Grape grain longitudinal and transverse diameter identification method based on intelligent image algorithm

Abstract

The invention provides a grape grain longitudinal and transverse diameter recognition method based on an intelligent image algorithm, and belongs to the technical field of image recognition, and the method specifically comprises the steps: determining a screening intelligent algorithm in the intelligent algorithm through the historical accuracy, determining the recognition processing results of the longitudinal and transverse diameters of grape grains through different screening intelligent algorithms, and carrying out the recognition processing of the longitudinal and transverse diameters of the grape grains. And determining distribution data of identification deviation grape grains in different screening intelligent algorithms by utilizing a longitudinal and transverse diameter preset interval of an identification target and the identification processing result, and determining identification reliability coefficients and matching intelligent algorithms of the different screening intelligent algorithms in combination with deviation data of the different identification deviation grape grains. Based on the identification processing results of different matching intelligent algorithms and the identification reliability coefficient, the output of the identification result of the longitudinal and transverse diameters of the grape grains of the identification target is carried out, and the accuracy of the identification result of the longitudinal and transverse diameters of the grape grains is improved.

Description

Technical Field

[0001] The invention belongs to the technical field of image recognition, and in particular relates to a method for recognizing the vertical and horizontal diameters of grape grains based on an intelligent image algorithm. Background Art

[0002] In order to realize the recognition and processing of the longitudinal and transverse diameters of grapes, the invention patent application CN202410627615.0 "Efficient Grape Quality Assessment Method Based on Machine Vision" uses the suspected sub-edges of grapes in the main image of grapes to evaluate the fullness and bad point degree of grapes, and then obtains the quality coefficient of grapes to determine the quality assessment result of grapes. However, the following technical problems exist through analysis: In the process of identifying the longitudinal and transverse diameters of grape grains, the distribution of grape grains is often too concentrated, which makes the identification of the longitudinal and transverse diameters of grape grains too difficult.

[0003] In response to the above technical problems, the present application specifically provides a method for identifying the longitudinal and transverse diameters of grape grains based on an intelligent image algorithm. Summary of the invention

[0004] To achieve the purpose of the present invention, the present invention adopts the following technical solutions: A method for identifying the longitudinal and transverse diameters of grape grains based on an intelligent image algorithm, specifically comprising: S1 determines the grape variety of the identification target based on the distribution of the spacing between adjacent grapes of the identification target and when the identification accuracy of the longitudinal and transverse diameters of the grapes meets the requirements; S2 determines the historical accuracy of different intelligent algorithms in longitudinal and transverse diameter recognition processing based on the grape variety of the recognition target and the distribution of the spacing between adjacent grape grains, and determines the screening intelligent algorithm among the intelligent algorithms using the historical accuracy; S3 determines the recognition processing results of the longitudinal and transverse diameters of grape grains of different screening intelligent algorithms, determines the distribution data of the recognition deviation grape grains in different screening intelligent algorithms by using the longitudinal and transverse diameter preset intervals of the recognition target and the recognition processing results, and determines the recognition reliability coefficients of different screening intelligent algorithms and matching intelligent algorithms in combination with the deviation data of the different recognition deviation grape grains; S4 outputs the recognition results of the longitudinal and transverse diameters of the grape grains of the recognition target based on the recognition processing results of different matching intelligent algorithms and the recognition reliability coefficients.

[0005] The beneficial effects of the present invention are: The distribution data and deviation data of grape grains with recognition deviations in different screening intelligent algorithms are used to determine the matching intelligent algorithm in the screening intelligent algorithm, thereby achieving the determination of a matching intelligent algorithm with higher recognition processing reliability based on the deviation between the recognition processing results of the screening intelligent algorithm and the preset intervals of the longitudinal and transverse diameters, thereby avoiding the influence of the screening intelligent algorithm with low recognition reliability on the final recognition results of the longitudinal and transverse diameters of the grape grains, and improving the accuracy of the recognition results.

