CN118038449B - Breeding method and application of edible fungus strain - Google Patents

Abstract

The invention claims a breeding method and application of edible fungus strains, which comprises the steps of dividing a colony growth image of edible fungus to be selected into grids, outputting a plurality of colony outline pixels of the edible fungus to form a colony outline pixel grid set of the edible fungus; training an edible fungus colony integrity recognition model, detecting metadata of a plurality of edible fungus colony contour pixels, and outputting attribute metadata sets of grids; matching degree comparison is carried out on the attribute metadata sets of the target edible fungus colonies between two grids in the outline pixels of the plurality of edible fungus colonies by using a comparison algorithm, and a first colony shape of the target edible fungus colonies is output; and carrying out superposition treatment and interruption monitoring treatment on the first colony shape, and outputting a second colony shape of the target edible fungus colony. According to the invention, the loss rate caused by errors and shielding overlapping is reduced, the edible fungus colony is predicted, interrupted and monitored by utilizing color calculation and historical appearance information, the robustness of the appearance graph is improved, and the dependence on hardware is reduced.

Description

A method for breeding edible fungi and its application

Technical Field

The invention belongs to the technical field of computer-aided edible fungus colony analysis, and in particular, relates to a breeding method and application of edible fungus strains.

Background technique

There are many types of edible fungi. Commonly used edible fungi include shiitake mushrooms, straw mushrooms, button mushrooms, black fungus, white fungus, hericium, bamboo fungus, etc. Edible fungi have various shapes. The parts of the fruiting body of umbrella-shaped edible fungi are called cap, gills, stipe and ring from top to bottom. The cap is the most important edible part of the umbrella-shaped edible fungi, and the integrity of the cap reflects the degree of opening and freshness of the umbrella-shaped edible fungi. The integrity of the cap is the integrity of the edible fungi. Therefore, the integrity of edible fungi is the most important indicator for evaluating the quality of edible fungi.

Different degrees of integrity of edible fungi correspond to different edible fungi grading, but the prior art cannot accurately and quickly calculate the integrity of edible fungi, mainly because the inner complete edge of the edible fungi can only identify a partial complete outline edge before the umbrella is fully opened, but cannot identify a complete complete outline edge, and the integrity of the edible fungi cannot be accurately calculated based on the partial complete outline edge, which makes it difficult to accurately grade the edible fungi, and it is impossible to accurately grade the edible fungi according to their integrity.

Summary of the invention

In view of the above-mentioned defects, the purpose of the present invention is to provide a method and device for breeding edible fungi species, aiming to propose a data engine to reduce the cost of data set production, improve comparison accuracy, reduce the impact of occlusion and overlap on target monitoring, reduce errors caused by missed detection and false detection, and reduce the algorithm's dependence on hardware.

According to the first aspect of the present invention, the present invention claims protection for a method for breeding edible fungi, characterized in that it comprises:

Divide the growth image of the edible fungus colony to be selected into a grid, output a plurality of edible fungus colony outline pixels, and form an edible fungus colony outline pixel grid set;

Training an edible mushroom colony integrity recognition model, performing metadata detection on the plurality of edible mushroom colony outline pixels, and outputting an attribute metadata data set of each grid;

Using a comparison algorithm, a matching comparison is performed on the attribute metadata data set of the target edible fungus colony between two grids in the plurality of edible fungus colony outline pixels, and a first colony shape of the target edible fungus colony is output;

Performing superposition processing and interruption monitoring processing on the first colony shape, and outputting a second colony shape of the target edible fungus colony;

The target edible fungus colony with the second colony appearance that meets the preset conditions is used as the edible fungus for breeding.

