ1 and λ2 of these coordinate points, and calculate the uniformity coefficient H according to the following formula. The larger H is, the more uneven the distribution is.

[0143]

[0144] When the uniformity coefficient H = 1: it means that the eigenvalues ​​are completely equal and the distribution is completely uniform;

[0145] When the uniformity coefficient H is close to 1: it means that the distribution tends to be uniform;

[0146] When the uniformity coefficient H is greater than 1 (e.g., 2.0 or higher): it indicates uneven distribution;

[0147] Mycelium of seafood mushroom during recovery period Figure 5 The calculation results of the uniformity coefficient are: λ1=2721.95, λ2=2968.32, H=1.09, indicating that the distribution tends to be uniform, and the visual scatter plot of the fitting results is as follows: Figure 6b shown.

[0148] The calculation method of the mycelium coverage of seafood mushroom is as follows: the number of pixels of the binary image of mycelium without background is identified, and the number of white pixels N in the mycelium area is calculated. V The total number of pixels in the bottle mouth area N P The ratio is calculated as:

[0149]

[0150] Among them, W F Indicates mycelium coverage, N V Indicates the number of pixels in the hyphae area, N P Indicates the number of pixels in the bottle mouth area;

[0151] Mycelium coverage rate of seafood mushroom W F Divided into four different intervals, namely 0%-20%, 20%-50%, 50%-70%, 70%-100%, corresponding to the four levels of coverage: no coverage, less coverage, more coverage, and basic coverage;

[0152] Figure 5 The calculation result of coverage is: N v =20141,N P =41734, 20%<W F =48.26%<50%, indicating less coverage;

[0153] The mycelium growth rate V is calculated by dividing the change in the white pixel distribution frequency Gt during the time (t2-t1) by the corresponding time. GSV ;

[0154] Figure 7a Shown Figure 3 The image of the culture bottle was taken after 6 hours. Figure 7b is the result of the top-hat adaptive binarization of the image, and its Figure 3 Medium Calculation Therefore, the growth rate of the culture bottle is Frequency / hour;

[0155] The blue intensity R was obtained through field experiments. B Compared with the 14 groups of sample data corresponding to the brightening factor L {(51.04%, 1.227), (22.86%, 1.14), (32.90%, 1.225), (70.53%, 1.228), (99.66%, 1.26), (99.49%, 1.25), (0.38%, 1.25), (0.24%, 1.24), (0.46%, 1.23), (87.99%, 1.23), (94.29%, 1.24), (99.50%, 1.246), (99.43%, 1.248), (99.44%, 1.248)}, the blue light brightening factor L is fitted with the blue light intensity R B The transformation function L(R B ), as shown below, is a sixth-order polynomial fitting formula, with a determination coefficient R 2 =0.9661, with high fitting accuracy, and realizes adaptive brightening adjustment for different blue light intensities. The fitting function curve of the seafood mushroom brightening factor is as follows: Fig.14 As shown:

[0156]

[0157] In another specific embodiment, if the image of the seafood mushroom recovery period has moderate blue light, the specific steps of intelligently identifying its growth characteristics are:

[0158] Take and obtain on-site growth images of seafood mushroom mycelium during the recovery period, such as Figure 8a As shown;

[0159] The mycelium growth image of the seafood mushroom recovery period is converted from an RGB image to a grayscale image, and then the grayscale image is processed by the Laplace operator, and the Laplace variance of the image is calculated to be Q=338.92>200, and it is determined that there are no small water droplets in the image and the image is clear, and further processing is performed based on the image later;

[0160] To deal with blue light interference: Figure 8a Convert the RGB image to an HSV image and calculate the number of blue pixels N in the image B =55398, blue light intensity R B =70.53%,50% <R B <99%, then Figure 8a The blue light level is judged as moderate blue light;

