Digital management system and method for kiwi fruit resource acquisition information

Abstract

The invention relates to the technical field of data storage, and discloses a digital management system and method for kiwi fruit resource acquisition information, and the system comprises an acquisition module which is used for receiving a kiwi fruit image and parameter data uploaded by a user; the initial detection evaluation module is used for calculating a quality score based on image definition, exposure and edge features and detecting parameter integrity; the feature comparison module is used for extracting image features to construct a feature route map and calculating the matching degree with a standard map; the abnormity analysis module is used for judging based on the fruit diameter deviation and the temperature disturbance; and the uploading module is used for judging whether to upload the data to the database according to the deviation degree and the environmental disturbance so as to ensure the validity and the reliability of the acquired data. According to the method, the accuracy and reliability of kiwi fruit resource data are effectively improved through image quality evaluation and parameter integrity detection in combination with feature comparison and environment anomaly analysis.

Description

Technical Field

[0001] The present invention relates to the technical field of data storage, and in particular to a digital management system and method for kiwifruit resource collection information. Background Art

[0002] When establishing a database of plant specimens such as kiwifruit, it is usually necessary to manually take images of the fruit and manually fill in information such as variety, growth parameters, and collection environment. Technical personnel then conduct manual screening and standardized filing.

[0003] Because image acquisition is often performed by non-professionals, common issues include blurred images, uneven exposure, and irregular shooting angles. This severely reduces the effectiveness of subsequent image recognition and comparison, affecting the accuracy of fruit feature extraction. Furthermore, field entry during the acquisition process is prone to errors, making it difficult to support the construction of a standardized resource library.

[0004] In summary, the current kiwifruit resource collection and information management have problems such as inaccurate image quality assessment and lack of parameter integrity verification. It is urgent to build an intelligent collection information management system to improve the accuracy, standardization and information credibility of kiwifruit resource collection. Summary of the Invention

[0005] In view of this, the present invention proposes a digital management system and method for kiwifruit resource collection information to solve the problems of inaccurate image quality assessment and lack of parameter integrity verification in the prior art.

[0006] On the one hand, the present invention proposes a digital management system for kiwifruit resource collection information, comprising:

[0007] An acquisition module is configured to receive kiwi fruit images and corresponding parameter data uploaded by users;

[0008] The initial inspection and evaluation module is configured to calculate an image quality score for the kiwifruit image based on clarity, exposure, and edge feature values, perform integrity detection on the parameter data based on the ratio of the number of uploaded fields to the expected number of fields, and determine whether to input the kiwifruit image and parameter data into the feature comparison module based on the image quality score and integrity detection results;

[0009] The feature comparison module is configured to extract the texture of the kiwifruit image and the features of the fruit crown area, construct an image feature path map based on a deep image trajectory fitting algorithm, calculate the matching degree between the image feature path and the standard reference map, and determine whether to perform environmental and parameter anomaly analysis based on the matching degree;

[0010] The anomaly analysis module is configured to perform environmental and parameter anomaly analysis, including: calculating the degree of deviation of the fruit diameter parameter from the standard variety data; retrieving the temperature records before and after the kiwi fruit image was taken, and determining whether there is an environmental disturbance based on the temperature change rate;

[0011] The uploading module is configured to determine whether the uploading standard is met based on the degree of deviation and whether there is environmental disturbance, and to upload the kiwi fruit image and parameter data to the database.

[0012] Furthermore, when calculating the image quality score based on the clarity, exposure and edge feature values, it includes:

[0013] After graying the kiwifruit image, an edge response image is extracted by 3*3 Laplace operator convolution, and the variance of the edge response image is calculated. The variance of the edge response image and a preset variance value are used as the clarity; the preset variance value is the average variance of the edge response image of the historical kiwifruit image;

[0014] Set a high brightness interval and a low brightness interval, and calculate the sum of the proportions of pixels in the kiwi fruit image located in the high brightness interval and the low brightness interval, where the sum of the proportions is the exposure;

[0015] After extracting the edges in the kiwifruit image based on the Canny edge detection algorithm, the ratio of the number of edge points in the kiwifruit target area to the number of valid pixels in the kiwifruit image is calculated, and the ratio is the edge feature value.