Furthermore, the first bacterial colony shape is subjected to superposition processing, and further includes:

Obtain the local shape where the inner contour of the shape appears in the middle of the area other than the first grid from the first colony shape, and output the local shape grid set ;

Detection , if the appearance/> In scenes where there is an incomplete mesh, the new shape will be output at the incomplete mesh and added to the mesh set/> ;

exist The neighborhood of the inner contour pixel is used to query the outer shape of the interrupted part of the grid and the outer shape of the terminated grid set. , for the appearance of the interrupted grid in the grid set, compare the interruption position with /> Whether the inner outline pixels are continuous and the number of interrupted grids and/> Whether the characteristic values are the same, if they are continuous and the same, breeding is carried out in/> In, otherwise clear;

For the shapes in the grid set, the ideal superimposed shape is output according to the strategy of the most similar growth curve posture and the closest inner and outer distance. Superimposed on the ideal superimposed shape, the output new shape is recorded as the superimposed shape/> ;

Detection The outline of the shape. If the outline is close to the edge or the grid where the outline is located is the last grid of the growth image, then / > Shape from mesh set/> in the clear, otherwise in /> In the next grid of the grid where the outer contour pixel is located, check whether there is an outer contour pixel at the adjacent position of the outer contour pixel. If there is an outer contour, output the ideal superimposed outer contour according to the strategy of the most similar growth curve posture and the closest inner and outer distance. The local shape is superimposed on the ideal superimposed shape, and the superimposed shape is recorded as the superimposed shape/> ;

Loop until Cleared when empty or the grid set length no longer changes /> The appearance of the .

Furthermore, before training the edible fungus colony integrity recognition model, the method further includes:

Model preprocessing, including:

Analyze and process the surface grid of the edible fungus colony integrity recognition model to extract the required edible fungus colony information;

Calculate the growth importance of the grid between the colony nodes and the attached colony nodes as well as the geometric characteristic values of the colony;

The processed edible fungus colony information is stored in a colony tree structure, and redundant information irrelevant to colony capture is eliminated.

Furthermore, the training of the edible fungus colony integrity recognition model further includes:

Acquire a sample edible mushroom colony image including a circular object segmentation result as training data;

Extracting the circumscribed rectangle of the circular object in the sample edible fungus colony image;

De-noising the bounding rectangle extracted in the bounding rectangle extraction step, and outputting the de-noised bounding rectangle;

A loss function is set to use a pixel matrix containing the denoised circumscribed rectangle as an importance matrix;

The training data is learned using the loss function set in the loss function setting step, thereby outputting a segmentation model for segmenting the circular object in the edible mushroom colony image.

Furthermore, the method of using a comparison algorithm to compare the matching degree of the attribute metadata data set of the target edible fungus colony between two grids in the plurality of edible fungus colony outline pixels and outputting a first colony shape of the target edible fungus colony also includes:

The comparison algorithm uses a multi-strategy comparison method, including the most similar independent metadata strategy and the average growth curve self-stabilization strategy;

The independent metadata most similar strategy is that the edible fungus colony grows from the k-1th grid to the kth grid through the edible fungus colony recognition model output with the highest feature matching degree;

The average growth curve self-stabilization strategy refers to the highest matching degree between the average growth curve length of the displacement of the edible fungus colony growing from the k-1 grid to the k grid and the posture characteristics;

The importance of the independent metadata most similar strategy and the average growth curve self-stabilizing strategy is different, wherein the importance of the independent metadata most similar strategy is greater than the importance of the average growth curve self-stabilizing strategy.

Furthermore, the method further comprises:

Calculate the edible mushroom colony in the k-1th grid With the kth grid/> Characteristic matching degree of edible fungus colonies/> , output the corresponding contrast/> , output the corresponding contrast according to the growth curve self-stabilization strategy/> , sum up and output the final contrast according to the preset importance/> ;

Combined with the grid missing comparison scenario when there is breeding monitoring, the ratio of all edible fungi colonies in the k-1th grid to the edible fungi colonies in the kth grid is calculated. Output a contrast matrix; use the contrast algorithm to compare the edible fungus colonies in the k-1th grid and the kth grid; perform a cyclic comparison until all adjacent grids are compared, and output a first colony shape of the target edible fungus colony.