[0006] Based on the recognition processing results and recognition reliability coefficients of different matching intelligent algorithms, the recognition results of the longitudinal and transverse diameters of the grape grains of the recognition target are output, and the recognition processing results of multiple matching intelligent algorithms are used to output the recognition results of the longitudinal and transverse diameters of the grape grains of the recognition target. This avoids the technical problem of inaccurate recognition results of the longitudinal and transverse diameters caused by the use of a single matching intelligent algorithm, thereby ensuring the accuracy of the recognition processing results.

[0007] A further technical solution is that the distribution of the spacing between adjacent grape grains is determined based on the image recognition results of the grape grains, and specifically based on the analysis results of the image of the recognition target.

[0008] A further technical solution is that the method for determining the recognition accuracy of the longitudinal and transverse diameters of the grape grains is: Determine the distribution amount of the spacing between adjacent grape grains based on the spacing distribution between adjacent grape grains of the identified target; Determine the matching accuracy of different spacing distribution intervals based on the preset recognition accuracy corresponding to different spacing distribution intervals; Based on the proportion of the number of grape grains in different spacing distribution intervals, the weight coefficients of different spacing distribution intervals are determined, and combined with the matching accuracy of different spacing distribution intervals, the recognition accuracy of the longitudinal and transverse diameters of the grape grains is determined.

[0009] A further technical solution is that the recognition accuracy of the longitudinal and transverse diameters of the grape particles ranges from 0 to 1, and when the recognition accuracy of the longitudinal and transverse diameters of the grape particles is greater than a preset accuracy threshold, it is determined that the recognition accuracy of the longitudinal and transverse diameters of the grape particles meets the requirements.

[0010] A further technical solution is that the method for outputting the recognition result of the longitudinal and transverse diameters of the grape grains of the recognition target is: Determine the recognition results of different grape grains of the recognition target in terms of the longitudinal and transverse diameters of the matching intelligent algorithm based on the recognition processing results of the matching intelligent algorithm; Based on the deviation of the recognition results of the longitudinal and transverse diameters of different grapes in different matching intelligent algorithms, the recognition results are divided into different similarity intervals; Determine the sum of the recognition reliability coefficients in different similarity intervals by the sum of the recognition reliability coefficients of the matching intelligent algorithm corresponding to the recognition results in different similarity intervals, and take the similarity interval with the largest sum of the recognition reliability coefficient as the matching similarity interval; The average value of the recognition results of the longitudinal and transverse diameters of different matching intelligent algorithms in the matching similar interval is output as the recognition result of the longitudinal and transverse diameters of the grape grains.

[0011] A further technical solution is to divide the recognition results into different similarity intervals, specifically including: The recognition results of the deviation of the longitudinal and transverse diameters within the preset deviation range are divided into the same similarity range.

[0012] Other features and advantages will be described in the following description. The objects and other advantages of the present invention are realized and obtained by the structures particularly pointed out in the description and drawings.

[0013] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The above and other features and advantages of the present invention will become more apparent by describing in detail exemplary embodiments thereof with reference to the accompanying drawings; Figure 1 It is a flow chart of a method for identifying the vertical and horizontal diameters of grape grains based on an intelligent image algorithm; Figure 2 It is a flow chart of a method for determining the recognition accuracy of the longitudinal and transverse diameters of grape grains; Figure 3 It is a flow chart of a method for determining a screening intelligent algorithm in an intelligent algorithm; Figure 4 It is a flow chart of a method for determining the identification reliability coefficient of a screening intelligent algorithm. DETAILED DESCRIPTION

[0015] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below in conjunction with the drawings in the embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of the embodiments. Based on the embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this specification.

[0016] In the process of identifying the longitudinal and transverse diameters of grape grains, the distribution of grape grains is often too concentrated, which makes the identification of the longitudinal and transverse diameters of grape grains too difficult. Therefore, it is necessary to combine multiple intelligent image algorithms to identify the aspect ratio of grape grains.