[0161] Through the blue light brightening factor function L(R B ), calculate R B The corresponding brightening factor L = 1.228, then the improved IAWB-GWPR-CLAHE algorithm f1(L) = 1, f2(L) = 0, that is, IGC(I) = IG(I), indicating that the IAWB-GWPR algorithm needs to be used for blue light processing. Figure 8b This is the result after blue light processing. Figure 8c for Figure 8a The growth image corresponding to the blue light is not turned on. As can be seen from the figure, the difference between the two images is small, which shows that the blue light interference treatment method proposed in the present invention has a better effect;

[0162] Using the top-hat adaptive binarization method Figure 8b Converted into a binary image, such as Figure 8d As shown;

[0163] Extract the contour with the largest area to obtain a background-free mycelium binary image, such as Fig. 9 As shown;

[0164] Count the frequency G of white pixels in the grayscale histogram of the binary image of hyphae without background t , further analyzed the growth characteristics of seafood mushrooms during the recovery period: the degree of mycelium density, mycelium distribution uniformity and mycelium coverage. The calculation results are as follows Fig.10a As shown: The degree of whiteness of seafood mushroom mycelium is: Fig. 9 The frequency of white pixel distribution G t , Gt =4640;500 <G t <5500, indicating mild rich white;

[0165] The uniformity of the mycelium distribution of seafood mushrooms is calculated as follows: Fig. 9 The eigenvalues ​​of the two-dimensional Gaussian matrix of all white pixel coordinates are λ1=2898.69, λ2=3426.04, and the uniformity coefficient H=1.18 is calculated, indicating that the distribution tends to be uniform. The visual scatter plot of the fitting result is as follows: Fig.10b As shown;

[0166] Mycelium coverage of seafood mushroom W F To: Calculate the background-free hyphae binary image Fig. 9 The number of white pixels N in V = 13442, the total number of pixels in the bottle mouth area N P =39189, then the coverage W F =34.3%, 20% <W F <50% indicates less coverage.

[0167] In another specific embodiment, when the growth image of seafood mushrooms in the recovery period is heavily exposed to blue light, the specific steps of intelligently identifying its growth characteristics are:

[0168] Take and obtain on-site growth images of seafood mushroom mycelium during the recovery period, such as Fig.11a As shown;

[0169] The mycelium growth image of the seafood mushroom in the recovery period is converted from an RGB image to a grayscale image, and then the grayscale image is processed by the Laplace operator, and the Laplace variance of the image is calculated to be Q=346.48>200, and it is determined that there are no small water droplets in the image and the image is clear, and further processing is performed based on the image later;

[0170] To deal with blue light interference, Fig.11a The RGB image is converted to an HSV image, and the number of blue pixels in the image is N B =58687, blue light intensity R B =99.39%,99% <R B <100%, then Fig.11b The blue light level is judged as heavy blue light;

[0171] Through the blue light brightening factor function L(R B ), calculate R B The corresponding brightening factor L = 1.247, then the improved IAWB-GWPR-CLAHE algorithm f1(L) = 0, f2(L) = 1, that is, IGC(I) = CLAHE(I), indicating that the CLAHE algorithm needs to be used for blue light processing. Fig.11bThis is the result after blue light processing. Fig.11c for Fig.11a The growth image without blue light, comparison Fig.11b As can be seen from Fig. 2 and Fig. 3 , the difference between the two figures where there are hyphae is small, which shows that the blue light interference treatment method proposed in the present invention has a better effect;

[0172] Using the top-hat adaptive binarization method Fig.11b Converted into a binary image, such as Fig.11d As shown;

[0173] Extract the contour with the largest area to obtain a background-free mycelium binary image, such as Fig.12 As shown;

[0174] Count the frequency G of white pixels in the grayscale histogram of the binary image of hyphae without background t , further analyzed the growth characteristics of seafood mushrooms during the recovery period: the degree of mycelium density, mycelium distribution uniformity and mycelium coverage. The calculation results are as follows Fig.13a As shown: The degree of whiteness of seafood mushroom mycelium is: Fig.12 The frequency of white pixel distribution G t , G t =13897;5500 <G t <15000 means it is medium-rich white.