Furthermore, the characteristic matching degree of the edible fungus colony is calculated, and the distance Calculated as follows:

in, 、/> is the edible mushroom colony in the k grid/> Positioning, /> 、/> is the location of the edible mushroom colony m in the k-1 grid;

The feature matching formula is: ;

in is the edible fungus colony in the k-1th grid/> A vector description of the feature, is the edible fungus colony in the i-th grid/> Features, correct the results and output /> ;

The average growth curve self-stabilization strategy is used to determine the edible fungus colony in the k-1 grid The kth grid appears at the point (/> ）distance/> Consider at the same time/> The attitude direction vector value/> No abnormal changes will occur in the next grid. Use the formula: /> Where/> is the hyperbolic cosine function,/> is the correlation coefficient, /> is the growth curve length of the edible fungus colony l in the k-1th grid, is the angle between the kth grid edible fungus colony l and the k-1th grid edible fungus colony k, /> is the growth curve attitude direction vector value of the edible fungus colony l in the k-1th grid;

According to importance Add and output total contrast /> ;

Further, interrupting monitoring of the first bacterial colony shape also includes:

Get the shape with interruption in the middle of the superimposed shape. If the kth grid is interrupted, calculate the edible fungus colony Color information, output color RGB mean /> and color boundary length/> And the growth curve attitude direction vector value of the edible fungus colony/> and growth curve length/> ;

According to the preset importance, color and growth curve, the pixel evaluation posture score is output as follows: , the pixel evaluation length is ;

and/> is the weight coefficient;

The evaluation pixels are added as the outline to the superimposed outline, and the cycle is repeated until the entire outline graphic is complete, and the second colony outline of the target edible fungus colony is output.

Further, the target edible fungus colony with the second colony shape that meets the preset conditions is used as the edible fungus for breeding, and further includes:

Detecting the composition quality of the target edible fungus colony of the second colony shape, wherein the composition quality includes at least one of a microbial content, a pollutant content, and a harmful chemical content;

Based on the component qualities of the edible fungi to be selected, the edible fungi to be bred are determined.

Furthermore, according to the second aspect of the present invention, the present invention seeks to protect the application of the method described in the breeding of edible fungi strains.

The scheme of the present invention trains an edible fungus colony target recognition model, uses the model to automatically detect the edible fungus colony targets in each grid in the image sequence, and saves the feature description vector of each edible fungus colony target for comparison. During the comparison, multi-strategy comparison is used to improve the registration rate, and the loss rate due to errors and occlusion overlap is reduced by using a breeding monitoring grid missing comparison method. The shape graphics are optimized by using overlay processing, and the edible fungus colony is predicted and interrupted for monitoring by using color calculation and historical shape information, thereby improving the robustness of the shape graphics and reducing the dependence on hardware.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a flowchart of a method for selecting and breeding edible fungi according to an embodiment of the present invention.

FIG. 2 is a second workflow diagram of a method for breeding edible fungi species according to an embodiment of the present invention.

FIG. 3 is a third workflow diagram of a method for selecting and breeding edible fungi according to an embodiment of the present invention.

Detailed ways

The following will be combined with the accompanying drawings to clearly and thoroughly describe the examples in 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.

According to the first embodiment of the present invention, referring to FIG1 , the present invention claims protection for a method for breeding edible fungi, comprising:

Divide the growth image of the edible fungus colony to be selected into a grid, output a plurality of edible fungus colony outline pixels, and form an edible fungus colony outline pixel grid set;

Training an edible mushroom colony integrity recognition model, performing metadata detection on the plurality of edible mushroom colony outline pixels, and outputting an attribute metadata data set of each grid;

Using a comparison algorithm, a matching comparison is performed on the attribute metadata data set of the target edible fungus colony between two grids in the plurality of edible fungus colony outline pixels, and a first colony shape of the target edible fungus colony is output;

Performing superposition processing and interruption monitoring processing on the first colony shape, and outputting a second colony shape of the target edible fungus colony;

The target edible fungus colony with the second colony appearance that meets the preset conditions is used as the edible fungus for breeding.