[0016] Furthermore, when calculating the image quality score based on the clarity, exposure and edge feature values, it also includes:

[0017] The image quality score is the sum of the products of the clarity, exposure and edge feature values ​​and their corresponding weight coefficients.

[0018] Furthermore, the integrity detection includes:

[0019] Match the parameter data with the field template, and calculate the ratio of the number of fields in the parameter data to the expected number of fields in the field template, which is recorded as the field ratio;

[0020] Set format rules for each field of parameter data and use regular expressions to verify whether there are any missing fields;

[0021] When the field ratio is 0.8-1.2 and there is no missing field, the integrity test result is determined to be passed.

[0022] Furthermore, when the image quality score is greater than or equal to the preset image score upper threshold and the integrity test result passes, the kiwi fruit image and parameter data are input into the upload module and output;

[0023] When the image quality score is less than the preset image score upper threshold, greater than or equal to the preset image score lower threshold, and the integrity test result passes, the kiwi fruit image and parameter data are input into the feature comparison module;

[0024] When the image quality score is lower than the preset image score lower threshold, or the integrity test result fails, the kiwi fruit image and parameter data will not be uploaded.

[0025] Furthermore, when calculating the matching degree based on the image feature route and the standard reference image, it includes:

[0026] The target area of ​​kiwifruit images is segmented, and image blocks of the fruit crown area, surrounding texture area and background boundary area are extracted, and feature point detection and directional gradient calculation are performed respectively.

[0027] Based on the SIFT algorithm, the feature descriptor of local key points is extracted for each image block, and a set of image partition feature vectors is constructed;

[0028] Perform the same processing on the corresponding area in the standard reference image to obtain a set of standard image feature vectors;

[0029] Each feature vector set of the kiwifruit image is spliced ​​according to the physical space order of the image blocks to form a two-dimensional feature trajectory matrix, and Bezier curve interpolation fitting is used to generate a continuous image feature path;

[0030] Calculate the DTW distance between the feature path of the kiwi image and the feature path of the standard image;

[0031] Map the DTW distance value to a matching degree in the interval [0,1].

[0032] Furthermore, when determining whether to perform an environment and parameter anomaly analysis based on the matching degree, it includes:

[0033] Divide the matching degree into low matching degree interval, medium matching degree interval and high matching degree interval in order from small to large;

[0034] When the matching degree is in the high matching degree range, the kiwi fruit image and the corresponding parameter data are transmitted to the upload module;

[0035] When the matching degree is in the medium matching degree range, the kiwi fruit image and the corresponding parameter data are transmitted to the abnormality analysis module;

[0036] When the matching degree is in a low matching degree range, the kiwi fruit image and the corresponding parameter data are not transmitted to the abnormality analysis module or the upload module.

[0037] Furthermore, the degree of deviation is the ratio of the fruit diameter in the kiwifruit image to the fruit diameter in the standard variety data; when judging whether there is environmental disturbance, the temperature record is obtained, the retrieval length of the temperature record is proportional to the current maximum temperature difference at the shooting time, and the standard deviation of all temperature values ​​in the temperature record is calculated. When the standard deviation is greater than or equal to the reference standard deviation, it is judged that there is environmental disturbance.

[0038] Furthermore, when the degree of deviation is less than a preset deviation threshold and there is no environmental disturbance, the kiwi fruit image and parameter data are input into an uploading module;

[0039] When the degree of deviation is greater than or equal to the deviation threshold, or when there is environmental disturbance, the kiwi fruit image and parameter data are returned.