[0017] Based on the proportion of the number of grape grains in different spacing distribution intervals, the weight coefficients of different spacing distribution intervals are determined, and the recognition accuracy of the longitudinal and transverse diameters of the grape grains is determined by using the sum of the products of the weight coefficients and the preset accuracy of the different spacing distribution intervals. When the recognition accuracy is greater than 0.6, it is determined that the recognition accuracy of the longitudinal and transverse diameters of the grape grains meets the requirements.

[0018] Based on the historical accuracy of the intelligent algorithm in longitudinal and transverse diameter recognition processing, the number of recognition times with a recognition accuracy greater than the preset accuracy is calculated. When the proportion of the number of recognition times with a recognition accuracy greater than the preset accuracy is greater than 0.5, the intelligent algorithm is determined to be a screening intelligent algorithm.

[0019] The distribution data of the recognition deviation grape grains in the screening intelligent algorithm is used to determine the distribution number of the recognition deviation grape grains corresponding to the screening intelligent algorithm. Based on the proportion of the number of recognition deviation grape grains in the number of recognition targets, the recognition reliability coefficient of the screening intelligent algorithm is determined. The screening intelligent algorithm with a recognition reliability coefficient greater than 0.9 is used as a matching intelligent algorithm.

[0020] The recognition result of the longitudinal and transverse diameters of the target grape grains is determined according to the average value of the recognition results of different matching intelligent algorithms on the longitudinal and transverse diameters of different grape grains.

[0021] Intelligent algorithms include Convolutional Neural Networks (CNN), Support Vector Machines (SVM), Decision Trees, Random Forests, and K-Nearest Neighbors (KNN).

[0022] Example 1 To solve the above problems, according to one aspect of the present invention, Figure 1 As shown, a method for identifying the longitudinal and transverse diameters of grape grains based on an intelligent image algorithm is provided, which specifically includes: S1 determines the grape variety of the identification target based on the distribution of the spacing between adjacent grapes of the identification target and when the identification accuracy of the longitudinal and transverse diameters of the grapes meets the requirements; Furthermore, the distribution of the spacing between adjacent grape grains is determined based on the image recognition results of the grape grains, and specifically based on the analysis results of the image of the recognition target.

[0023] Specifically, Figure 2 As shown, the method for determining the recognition accuracy of the longitudinal and transverse diameters of the grape grains is: Determine the distribution amount of the spacing between adjacent grape grains based on the spacing distribution between adjacent grape grains of the identified target; Determine the matching accuracy of different spacing distribution intervals based on the preset recognition accuracy corresponding to different spacing distribution intervals; Based on the proportion of the number of grape grains in different spacing distribution intervals, the weight coefficients of different spacing distribution intervals are determined, and combined with the matching accuracy of different spacing distribution intervals, the recognition accuracy of the longitudinal and transverse diameters of the grape grains is determined.

[0024] Optionally, the recognition accuracy of the longitudinal and transverse diameters of the grape particles ranges from 0 to 1. When the recognition accuracy of the longitudinal and transverse diameters of the grape particles is greater than a preset accuracy threshold, it is determined that the recognition accuracy of the longitudinal and transverse diameters of the grape particles meets the requirements.