Furthermore, the first bacterial colony shape is subjected to superposition processing, and further includes:

Obtain the local shape where the inner contour of the shape appears in the middle of the area other than the first grid from the first colony shape, and output the local shape grid set ;

Detection , if the appearance/> In scenes where there is an incomplete mesh, the new shape will be output at the incomplete mesh and added to the mesh set/> ;

exist The neighborhood of the inner contour pixel is used to query the outer shape of the interrupted part of the grid and the outer shape of the terminated grid set. , for the appearance of the interrupted grid in the grid set, compare the interruption position with /> Whether the inner outline pixels are continuous and the number of interrupted grids and/> Whether the characteristic values are the same, if they are continuous and the same, breeding is carried out in/> In, otherwise clear;

For the shapes in the grid set, the ideal superimposed shape is output according to the strategy of the most similar growth curve posture and the closest inner and outer distance. Superimposed on the ideal superimposed shape, the output new shape is recorded as the superimposed shape/> ;

Detection The outline of the shape. If the outline is close to the edge or the grid where the outline is located is the last grid of the growth image, then / > Shape from mesh set/> in the clear, otherwise in /> In the next grid of the grid where the outer contour pixel is located, check whether there is an outer contour pixel at the adjacent position of the outer contour pixel. If there is an outer contour, output the ideal superimposed outer contour according to the strategy of the most similar growth curve posture and the closest inner and outer distance. The local shape is superimposed on the ideal superimposed shape, and the superimposed shape is recorded as the superimposed shape/> ;

Loop until Cleared when empty or the grid set length no longer changes /> The appearance of the .

Furthermore, before training the edible fungus colony integrity recognition model, the method further includes:

Referring to Figure 2, model preprocessing includes:

Analyze and process the surface grid of the edible fungus colony integrity recognition model to extract the required edible fungus colony information;

Calculate the growth importance of the grid between the colony nodes and the attached colony nodes as well as the geometric characteristic values of the colony;

The processed edible fungus colony information is stored in a colony tree structure, and redundant information irrelevant to colony capture is eliminated.

Further, referring to FIG. 3 , the training edible fungus colony integrity recognition model further includes:

Acquire a sample edible mushroom colony image including a circular object segmentation result as training data;

Extracting the circumscribed rectangle of the circular object in the sample edible fungus colony image;

De-noising the bounding rectangle extracted in the bounding rectangle extraction step, and outputting the de-noised bounding rectangle;

A loss function is set to use a pixel matrix containing the denoised circumscribed rectangle as an importance matrix;

The training data is learned using the loss function set in the loss function setting step, thereby outputting a segmentation model for segmenting the circular object in the edible mushroom colony image.

Furthermore, the method of using a comparison algorithm to compare the matching degree of the attribute metadata data set of the target edible fungus colony between two grids in the plurality of edible fungus colony outline pixels and outputting a first colony shape of the target edible fungus colony also includes:

The comparison algorithm uses a multi-strategy comparison method, including the most similar independent metadata strategy and the average growth curve self-stabilization strategy;

The independent metadata most similar strategy is that the edible fungus colony grows from the k-1th grid to the kth grid through the edible fungus colony recognition model output with the highest feature matching degree;

The average growth curve self-stabilization strategy refers to the highest matching degree between the average growth curve length of the displacement of the edible fungus colony growing from the k-1 grid to the k grid and the posture characteristics;

The importance of the independent metadata most similar strategy and the average growth curve self-stabilizing strategy is different, wherein the importance of the independent metadata most similar strategy is greater than the importance of the average growth curve self-stabilizing strategy.

Furthermore, the method further comprises:

Calculate the edible mushroom colony in the k-1th grid With the kth grid/> Characteristic matching degree of edible fungus colonies/> , output the corresponding contrast/> , output the corresponding contrast according to the growth curve self-stabilization strategy/> , sum up and output the final contrast according to the preset importance/> ;

Combined with the grid missing comparison scenario when there is breeding monitoring, the ratio of all edible fungi colonies in the k-1th grid to the edible fungi colonies in the kth grid is calculated. Output a contrast matrix; use the contrast algorithm to compare the edible fungus colonies in the k-1th grid and the kth grid; perform a cyclic comparison until all adjacent grids are compared, and output a first colony shape of the target edible fungus colony.