[0175] The uniformity of the mycelium distribution of seafood mushrooms is calculated as follows: Fig.12 The eigenvalues ​​of the two-dimensional Gaussian matrix of all white pixel coordinates are λ1=2377.00, λ2=3104.57, and the uniformity coefficient H=1.31 is calculated, indicating that the distribution is uneven. The visual scatter plot of the fitting result is as follows: Fig.13b As shown;

[0176] Mycelium coverage of seafood mushroom W F To: Calculate the background-free hyphae binary image Fig.12 The number of white pixels N in V =184525, total number of pixels in the bottle mouth area N P =636078, then the coverage W F =29.01%, 20% <W F <50% indicates less coverage.

[0177] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.

[0178] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An intelligent identification method for mycelial growth characteristics of edible fungi in the recovery period of factory-scale bottle cultivation, characterized in that: include: S1, photographing and obtaining the on-site growth RGB image of mycelium at time t during the recovery period of bottle-grown edible fungi; S2, performing water drop interference and blur detection and processing on the acquired RGB image to obtain a clear RGB image; S3, performing blue light interference processing on the clear RGB image; S4, converting the image processed by blue light interference into a binary image with enhanced foreground objects by using a top-hat adaptive binarization method; S5, searching for all contours in the binary image of the enhanced foreground object, selecting the contour with the largest area, and calculating the circumscribed rectangle and the center point of the rectangle based on the contour with the largest area; S6, draw a circle with the longest distance from the center point of the rectangle to the rectangular border as the radius, retain the white pixels in the circular area, and replace all the pixels outside the circular area with black pixels to obtain a background-free hyphae binary image; S7, generating a grayscale histogram of the background-free mycelium binary image, and reading the distribution frequency number G of white pixels with a pixel value of 255 in the histogram t , according to G t The mycelial growth characteristics during the recovery period were calculated.

2. The method for intelligently identifying the growth characteristics of mycelium in the recovery period of edible fungi produced in bottle cultivation in factories according to claim 1 is characterized in that: In S2, water drop interference and blur detection are specifically as follows: the RGB image of the mycelium on-site growth in the recovery period is converted into a grayscale image, the grayscale image is processed using the Laplace operator, and the Laplace variance Q of the image is calculated; If Q max When it is determined that there are small water droplets in the image, an alarm is prompted and the lens is scraped, and then the image is re-taken to obtain the RGB image of the mycelium growth in the recovery period;​ If Q>Q max , determine that the image is clear.

3. The method for intelligently identifying the growth characteristics of mycelium in the recovery period of edible fungi produced in bottle cultivation in factories according to claim 1 is characterized in that: In S3, the improved IAWB-GWPR-CLAHE algorithm is used to process the blue light interference of the clear RGB image, specifically: S301, converting a clear RGB image from an RGB color space to an HSV color space to obtain an HSV image; S302, calculate blue light intensity R B : Among them, R B Indicates the proportion of blue pixels in the HSV image, N B Indicates the number of blue pixels in the HSV image, N Total Indicates the total number of pixels in the HSV image; S303, by brightening factor function L(R B ) Calculate the brightening factor L; S304, using the improved IAWB-GWPR-CLAHE algorithm to process the image and calculate IGC(I): IGC(I)=f1(L)×IG(I)+f2(L)×CLAHE(I) Where I represents the input image, IG(I) and CLAHE(I) represent the IAWB-GWPR algorithm and CLAHE algorithm respectively, 0<R B <R Bmin , R Bmin <R B <R Bmid , R Bmid <R B <R Bmax , R Bmax <R B <1 indicates the level of blue light intensity: no blue light, mild blue light, moderate blue light and severe blue light; If IGC(I) = 0, no blue light processing is required; If IGC(I)=IG(I), use the improved automatic white balance algorithm IAWB-GWPR to process blue light; If IGC(I)=CLAHE(I), use contrast-limited adaptive histogram equalization algorithm CLAHE for blue light processing; If IGC(I)=IG(I)+CLAHE(I), the improved automatic white balance algorithm IAWB-GWPR and the contrast-limited adaptive histogram equalization algorithm CLAHE are used to process blue light.