[0025] In one possible embodiment, the method for determining the recognition accuracy of the longitudinal and transverse diameters of the grape grains is: The spacing distribution between adjacent grape grains of the identification target is used to determine the spacing distribution between adjacent grape grains. When the spacing distribution between different adjacent grape grains is greater than a preset distance threshold, it is determined that the recognition accuracy of the longitudinal and transverse diameters of the grape grains meets the requirements. When there are grapes whose spacing distribution is not greater than the preset distance threshold: Grapes with a spacing distribution amount not greater than a preset distance threshold are regarded as adjacent grapes. When the number of the adjacent grapes is greater than the preset number of grapes, it is determined that the recognition accuracy of the longitudinal and transverse diameters of the grapes does not meet the requirements; When the number of adjacent grapes is not greater than the preset number of grapes: Based on the preset recognition accuracy rates corresponding to different spacing distribution intervals, the matching accuracy rates of different spacing distribution intervals are determined, and when the number of grape grains in the preset spacing distribution interval does not meet the requirement, it is determined that the recognition accuracy rates of the longitudinal and transverse diameters of the grape grains do not meet the requirement; When the number of grapes in the preset distribution interval does not meet the requirements: Based on the proportion of the number of grape grains in different spacing distribution intervals, the weight coefficients of different spacing distribution intervals are determined, and the ratio of the matching accuracy rate to the weight coefficient is used to determine the corrected recognition accuracy rate. When the corrected recognition accuracy rates of different spacing distribution intervals all meet the requirements, it is determined that the recognition accuracy rates of the longitudinal and transverse diameters of the grape grains meet the requirements; When there is a gap distribution interval where the correction recognition accuracy does not meet the requirements: The intervals of the spacing distribution amount in which the corrected recognition accuracy does not meet the requirement are used as recognition deviation intervals. When the number of the recognition deviation intervals is greater than the preset number of deviation intervals, it is determined that the recognition accuracy of the longitudinal and transverse diameters of the grape grains does not meet the requirement. When the number of the identified deviation intervals is not greater than the number of preset deviation intervals: The recognition accuracy of the longitudinal and transverse diameters of the grape particles is determined by using the weight coefficients of different spacing distribution intervals and the matching accuracy.

[0026] It should be noted that the grape variety of the recognition target is determined according to the image recognition result of the recognition target, the image of the recognition target is used as input, and the output of a preset grape variety recognition model is used as the image recognition result of the recognition target.

[0027] S2 determines the historical accuracy of different intelligent algorithms in longitudinal and transverse diameter recognition processing based on the grape variety of the recognition target and the distribution of the spacing between adjacent grape grains, and determines the screening intelligent algorithm among the intelligent algorithms using the historical accuracy; Furthermore, the historical accuracy of the intelligent algorithm in performing longitudinal and transverse diameter recognition processing is determined according to the recognition accuracy of similar recognition times of the intelligent algorithm in the grape variety.

[0028] It should be noted that the similar identification times are the identification times in which the deviations of the number of grape grains in different spacing distribution intervals are all within the preset deviation number interval.

[0029] Specifically, Figure 3 As shown, the method for determining the screening intelligent algorithm in the intelligent algorithm is: Determining recognition accuracy rates at different similar recognition times based on recognition accuracy rates of similar recognition times in the grape variety; Using the recognition accuracy rates at different similar recognition times, a similar recognition time with a recognition accuracy rate greater than a preset accuracy rate is performed; Whether the intelligent algorithm is a screening intelligent algorithm is determined based on the proportion of similar identification times with an identification accuracy greater than a preset accuracy.

[0030] Optionally, the method for determining the screening intelligent algorithm in the intelligent algorithm is: Determining recognition accuracy rates at different similar recognition times based on recognition accuracy rates of similar recognition times in the grape variety; The average value of the accuracy of the intelligent algorithm is determined by the average value of the recognition accuracy of different similar recognition times, and the accuracy evaluation value is used to determine whether the intelligent algorithm is a screening intelligent algorithm.

[0031] In another embodiment, the method for determining the screening intelligent algorithm in the intelligent algorithm is: S21 determines the recognition accuracy rates at different similar recognition times based on the recognition accuracy rates of the similar recognition times in the grape variety; S22 determines similarity weight coefficients of different similarity identification times based on the average value of the deviation coefficients of the number of grape grains in different spacing distribution intervals in different similarity identification times; S23 determines the accuracy evaluation value of the intelligent algorithm based on the average value of the product of the similarity weight coefficient and the recognition accuracy rate of different similar recognition times, and uses the accuracy evaluation value to determine whether the intelligent algorithm is a screening intelligent algorithm.