Furthermore, the characteristic matching degree of the edible fungus colony is calculated, and the distance Calculated as follows: ;

in, 、/> is the edible mushroom colony in the k grid/> Positioning, /> 、/> is the location of the edible mushroom colony m in the k-1 grid;

The feature matching formula is: ;

in is the edible fungus colony in the k-1th grid/> A vector description of the feature, is the edible fungus colony in the i-th grid/> Features, correct the results and output /> ;

The average growth curve self-stabilization strategy is used to determine the edible fungus colony in the k-1 grid The kth grid appears at the point (/> ) is at a distance of/> Consider at the same time/> The attitude direction vector value/> No abnormal changes will occur in the next grid. Use the formula: /> , where/> is the hyperbolic cosine function,/> is the correlation coefficient, /> is the growth curve length of the edible fungus colony l in the k-1th grid, is the angle between the kth grid edible fungus colony l and the k-1th grid edible fungus colony k, /> is the growth curve attitude direction vector value of the edible fungus colony l in the k-1th grid;

According to importance Add and output total contrast /> .

Further, interrupting monitoring of the first bacterial colony shape also includes:

Get the shape with interruption in the middle of the superimposed shape. If the kth grid is interrupted, calculate the edible fungus colony Color information, output color RGB mean /> and color boundary length/> And the growth curve attitude direction vector value of the edible fungus colony/> and growth curve length/> ;

According to the preset importance, color and growth curve, the pixel evaluation posture score is output as follows: , the pixel evaluation length is ;

and/> is the weight coefficient;

The evaluation pixels are added as the outline to the superimposed outline, and the cycle is repeated until the entire outline graphic is complete, and the second colony outline of the target edible fungus colony is output.

Furthermore, taking the target edible fungi colony with the second colony appearance that meets the preset conditions as the edible fungi for breeding also includes: detecting the composition quality of the target edible fungi colony with the second colony appearance, the composition quality including at least one of the microbial content, pollutant content and harmful chemical content; and determining the edible fungi for breeding based on the composition quality of the edible fungi to be selected.

In this embodiment, when the comparison algorithm is used to compare the edible fungus colonies in the k-1th grid and the kth grid, it includes:

According to the contrast matrix, each edible fungus colony is compared with an ideal contrast object. Its ideal comparison object/> The contrast is compared with the threshold value Q. When it is greater than the threshold value Q, / > Add to /> In the monitored appearance, if the value is less than the threshold value Q, it cannot be compared.

If there is a scene that cannot be compared:

Scenario 1: For a colony of edible mushrooms in a k-1 grid There is no edible mushroom colony in the kth grid for comparison. At this time, the edible mushroom colony in the actual edible mushroom colony scene is out of the area and cannot be compared. At this time, it is detected whether the position of the edible mushroom colony appears at the edge of the area. If so, the edible mushroom colony is no longer monitored.

Scenario 2: For the extra edible fungus colonies in the k grid that have not been compared, set them as new monitoring objects. If there is no edible mushroom colony in the grid for comparison, the object is not being monitored.

The steps are repeated until the comparison step is completed for all edible fungi colonies in the k-1 grid.

Assume that there are F grids in the growth image, and each grid is identified and then the grid set of edible fungi colonies is output. ; />

in, , monitoring is to / > With/> ,/> With/> ,……,/> With/> The process of comparing edible fungi colonies in the The number of edible fungi colonies detected in each grid may be different.

Assume that the breeding monitoring grid set is

To carry out With/> When comparing, all the edible fungus colony grid sets of the comparison object are not output. Where c is the current grid set to be compared. With/> The comparison, /> For the a-th edible mushroom colony in the x-th grid, no comparison object has been output so far, and/> ,/> When conducting/> With/> When comparing the All edible fungi colonies in/> For comparison, due to the loss of contrast in the grid, it is necessary to consider how many grids the edible fungus colony actually grows in the comparison, and its contrast /> Should be: ;

For example: Assume No with/> If the comparison of edible fungi colonies in any grid set is successful, it will be added to the breeding monitoring grid set/> In/> With/> When comparing, the same With/> Compare; therefore, when calculating the matching degree during the comparison/> Need to consider Actually it grows 6 grid intervals, so .