[0040] Compared with the prior art, the present invention has the following beneficial effects:

[0041] This invention effectively improves the accuracy of kiwifruit image screening through a multi-dimensional image quality assessment method. Grayscaling and the Laplacian operator are used to calculate the variance of edge response images, accurately reflecting image clarity. Exposure is comprehensively assessed by combining the ratio of high- and low-brightness pixels, and Canny edge detection is used to extract target edge feature values, ensuring that only images that meet quality standards are processed. Parameter data is integrity-checked using field ratios and formatting rules to avoid missing or incorrect data and ensure standard information entry. Based on a deep image trajectory fitting algorithm and DTW distance calculation, kiwifruit image feature paths are precisely constructed and matched against a standard reference image, effectively improving the accuracy and stability of feature comparison. Matching degree zoning controls subsequent processes, rationally screening out low-quality or abnormal samples and reducing the risk of erroneous uploads. The anomaly analysis module integrates fruit diameter deviation and ambient temperature disturbances to enhance the ability to identify abnormal data and prevent data contaminated by environmental factors from being mistakenly entered into the database. The overall system boasts a high degree of automation, reducing manual intervention, improving data collection efficiency and data reliability, and enhancing the standardization, integrity, and practical value of the kiwifruit resource database.

[0042] On the other hand, the present invention proposes a digital management method for kiwifruit resource collection information, comprising:

[0043] Receive kiwi fruit images and corresponding parameter data uploaded by users;

[0044] Calculate the image quality score of kiwifruit images based on clarity, exposure, and edge feature values. Perform integrity checks on parameter data based on the ratio of the number of uploaded fields to the expected number of fields. Determine whether to construct an image feature roadmap based on the image quality score and integrity test results.

[0045] When it is necessary to construct an image feature roadmap, the texture of the kiwifruit image and the features of the fruit crown area are extracted, and the image feature roadmap is constructed based on the deep image trajectory fitting algorithm. The matching degree between the image feature roadmap and the standard reference image is calculated, and the decision on whether to perform environmental and parameter anomaly analysis is made based on the matching degree.

[0046] When conducting environmental and parameter anomaly analysis, it includes: comparing the degree of deviation of the fruit diameter parameter with the standard variety data; retrieving the temperature records before and after the kiwifruit image was taken, and judging whether there is environmental disturbance based on the temperature change rate;

[0047] Based on the degree of deviation and whether there is environmental disturbance, determine whether the upload criteria are met and upload the kiwifruit image and parameter data to the database.

[0048] It should be noted that the digital management method and system for kiwifruit resource collection information of the present invention have the same beneficial effects and will not be described in detail here. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] Various other advantages and benefits will become apparent to those skilled in the art upon reading the detailed description of the preferred embodiment below. The accompanying drawings are for illustration purposes only and are not to be considered as limiting the present invention. The same reference symbols are used throughout the drawings to represent the same components. In the drawings:

[0050] Figure 1 This is a functional block diagram of a digital management system for kiwifruit resource collection information provided by an embodiment of the present invention.

[0051] Figure 2 The present invention provides a flowchart of a digital management method for kiwifruit resource collection information. DETAILED DESCRIPTION

[0052] The exemplary embodiments disclosed in the present application will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, unless there is a conflict, the embodiments of the present invention and the features described in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0053] See Figure 1 As shown, the embodiment of the present invention provides a digital management system for kiwifruit resource collection information, including:

[0054] An acquisition module is configured to receive kiwi fruit images and corresponding parameter data uploaded by users;

[0055] The initial inspection and evaluation module is configured to calculate an image quality score for the kiwifruit image based on clarity, exposure, and edge feature values, perform integrity detection on the parameter data based on the ratio of the number of uploaded fields to the expected number of fields, and determine whether to input the kiwifruit image and parameter data into the feature comparison module based on the image quality score and integrity detection results;