[0032] Furthermore, when the accuracy evaluation value is greater than a preset evaluation value threshold, it is determined that the intelligent algorithm does not belong to the screening intelligent algorithm.

[0033] Optionally, the above step S21 includes the following contents: S211 determines the recognition accuracy rates at different similar recognition times based on the recognition accuracy rates of the similar recognition times in the grape variety. When the recognition accuracy rates of different similar recognition times are all less than a preset accuracy threshold, it is determined that the intelligent algorithm does not belong to the screening intelligent algorithm. When there is a similar recognition time with a recognition accuracy rate not less than the preset accuracy threshold, the process proceeds to step S212. S212: The number of similar recognitions with a recognition accuracy rate not less than a preset accuracy rate threshold is regarded as the number of accurate recognitions. When the number of accurate recognitions is greater than the preset accuracy threshold, the process proceeds to step S213. When the number of accurate recognitions is not greater than the preset accuracy threshold, the process proceeds to step S22. S213 When the proportion of accurate recognition times in the similar recognition times is greater than the preset proportion threshold, it is determined that the intelligent algorithm belongs to a screening intelligent algorithm. When the proportion of accurate recognition times in the similar recognition times is not greater than the preset proportion threshold, the process proceeds to step S23.

[0034] Optionally, the above step S22 includes the following contents: S221 determines the basic recognition reliability coefficient based on the number of accurate recognition times and the proportion of the number of accurate recognition times in the number of similar recognition times. When the basic recognition reliability coefficient is greater than the preset reliability coefficient threshold, it is determined that the intelligent algorithm belongs to the screening intelligent algorithm. When the basic recognition reliability coefficient is not greater than the preset reliability coefficient threshold, it proceeds to step S222; S222: when the basic identification reliability coefficient is within the preset reliability coefficient interval, the process proceeds to step S223; when the basic identification reliability coefficient is not within the preset reliability coefficient interval, it is determined that the intelligent algorithm does not belong to the screening intelligent algorithm; S223: Determine the similarity weight coefficients of different similarity identification times based on the average value of the deviation coefficients of the number of grape grains in different spacing distribution intervals in different similarity identification times, and determine the correction accuracy of different similarity identification times by multiplying the similarity weight coefficients of different similarity identification times by the recognition accuracy rate. When there is no similarity identification times with a correction accuracy rate greater than a preset correction threshold, it is determined that the intelligent algorithm does not belong to the screening intelligent algorithm. When there is a similarity identification times with a correction accuracy rate greater than the preset correction threshold, it proceeds to step S224. S224 When the number of similar recognitions with a correction accuracy greater than the preset correction threshold is greater than the preset recognition number threshold, it is determined that the intelligent algorithm does not belong to the screening intelligent algorithm; when the number of similar recognitions with a correction accuracy greater than the preset correction threshold is not greater than the preset recognition number threshold, proceed to step S23.

[0035] S3 determines the recognition processing results of the longitudinal and transverse diameters of grape grains of different screening intelligent algorithms, determines the distribution data of the recognition deviation grape grains in different screening intelligent algorithms by using the longitudinal and transverse diameter preset intervals of the recognition target and the recognition processing results, and determines the recognition reliability coefficients of different screening intelligent algorithms and matching intelligent algorithms in combination with the deviation data of the different recognition deviation grape grains; It should be noted that the recognition processing result of the longitudinal and transverse diameters of the grape grains includes the recognition result of the longitudinal and transverse diameters of the grape grains.

[0036] Furthermore, the preset longitudinal and transverse diameter intervals of the identified target are determined according to the grape variety of the identified target, specifically according to the longitudinal diameter interval and transverse diameter interval corresponding to the grape variety.