In this embodiment, the total contrast matrix needs to be calculated first. The distance strategy is not explicitly used in the present invention. This strategy is included in the growth curve self-stabilization strategy. When the average growth curve is very small, it is equivalent to the minimum distance strategy. This can avoid the interference of the distance strategy on the contrast when the edible fungus colonies intersect in the area.

Take the 10th and 11th grids as examples. For the 9th grid with /> The associated edible fungus colonies, at this time, are compared/> With/> , their positions are respectively/> , the attribute metadata is/> , calculate the distance , through/> Correct its value to between 0 and 1, usually/> Take 0.2 and calculate its matching degree/> By correcting /> , usually/> Take 0.1, and calculate the displacement of the edible fungus colony from the 9th grid to the 10th grid according to the above distance formula as its growth curve/> , calculate its growth curve posture/> , calculate/> With/> Posture , calculate/> , calculate the total contrast/> , continue the cycle calculation Until the total contrast between all monitored edible fungi colonies and the edible fungi colonies in the 11th grid is calculated, the contrast matrix is output. This breeding monitoring and fixed edible fungi colonies in situ grid missing contrast method can reduce the error accumulation problem caused by the prediction algorithm and improve the registration rate.

After calculating the contrast matrix, the most relevant edible fungus colony in the 11th grid is obtained through the comparison process. It is obtained from the edible fungus colony/> The most contrasting object will be Compared with the threshold value Q, if it is less than this value, it means /> No/> Edible mushroom colonies, which also means There is no contrasting edible fungus colony in the 11th grid, and the threshold value Q is usually 0.6. It falls in the first quadrant with the upper left corner of the image as the origin. So the right and bottom sides of the image are the edges of the area where the edible fungus colony may swim out. Calculate/> and/> And in/> When the comparison is less than the growth curve, it means that it is at the edge, and the edible fungus colony will no longer be monitored. Otherwise, it will be fixed at the position of the 10th grid and selected for monitoring. All objects to be monitored are detected in a loop. In this comparison algorithm, the overall situation can be taken into account, and the two most similar ones are output in the entire comparison process. There will be no preference for the first detected edible fungus colony in the non-maximum suppression method, making it easier for the first detected edible fungus colony to obtain the comparison object. And considering whether the edible fungus colony is close to the edge of the area, the objects of selective monitoring are reduced, which reduces the complexity of calculation.

Furthermore, the first bacterial colony shape is subjected to superposition processing, and further includes:

According to the output shape graphics of edible fungi colonies, it can be used for edible fungi colony analysis. Firstly, the edible fungi colonies output by the edible fungi colony recognition model can output the average number of edible fungi colonies in the area within the growth image time to assist detection personnel in counting. Secondly, the growth parameters of the edible fungi colonies can be calculated according to the shape graphics, and the growth category of the edible fungi colony and the growth curve of the edible fungi colony can be detected.

Further, according to the second embodiment of the present invention, the present invention seeks to protect the application of the method described in the breeding of edible fungi strains.

Those skilled in the art will appreciate that the contents disclosed in this disclosure may be subject to various variations and improvements. For example, the various devices or components described above may be implemented by hardware, or by software, firmware, or a combination of some or all of the three.

Flowcharts are used in this disclosure to illustrate the steps of the method according to the embodiments of the present disclosure. It should be understood that the preceding or following steps are not necessarily performed precisely in order. On the contrary, various steps may be processed in reverse order or simultaneously. At the same time, other processing may also be added to these processes.

Those skilled in the art will appreciate that all or part of the steps in the above method may be completed by instructing the relevant hardware through a computer program, and the program may be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk. Optionally, all or part of the steps in the above embodiment may also be implemented using one or more integrated circuits. Accordingly, each module/unit in the above embodiment may be implemented in the form of hardware or in the form of a software functional module. The present disclosure is not limited to any particular form of combination of hardware and software.

Unless otherwise defined, all terms used herein have the same meaning as commonly understood by those of ordinary skill in the art to which the present disclosure belongs. It should also be understood that terms such as those defined in common dictionaries should be interpreted as having a meaning consistent with their meaning in the context of the relevant technology, and should not be interpreted in an idealized or highly formal sense, unless explicitly defined as such herein.