[0056] The feature comparison module is configured to extract the texture of the kiwifruit image and the features of the fruit crown area, construct an image feature path map based on a deep image trajectory fitting algorithm, calculate the matching degree between the image feature path and the standard reference map, and determine whether to perform environmental and parameter anomaly analysis based on the matching degree;

[0057] The anomaly analysis module is configured to perform environmental and parameter anomaly analysis, including: calculating the degree of deviation of the fruit diameter parameter from the standard variety data; retrieving the temperature records before and after the kiwi fruit image was taken, and determining whether there is an environmental disturbance based on the temperature change rate;

[0058] The uploading module is configured to determine whether the uploading standard is met based on the degree of deviation and whether there is environmental disturbance, and to upload the kiwi fruit image and parameter data to the database.

[0059] It can be understood that the present invention realizes the intelligent digital management of kiwifruit resource collection information through the collaborative work of multiple modules. The acquisition module effectively integrates image and parameter data, and the initial inspection and evaluation module scientifically quantifies the image quality based on clarity, exposure and edge features, while strictly checking the parameter integrity to ensure that the data is accurate and reliable. The feature comparison module uses a deep image trajectory fitting algorithm to accurately extract key features and match them with the standard reference image, thereby improving the recognition accuracy and objectivity of the judgment. The abnormal analysis module combines the fruit diameter deviation with the ambient temperature change rate to dynamically identify abnormal data and effectively prevent abnormal samples from being uploaded by mistake. The upload module makes intelligent decisions on data upload based on the comprehensive judgment results to ensure the standardization and credibility of the information in the database. The overall system improves the automation level and management efficiency of kiwifruit resource data collection, and provides a solid guarantee for the establishment of a high-quality, standardized kiwifruit resource database.

[0060] It should be noted that, before and after the shooting time refers to n days before or after the shooting time.

[0061] In some embodiments of the present application, when calculating the image quality score based on the clarity, exposure, and edge feature values, the following steps are included:

[0062] After graying the kiwifruit image, the edge response image is extracted through 3*3 Laplace operator convolution, and the variance of the edge response image is calculated. The variance of the edge response image and the preset variance value are the clarity; the preset variance value is the average variance of the edge response image of the historical kiwifruit image;

[0063] Set high brightness range and low brightness range, and calculate the sum of the proportions of pixels in the kiwi fruit image that are in the high brightness range and the low brightness range. The total proportion is the exposure;

[0064] After extracting the edges in the kiwifruit image based on the Canny edge detection algorithm, the ratio of the number of edge points in the kiwifruit target area to the number of valid pixels in the kiwifruit image is calculated, and the ratio is the edge feature value.

[0065] It should be noted that the collected color kiwifruit image is first converted to a grayscale image. During the conversion, the color information of each pixel is calculated as a corresponding grayscale value based on the different weights of the three channels (red, green, and blue), forming a single-channel grayscale image.

[0066] Next, a three-by-three Laplacian operator is used to convolve the grayscale image. Specifically, the Laplacian operator's matrix template is multiplied and accumulated with the corresponding pixel values ​​of each pixel in the image and its surrounding neighborhood to obtain a new image, which is the edge response image.

[0067] Next, statistics are taken for all pixel values ​​in the edge response image, and the variance of these pixel values ​​is calculated. Variance reflects the degree of dispersion in the distribution of pixel values ​​and can quantify the edge clarity of the image. To set a standard, the variances of the edge response images of a large number of historical kiwifruit images are averaged to obtain a preset reference variance value.

[0068] The image clarity is evaluated by comparing the variance of the edge response image of the current image with the preset reference variance. If the edge response image variance is close to or exceeds the preset value, it means that the image has high clarity.

[0069] Next, define high-brightness intervals and low-brightness intervals within the image's grayscale range. For example, pixels with grayscale values ​​in the higher range are defined as the high-brightness interval, while pixels with grayscale values ​​in the lower range are defined as the low-brightness interval. Then, we iterate over all pixels in the grayscale image, counting the number of pixels that fall into these two intervals. The sum of these two values ​​is compared with the total number of pixels in the image to obtain a single value representing the image's exposure. A higher value indicates overexposure or underexposure.