[0037] It should be noted that the identified deviation grape grains are grape grains whose identification results of longitudinal and transverse diameters are not within the preset longitudinal and transverse diameter ranges.

[0038] Specifically, Figure 4 As shown, the method for determining the identification reliability coefficient of the screening intelligent algorithm is: Determine the distribution number of the identified deviation grape grains corresponding to the screening intelligent algorithm using the distribution data of the identified deviation grape grains in the screening intelligent algorithm; Based on the proportion of the distribution number of grape grains with identification deviations in the number of identification targets, the identification reliability coefficient of the screening intelligent algorithm is determined.

[0039] Furthermore, the recognition reliability coefficient is the difference between a preset value and the recognition deviation, which is the proportion of the distribution number of grape grains in the number of the recognition target.

[0040] It should also be noted that when the recognition reliability coefficient of the screening intelligent algorithm is greater than a preset reliability coefficient threshold, the screening intelligent algorithm is determined to be a matching intelligent algorithm.

[0041] Optionally, the method for determining the identification reliability coefficient of the screening intelligent algorithm is: Determine the number of identification deviation grape grains in different images of the identification target by using the distribution data of the identification deviation grape grains in the screening intelligent algorithm; Determine the recognition deviation coefficients of different images based on the number ratio of recognition deviation grape grains in different images; The recognition reliability coefficient of the screening intelligent algorithm is determined according to the average value of the recognition deviation coefficients of different images.

[0042] Optionally, the method for determining the identification reliability coefficient of the screening intelligent algorithm is: S31 determines the number of identification deviation grape grains in different images of the identification target by using the distribution data of the identification deviation grape grains in the screening intelligent algorithm, and determines the deviation coefficients of different identification deviation grape grains by using the deviation data of different identification deviation grape grains; S32 determines recognition deviation coefficients of different images based on the number of recognition deviation grape grains in different images, the number proportion, and the deviation coefficients of different recognition deviation grape grains; S33 determines the recognition reliability coefficient of the screening intelligent algorithm according to the recognition deviation coefficients of different images.

[0043] Optionally, the above step S31 includes the following contents: S311 determines the number of recognized deviation grapes in different images in the recognition target based on the distribution data of the recognized deviation grapes in the screening intelligent algorithm. When the total number of recognized deviation grapes in different images does not meet the requirement, it is determined that the screening intelligent algorithm does not belong to the matching intelligent algorithm. When the total number of recognized deviation grapes in different images meets the requirement, it proceeds to step S312. S312 uses the number of recognized deviation grapes in different images to identify images whose number of recognized deviation grapes does not meet the requirement. When there is an image whose number of recognized deviation grapes does not meet the requirement, it is determined that the screening intelligent algorithm does not belong to the matching intelligent algorithm. When there is no image whose number of recognized deviation grapes does not meet the requirement, the process proceeds to step S313. S313 uses the deviation data of different identification deviation grape grains to determine the deviation coefficients of different identification deviation grape grains. When there are identification deviation grape grains whose deviation coefficients do not meet the requirements, the process proceeds to step S314. When there are no identification deviation grape grains whose deviation coefficients do not meet the requirements, the process proceeds to step S32. S314 When the number of identified deviation grape grains whose deviation coefficient does not meet the requirements is greater than the preset deviation grape grain number, it is determined that the screening intelligent algorithm does not belong to the matching intelligent algorithm. When the number of identified deviation grape grains whose deviation coefficient does not meet the requirements is not greater than the preset deviation grape grain number, proceed to step S32.