The above is an explanation of the present disclosure and should not be considered as a limitation thereof. Although several exemplary embodiments of the present disclosure are described, it will be readily understood by those skilled in the art that many modifications may be made to the exemplary embodiments without departing from the novel teachings and advantages of the present disclosure. Therefore, all such modifications are intended to be included within the scope of the present disclosure as defined in the claims. It should be understood that the above is an explanation of the present disclosure and should not be considered to be limited to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The present disclosure is defined by the claims and their equivalents.

In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "illustrative embodiments", "examples", "specific examples", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

Although embodiments of the present invention have been shown and described, those skilled in the art will appreciate that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims (8)

1. A method for breeding edible fungi, comprising: Divide the growth image of the edible fungus colony to be selected into a grid, output a plurality of edible fungus colony outline pixels, and form an edible fungus colony outline pixel grid set; Training an edible mushroom colony integrity recognition model, performing metadata detection on the plurality of edible mushroom colony outline pixels, and outputting an attribute metadata data set of each grid; Using a comparison algorithm, a matching comparison is performed on the attribute metadata data set of the target edible fungus colony between two grids in the plurality of edible fungus colony outline pixels, and a first colony shape of the target edible fungus colony is output; Performing superposition processing and interruption monitoring processing on the first colony shape, and outputting a second colony shape of the target edible fungus colony; The target edible fungus colony with the second colony shape that meets the preset conditions is used as the edible fungus for breeding; The first bacterial colony shape is subjected to superposition processing, further comprising: Obtain the local shape where the inner contour of the shape appears in the middle of the area other than the first grid from the first colony shape, and output the local shape grid set ; Detection , if the appearance/> In scenes where there is an incomplete mesh, the new shape will be output at the incomplete mesh and added to the mesh set/> ; exist The neighborhood of the inner contour pixel is used to query the shape of the interrupted part of the grid and the shape of the terminated shape to output the grid set , for the appearance of the interrupted grid in the grid set, compare the interruption position with /> Whether the inner outline pixels are continuous and the number of interrupted grids and/> Whether the characteristic values are the same, if they are continuous and the same, breeding is carried out in/> Otherwise clear; For the shapes in the grid set, the ideal superimposed shape is output according to the strategy of the most similar growth curve posture and the closest inner and outer distance. Superimposed on the ideal superimposed shape, the output new shape is recorded as the superimposed shape/> ; Detection The outline of the shape. If the outline is close to the edge or the grid where the outline is located is the last grid of the growth image, then / > Shape from mesh set/> in the clear, otherwise in /> In the next grid of the grid where the outer contour pixel is located, check whether there is an outer contour pixel at the adjacent position of the outer contour pixel. If there is an outer contour, output the ideal superimposed outer contour according to the strategy of the most similar growth curve posture and the closest inner and outer distance. The local shape is superimposed on the ideal superimposed shape, and the superimposed shape is recorded as the superimposed shape/> ; Loop until Cleared when empty or the grid set length no longer changes /> The appearance of The interruption monitoring process of the first bacterial colony shape also includes: Get the shape with interruption in the middle of the superimposed shape. If the kth grid is interrupted, calculate the edible fungus colony Color information, output color RGB mean /> and color boundary length/> And the growth curve attitude direction vector value of the edible fungus colony/> and growth curve length/> ; According to the preset importance, color and growth curve, the pixel evaluation posture score is output as follows: ; The pixel evaluation length is ; and/> is the weight coefficient; The evaluation pixels are added as the outline to the superimposed outline, and the cycle is repeated until the entire outline graphic is complete, and the second colony outline of the target edible fungus colony is output. 2. The method for selecting and breeding edible fungi according to claim 1, characterized in that before the training of the edible fungi colony integrity recognition model, it also includes: model preprocessing, including: analyzing and processing the surface grid of the edible fungi colony integrity recognition model to extract the required edible fungi colony information; Calculate the growth importance of the grid between the colony nodes and the attached colony nodes as well as the geometric characteristic values of the colony; The processed edible fungus colony information is stored in a colony tree structure, and redundant information irrelevant to colony capture is removed. 