[0070] Next, the grayscale image is processed using the Canny edge detection algorithm to extract edge contours. This algorithm uses a multi-step process, including noise reduction, gradient calculation, non-maximum suppression, and edge connection, to ultimately generate a binary image of edge points.

[0071] On this basis, the target area of ​​the kiwi fruit is determined, which can be achieved by color thresholding or other simple image segmentation techniques. Then the number of edge points in the target area is counted, and the number of valid pixels in the entire kiwi fruit image is counted, excluding the background part.

[0072] Finally, the ratio of the number of edge points in the target area to the number of valid pixels is calculated to obtain the edge feature value, which reflects the richness and clarity of the edge of the fruit area in the image.

[0073] In some embodiments of the present application, when calculating the image quality score based on the clarity, exposure, and edge feature values, the following steps are further included:

[0074] The image quality score is the sum of the product of the clarity, exposure and edge feature values ​​and their corresponding weight coefficients.

[0075] In some embodiments of the present application, integrity detection includes:

[0076] Match the parameter data with the field template, and calculate the ratio of the number of fields in the parameter data to the expected number of fields in the field template, which is recorded as the field ratio;

[0077] Set format rules for each field of parameter data and use regular expressions to verify whether there are any missing fields;

[0078] When the field ratio is 0.8-1.2 and there is no missing field, the integrity test result is judged to be passed.

[0079] It should be noted that the field template defines the names and order of all the fields that should be included in the kiwifruit parameter data, including variety, harvest time, fruit diameter, and growth environment temperature. The field template represents the complete specification standard of the data.

[0080] When receiving parameter data uploaded by a user, the system compares the number of fields in the data with the expected number of fields in the field template and calculates the ratio between the two: the number of uploaded fields divided by the expected number of fields. This ratio reflects the degree of field coverage in the uploaded data. If the ratio is too low, it means that the uploaded data is missing many fields and the data structure is incomplete; if it is too high, it may contain irrelevant or redundant fields.

[0081] Next, the system performs format verification on each field in the uploaded parameter data. Format rules include the field's data type (e.g., string, number, date), the allowed value range, and the data format specifications. To achieve this, the system uses regular expressions to match the content of each field. For example, date fields must conform to the "year-month-day" format, numeric fields must be positive and within a reasonable range, and string fields cannot contain illegal characters.

[0082] The system uses regular expression validation to determine whether a field is missing content or has an incorrect format. If a field is found to be empty, formatted incorrectly, or unreasonable, it is considered missing.

[0083] The integrity check is considered passed when the field ratio is within a reasonable range, that is, the number of uploaded fields is close to 80% to 120% of the expected number of fields, and all fields pass format verification with no missing or incorrect fields. If these conditions are not met, the parameter data is considered incomplete and fails the test.

[0084] In some embodiments of the present application, when the image quality score is greater than or equal to a preset image score upper threshold and the integrity test result passes, the kiwi fruit image and parameter data are input into the upload module and output;

[0085] When the image quality score is less than the preset image score upper threshold, greater than or equal to the preset image score lower threshold, and the integrity test result passes, the kiwi fruit image and parameter data are input into the feature comparison module;

[0086] When the image quality score is lower than the preset image score lower threshold, or the integrity test result fails, the kiwi fruit image and parameter data will not be uploaded.

[0087] It should be noted that this embodiment determines the subsequent processing flow of the kiwi fruit image and its parameter data by setting upper and lower thresholds of the image quality score and combining the parameter data integrity detection results, thereby ensuring the quality and validity of the uploaded data.