[0044] Optionally, the above step S32 includes the following contents: S321 determines the recognition deviation coefficients of different images based on the number of recognition deviation grapes in different images, the proportion of the number and the deviation coefficients of different recognition deviation grapes. When there is an image whose recognition deviation coefficient does not meet the requirements, the process proceeds to step S322. When there is no image whose recognition deviation coefficient does not meet the requirements, the process proceeds to step S33. S322: the images whose recognition deviation coefficients do not meet the requirements are used as recognition deviation images. When the number of the recognition deviation images does not meet the requirements, it is determined that the screening intelligent algorithm does not belong to the matching intelligent algorithm. When the number of the recognition deviation images meets the requirements, the process proceeds to step S323. S323 determines the image weight coefficients of different recognition deviation images based on the number of grape grains in different recognition deviation images. When the sum of the image weight coefficients of different recognition deviation images does not meet the requirements, it is determined that the screening intelligent algorithm does not belong to the matching intelligent algorithm. When the sum of the image weight coefficients of different recognition deviation images meets the requirements, it proceeds to step S33.

[0045] S4 outputs the recognition results of the longitudinal and transverse diameters of the grape grains of the recognition target based on the recognition processing results of different matching intelligent algorithms and the recognition reliability coefficients.

[0046] Specifically, the method for outputting the recognition result of the longitudinal and transverse diameters of the target grape grains is as follows: Determine the recognition results of different grape grains of the recognition target in terms of the longitudinal and transverse diameters of the matching intelligent algorithm based on the recognition processing results of the matching intelligent algorithm; Based on the deviation of the recognition results of the longitudinal and transverse diameters of different grapes in different matching intelligent algorithms, the recognition results are divided into different similarity intervals; Determine the sum of the recognition reliability coefficients in different similarity intervals by the sum of the recognition reliability coefficients of the matching intelligent algorithm corresponding to the recognition results in different similarity intervals, and take the similarity interval with the largest sum of the recognition reliability coefficient as the matching similarity interval; The average value of the recognition results of the longitudinal and transverse diameters of different matching intelligent algorithms in the matching similar interval is output as the recognition result of the longitudinal and transverse diameters of the grape grains.

[0047] Furthermore, the recognition results are divided into different similarity intervals, specifically including: The recognition results of the deviation of the longitudinal and transverse diameters within the preset deviation range are divided into the same similarity range.

[0048] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, equipment, and non-volatile computer storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0049] The above is a description of a specific embodiment of the present specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0050] The above description is only one or more embodiments of this specification and is not intended to limit this specification. For those skilled in the art, one or more embodiments of this specification may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of one or more embodiments of this specification shall be included in the scope of the claims of this specification.

Claims

1. A method for identifying the longitudinal and transverse diameters of grapes based on an intelligent image algorithm, characterized in that: Specifically include: When the recognition accuracy of the longitudinal and transverse diameters of the grape grains meets the requirements based on the distribution of the spacing between adjacent grape grains of the recognition target, the grape variety of the recognition target is determined; Based on the grape variety of the identification target and the distribution of the spacing between adjacent grapes, determine the historical accuracy of different intelligent algorithms when performing longitudinal and transverse diameter identification processing, and use the historical accuracy to determine the screening intelligent algorithm among the intelligent algorithms; Determine the recognition processing results of the longitudinal and transverse diameters of grape grains of different screening intelligent algorithms, determine the distribution data of the recognition deviation grape grains in different screening intelligent algorithms by using the longitudinal and transverse diameter preset intervals of the recognition target and the recognition processing results, and determine the recognition reliability coefficients of different screening intelligent algorithms and matching intelligent algorithms in combination with the deviation data of different recognition deviation grape grains; Based on the recognition processing results and recognition reliability coefficients of different matching intelligent algorithms, the recognition results of the longitudinal and transverse diameters of the grape grains of the recognition target are output.

2. The method for identifying the longitudinal and transverse diameters of grapes based on an intelligent image algorithm as claimed in claim 1, characterized in that: The distribution of the spacing between adjacent grape grains is determined based on the image recognition results of the grape grains, and specifically based on the analysis results of the image of the recognition target.