3. The method for selecting and breeding edible fungi according to claim 2, characterized in that the training edible fungi colony integrity recognition model further comprises: Acquire a sample edible mushroom colony image including a circular object segmentation result as training data; Extracting the circumscribed rectangle of the circular object in the sample edible fungus colony image; De-noising the bounding rectangle extracted in the bounding rectangle extraction step, and outputting the de-noised bounding rectangle; A loss function is set to use a pixel matrix containing the denoised circumscribed rectangle as an importance matrix; The training data is learned using the loss function set in the loss function setting step, thereby outputting a segmentation model for segmenting the circular object in the edible mushroom colony image. 4. The method for selecting and breeding edible fungi according to claim 3 is characterized in that the comparison algorithm is used to compare the matching degree of the attribute metadata data set of the target edible fungi colony between two grids in the plurality of edible fungi colony contour pixels, and output the first colony shape of the target edible fungi colony, and further comprises: the comparison algorithm is a multi-strategy comparison method, including the independent metadata most similarity strategy and the average growth curve self-stabilization strategy; The independent metadata most similar strategy is that the edible fungus colony grows from the k-1th grid to the kth grid through the edible fungus colony recognition model output with the highest feature matching degree; The average growth curve self-stabilization strategy refers to the highest matching degree between the average growth curve length of the displacement of the edible fungus colony growing from the k-1 grid to the k grid and the posture characteristics; The importance of the independent metadata most similar strategy and the average growth curve self-stabilizing strategy is different, wherein the importance of the independent metadata most similar strategy is greater than the importance of the average growth curve self-stabilizing strategy. 5. The method for breeding edible fungi according to claim 4, further comprising: Calculate the edible mushroom colony in the k-1th grid With the kth grid/> Characteristic matching degree of edible fungus colonies/> , output the corresponding contrast/> , output the corresponding contrast according to the growth curve self-stabilization strategy/> , sum up and output the final contrast according to the preset importance/> ; Combined with the grid missing comparison scenario when there is breeding monitoring, the ratio of all edible fungi colonies in the k-1th grid to the edible fungi colonies in the kth grid is calculated. Output contrast matrix; Using the comparison algorithm to compare the edible fungus colonies in the k-1th grid and the kth grid; The comparison is repeated until all adjacent grids have been compared, and the first colony shape of the target edible fungus colony is output. 6. The method for selecting and breeding edible fungi according to claim 5, characterized in that the step of calculating the characteristic matching degree of the edible fungi colony further comprises: distance Calculated as follows: ; in, 、/> are the edible fungus colonies in the k grids respectively/> Positioning, /> ,/> is the location of the edible mushroom colony m in the k-1 grid; The feature matching formula is: ; in is the edible fungus colony in the k-1th grid/> Vector description of features, /> is the edible fungus colony in the i-th grid/> Features, correct the results and output /> ; The average growth curve self-stabilization strategy is used to determine the edible fungus colony in the k-1 grid The kth grid appears at the point ( ）distance/> Consider at the same time/> The attitude direction vector value/> No abnormal changes will occur in the next grid. Use the formula: /> ; in is the hyperbolic cosine function,/> is the correlation coefficient, /> is the growth curve length of the edible fungus colony l in the k-1th grid, /> is the angle between the kth grid edible fungus colony l and the k-1th grid edible fungus colony k, is the growth curve attitude direction vector value of the edible fungus colony l in the k-1th grid; According to importance Add and output total contrast /> . 7. The method for selecting edible fungi according to claim 6, characterized in that the target edible fungi colonies with the second colony shape that meets the preset conditions are used as the selected edible fungi, and further comprising: Detecting the composition quality of the target edible fungus colony of the second colony shape, wherein the composition quality includes at least one of a microbial content, a pollutant content, and a harmful chemical content; Based on the component qualities of the edible fungi to be selected, the edible fungi to be selected are determined. 8. Use of the method according to any one of claims 1 to 7 in the breeding of edible fungi strains.