[0088] First, when the kiwifruit image's quality score reaches or exceeds the preset high-quality threshold and the corresponding parameter data integrity check passes, it indicates that the image is clear, properly exposed, has good edge features, and that the parameter data fields are complete and correctly formatted, indicating excellent overall data quality. At this point, the system directly inputs the image and parameter data into the upload module, completing data upload and storage without requiring additional feature comparison or anomaly analysis, improving processing efficiency and ensuring rapid storage of high-quality data.

[0089] Secondly, when the image quality score is below the high-quality threshold but not below the preset low-quality threshold, and the parameter data integrity test also passes, it indicates that although the image quality is not perfect, it is still within an acceptable range and the parameter data is complete and reliable. At this time, the system will not upload the data directly, but will instead send the data to the feature comparison module for further image feature extraction and matching with the standard reference image. The image's texture and crown area features are analyzed using a deep learning algorithm to confirm the image's match with the standard kiwifruit image and determine its validity. This step can effectively eliminate recognition errors that may be caused by poor image quality and improve overall data accuracy.

[0090] Finally, if the image quality score falls below the preset low-quality threshold or fails the parameter data integrity check, it indicates that the image is blurry, underexposed, has unclear edges, or that parameter fields are missing or have serious formatting errors, indicating poor data quality. To prevent erroneous or invalid data from entering the database, the system directly rejects the upload, avoiding contamination of database resources and impacting subsequent data analysis and management.

[0091] Through the above hierarchical judgment mechanism, the system can reasonably distinguish the data quality levels and adopt different processing strategies in a targeted manner, which not only ensures data quality but also improves processing efficiency and ensures the reliability and practicality of the kiwifruit resource database.

[0092] In some embodiments of the present application, when calculating the matching degree based on the image feature route and the standard reference image, the process includes:

[0093] The target area of ​​kiwifruit images is segmented, and image blocks of the fruit crown area, surrounding texture area and background boundary area are extracted, and feature point detection and directional gradient calculation are performed respectively.

[0094] Based on the SIFT algorithm, the feature descriptor of local key points is extracted for each image block, and a set of image partition feature vectors is constructed;

[0095] Perform the same processing on the corresponding area in the standard reference image to obtain a set of standard image feature vectors;

[0096] Each feature vector set of the kiwifruit image is spliced ​​according to the physical space order of the image blocks to form a two-dimensional feature trajectory matrix, and Bezier curve interpolation fitting is used to generate a continuous image feature path;

[0097] Calculate the DTW distance between the feature path of the kiwi image and the feature path of the standard image;

[0098] Map the DTW distance value to a matching degree in the range [0,1].

[0099] It should be noted that the kiwifruit image is segmented into several key areas: the crown of the fruit, the surrounding texture area, and the boundary between the fruit and the background. Feature points are detected in each of these three areas, and the directional gradient information of each feature point is calculated to capture the local shape and texture characteristics of the image.

[0100] Next, the SIFT (Scale-Invariant Feature Transform) algorithm is used to extract feature descriptors for the local key points in each image block. These descriptors represent the local image information of the key points and are scale- and rotation-invariant, helping to improve the robustness of the matching. The feature descriptors extracted from each image block are aggregated to form a corresponding set of feature vectors, which are used to describe the local features of that image block.

[0101] Subsequently, the same processing steps are performed on the corresponding areas of the standard reference image to obtain a set of feature vectors for each area of ​​the standard image as a basis for comparison.

[0102] The feature vectors of each kiwifruit image block are then concatenated according to their physical spatial order within the image to construct a two-dimensional feature trajectory matrix. To obtain a more continuous and smooth feature path, Bezier curve interpolation is used to fit the trajectory, generating a continuous image feature path.

[0103] Finally, the dynamic time warping distance (DTW distance) is calculated between the feature paths of the kiwifruit image and the feature paths of the standard reference image. This distance measures the similarity of the two feature paths over time, adapting to possible deformations and local differences in the image. The calculated DTW distance is then transformed through a mapping to obtain a matching value between zero and one. The closer the value is to one, the higher the matching degree and the greater the similarity between the image and the standard reference image.