3. The method for identifying the longitudinal and transverse diameters of grapes based on an intelligent image algorithm as claimed in claim 1, characterized in that: The method for determining the recognition accuracy of the longitudinal and transverse diameters of the grape grains is: Determine the distribution amount of the spacing between adjacent grape grains based on the spacing distribution between adjacent grape grains of the identified target; Determine the matching accuracy of different spacing distribution intervals based on the preset recognition accuracy corresponding to different spacing distribution intervals; Based on the proportion of the number of grape grains in different spacing distribution intervals, the weight coefficients of different spacing distribution intervals are determined, and combined with the matching accuracy of different spacing distribution intervals, the recognition accuracy of the longitudinal and transverse diameters of the grape grains is determined.

4. The method for identifying the longitudinal and transverse diameters of grapes based on an intelligent image algorithm as claimed in claim 3, characterized in that: The recognition accuracy of the longitudinal and transverse diameters of the grape particles ranges from 0 to 1. When the recognition accuracy of the longitudinal and transverse diameters of the grape particles is greater than a preset accuracy threshold, it is determined that the recognition accuracy of the longitudinal and transverse diameters of the grape particles meets the requirements.

5. The method for identifying the longitudinal and transverse diameters of grapes based on intelligent image algorithm according to claim 1, characterized in that: The grape variety of the recognition target is determined according to the image recognition result of the recognition target, the image of the recognition target is used as input, and the output of a preset grape variety recognition model is used as the image recognition result of the recognition target.

6. The method for identifying the longitudinal and transverse diameters of grapes based on intelligent image algorithm according to claim 1, characterized in that: The historical accuracy of the intelligent algorithm when performing longitudinal and transverse diameter recognition processing is determined according to the recognition accuracy of similar recognition times of the intelligent algorithm in the grape variety.

7. The method for identifying the longitudinal and transverse diameters of grapes based on intelligent image algorithm according to claim 6, characterized in that: The similar identification times are the identification times in which the deviations of the number of grape grains in different spacing distribution intervals are all within the preset deviation number interval.

8. The method for identifying the longitudinal and transverse diameters of grapes based on intelligent image algorithm according to claim 6, characterized in that: The method for determining the screening intelligent algorithm in the intelligent algorithm is: Determining recognition accuracy rates at different similar recognition times based on recognition accuracy rates of similar recognition times in the grape variety; Using the recognition accuracy rates at different similar recognition times, a similar recognition time with a recognition accuracy rate greater than a preset accuracy rate is performed; Whether the intelligent algorithm is a screening intelligent algorithm is determined based on the proportion of similar identification times with an identification accuracy greater than a preset accuracy.

9. The method for identifying the longitudinal and transverse diameters of grapes based on intelligent image algorithm according to claim 1, characterized in that: The method for outputting the recognition result of the longitudinal and transverse diameters of the grape grains of the recognition target is: Determine the recognition results of different grape grains of the recognition target in the longitudinal and transverse diameters of the matching intelligent algorithm based on the recognition processing results of the matching intelligent algorithm, and divide the recognition results into different similarity intervals based on the deviations of the recognition results of different grape grains in the longitudinal and transverse diameters of different matching intelligent algorithms; Determine the sum of the recognition reliability coefficients in different similarity intervals by the sum of the recognition reliability coefficients of the matching intelligent algorithm corresponding to the recognition results in different similarity intervals, and take the similarity interval with the largest sum of the recognition reliability coefficient as the matching similarity interval; The average value of the recognition results of the longitudinal and transverse diameters of different matching intelligent algorithms in the matching similar interval is output as the recognition result of the longitudinal and transverse diameters of the grape grains.

10. The method for identifying the longitudinal and transverse diameters of grapes based on intelligent image algorithm according to claim 9, characterized in that: The recognition results are divided into different similarity intervals, specifically including: The recognition results of the deviation of the longitudinal and transverse diameters within the preset deviation range are divided into the same similarity range.

Images