4. The method for intelligently identifying the growth characteristics of mycelium in the recovery period of edible fungi produced by factory-scale bottle cultivation according to claim 3 is characterized in that: L(R B ) function is obtained by adjusting different blue light intensities R B , adjust the brightening factor to obtain R B The corresponding ideal brightening factor L is collected through experiments. B ,L) sample data, the blue light brightening factor L is fitted by polynomial as the blue light intensity R B of the changing relationship.

5. The method for intelligently identifying the growth characteristics of mycelium in the recovery period of edible fungi produced by factory-scale bottle cultivation according to claim 1 is characterized in that: In S4, the top-hat adaptive binarization method is specifically: S401, performing a top-hat transformation on the image after the blue light interference processing to remove the uneven illumination, and obtaining I top ; S402, performing Gaussian smoothing with a 15×15 Gaussian kernel; S403, calculating the local adaptive binarization image threshold T of the image P(x, y) after top-hat smoothing local (x,y), according to T local (x, y) to achieve accurate binarization processing, and convert T local (x, y) is compared with the gray value T(x, y) of the pixel in the binary image to obtain the binary image P'(x, y) of the enhanced foreground object; Among them, P'(x,y) represents the final output top-hat adaptive enhanced foreground object binary image, T(x,y) represents the gray value corresponding to each pixel in the binary image, B(x,y) represents the offset for adjusting the image brightness or contrast, T local represents the local adaptive binary image threshold, P(x,y) represents the image after Gaussian smoothing, I top (x, y) represents the image after top-hat transformation, I(x, y) represents the original image, O(I(x, y)) represents the image after opening the original image, K represents the normalization factor, k represents the radius of the Gaussian kernel, G 15×15 (i, j) represents a Gaussian kernel of size 15×15, and w represents a rectangular local window of size w×w.

6. The method for intelligently identifying the growth characteristics of mycelium in the recovery period of edible fungi produced by factory-scale bottle cultivation according to claim 1 is characterized in that: In S7, the mycelium growth characteristics during the recovery period include: mycelium density, mycelium distribution uniformity, mycelium coverage, and mycelium growth rate.

7. The method for intelligently identifying the growth characteristics of mycelium in the recovery period of edible fungi produced by factory-scale bottle cultivation according to claim 6 is characterized in that: The degree of mycelium whiteness is expressed in G t It indicates the degree of recovery of mycelium during the recovery period.

8. The method for intelligently identifying the growth characteristics of mycelium in the recovery period of edible fungi produced in bottle cultivation in factories according to claim 6 is characterized in that: The method for calculating the uniformity of mycelium distribution is as follows: extract the coordinates of each white pixel representing mycelium in the binary image of mycelium without background, calculate the eigenvalues ​​λ1 and λ2 of the two-dimensional Gaussian matrix of all coordinate points, and calculate the uniformity coefficient H: When the uniformity coefficient H = 1: it means that the eigenvalues ​​are completely equal and the distribution is completely uniform; When the uniformity coefficient H is close to 1: it means that the distribution tends to be uniform; When the uniformity coefficient H is greater than 1: it indicates uneven distribution.

9. The method for intelligently identifying the growth characteristics of mycelium in the recovery period of edible fungi produced in factory-scale bottle cultivation according to claim 6 is characterized in that: Mycelium coverage W F The calculation method is: the number of white pixels N in the binary image of hyphae without background V and the total number of pixels N in the area P The ratio is calculated as:

10. The method for intelligently identifying the growth characteristics of mycelium in the recovery period of edible fungi produced in bottle cultivation in factories according to claim 6, characterized in that: The calculation method of mycelium growth rate is: count the distribution frequency of white pixels in the time (t2-t1) G t The rate of change is calculated as: Among them, V GSV represents the mycelial growth rate, G t2 , G t1 They represent the distribution frequency of white pixels at time t2 and time t1 respectively, and t2 and t1 represent the time points when automatic photography is taken.

Images