[0104] In some embodiments of the present application, when determining whether to perform an environment and parameter anomaly analysis based on the matching degree, it includes:

[0105] Divide the matching degree into low matching degree interval, medium matching degree interval and high matching degree interval in order from small to large;

[0106] When the matching degree is in the high matching degree range, the kiwi fruit image and the corresponding parameter data are transmitted to the upload module;

[0107] When the matching degree is in the medium matching degree range, the kiwi fruit image and the corresponding parameter data are transmitted to the abnormality analysis module;

[0108] When the matching degree is in a low matching degree range, the kiwi fruit image and the corresponding parameter data are not transmitted to the abnormality analysis module or the upload module.

[0109] It should be noted that, first, according to the matching degree values ​​between the kiwi fruit image and the standard reference image, all matching degrees are divided into three intervals from small to large, namely, low matching degree interval, medium matching degree interval and high matching degree interval. This partitioning strategy helps to take differentiated processing measures according to the different matching degrees. In this embodiment, the low matching degree interval is [0, 0.3], the medium matching degree interval is (0.3, 0.8], and the high matching degree interval is (0.8, 1.0].

[0110] When the match is high, the kiwifruit image is highly similar to the standard reference image, indicating good image quality and reliable data. At this point, the system directly transmits the kiwifruit image and its corresponding parameter data to the upload module for storage and archiving, eliminating the need for further anomaly detection and improving processing efficiency.

[0111] A medium match indicates that the image differs from the reference image. This may be due to slightly poor image quality, changes in the acquisition environment, or parameter anomalies. To prevent potential data anomalies from affecting the database, the system sends the image and parameter data to the Anomaly Analysis Module, which further analyzes deviations in the fruit diameter parameter and temperature changes in the shooting environment to determine if any anomalies exist, ensuring the accuracy and reliability of the uploaded data.

[0112] When the match is in the low range, it indicates that the kiwifruit image has a low similarity to the standard reference image, the data quality is obviously insufficient, or the parameters are seriously abnormal. To prevent invalid or erroneous data from entering the database, the system will not transmit these images and parameter data to the abnormality analysis module or upload module, but will directly discard or return them to ensure data quality control.

[0113] In some embodiments of the present application, the degree of deviation is the ratio of the fruit diameter in the kiwi fruit image to the fruit diameter in the standard variety data; when determining whether there is an environmental disturbance, a temperature record is obtained, and the retrieval length of the temperature record is proportional to the current maximum temperature difference at the shooting time. The standard deviation of all temperature values ​​in the temperature record is calculated, and when the standard deviation is greater than or equal to the reference standard deviation, it is determined that there is an environmental disturbance.

[0114] It should be noted that the degree of deviation is determined by comparing the fruit diameter measured in the kiwifruit image with the standard fruit diameter for that variety. Specifically, the system first extracts the fruit diameter parameter from the image and then calculates the ratio with the standard diameter for that variety preset in the database to determine the degree of deviation, which is used to assess whether the fruit size conforms to the variety's characteristics. When determining whether there is environmental disturbance, the system retrieves temperature records related to the time the kiwifruit image was captured. The length of the retrieved temperature data is proportional to the maximum temperature fluctuation at the time of the capture. That is, the greater the temperature fluctuation, the longer the time span of temperature data retrieved. The system then performs statistical analysis on these temperature records to calculate the degree of temperature fluctuation, using the standard deviation of all retrieved temperature data as a fluctuation indicator. If this temperature standard deviation reaches or exceeds a pre-set reference threshold, the system determines that a disturbance is present in the environment, potentially affecting fruit growth or image acquisition accuracy, thus providing an important reference for subsequent data processing.

[0115] In some embodiments of the present application, when the degree of deviation is less than a preset deviation threshold and there is no environmental disturbance, the kiwi fruit image and parameter data are input into the upload module;

[0116] When the degree of deviation is greater than or equal to the deviation threshold, or when there is environmental disturbance, the kiwi fruit image and parameter data are returned.

[0117] It should be noted that after completing image feature comparison and anomaly analysis, the kiwifruit image and its parameter data are further evaluated for upload eligibility based on the two dimensions of "deviation degree" and "environmental disturbance." Specifically, if the fruit diameter extracted from the image differs slightly from the diameter in the standard variety database—that is, the deviation is below the preset deviation threshold—and there are no significant ambient temperature fluctuations before and after the image was captured, meaning no environmental disturbance was detected, the system determines that the image and its parameter data meet the standard requirements in terms of both physiological characteristics and acquisition conditions. Therefore, the image and its parameter data are directly transferred to the upload module for storage. Conversely, if the fruit diameter in the image deviates significantly from the set value for the standard variety, indicating possible cultivar confusion or abnormal individuals in the sample, or if there are significant temperature fluctuations before and after the capture time, indicating that the acquisition process may have been affected by environmental interference, affecting data reliability, the system will determine that the image and parameter data do not meet the storage requirements and will be returned without entering the database. Through this judgment process, the system effectively improves the authenticity and representativeness of the information in the kiwifruit specimen database.

[0118] It should be noted that the threshold setting method in this embodiment is determined based on statistical analysis and historical data collection. Specifically, the upper and lower thresholds of the image quality score are obtained by cluster analysis and density distribution analysis of the clarity, exposure, and edge feature values ​​of a large number of historically uploaded images. The transition interval between the pass and fail samples of the image quality score is used as the discrimination threshold. The field ratio threshold in the field integrity test is fitted by the normal distribution of the number of field completions in all historical collection tasks, and the 95% confidence interval boundary is set as the reasonable range of the field ratio. The matching degree division interval is constructed using the matching distance data between the standard image samples and the historically collected images. The boundary values ​​of the low, medium, and high matching degree intervals are set under the premise of ensuring that the false rejection rate and false acceptance rate are both at the lowest level. The fruit diameter deviation threshold is constructed by constructing a standard deviation model based on the fruit diameters of all varieties in the standard database, and the range of multiples of the standard deviation is used as the judgment limit. The judgment of temperature disturbance refers to the temperature data set of historical sampling points before and after the shooting time. The standard deviation is calculated and the mean of the overall temperature fluctuation sample set plus two standard deviations is used as the reference standard deviation threshold.

[0119] See Figure 2 As shown, an embodiment of the present invention provides a digital management method for kiwifruit resource collection information, comprising:

[0120] S1: Receive the kiwi fruit image and corresponding parameter data uploaded by the user;

[0121] S2: Calculate the image quality score of the kiwifruit image based on clarity, exposure, and edge feature values. Perform integrity check on the parameter data based on the ratio of the number of uploaded fields to the expected number of fields. Determine whether to construct an image feature roadmap based on the image quality score and integrity check results.

[0122] S3: When it is necessary to construct an image feature roadmap, the texture of the kiwifruit image and the features of the fruit crown area are extracted, and the image feature roadmap is constructed based on the deep image trajectory fitting algorithm. The matching degree between the image feature roadmap and the standard reference image is calculated, and the environmental and parameter anomaly analysis is determined based on the matching degree.

[0123] S4: When performing environmental and parameter anomaly analysis, it includes: comparing the degree of deviation of the fruit diameter parameter with the standard variety data; retrieving the temperature records before and after the kiwifruit image was taken, and judging whether there is environmental disturbance based on the temperature change rate;

[0124] S5: Determine whether the upload criteria are met based on the degree of deviation and whether there is environmental disturbance, and upload the kiwifruit image and parameter data to the database.

[0125] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A digital management system for kiwifruit resource collection information, characterized in that: include: An acquisition module is configured to receive kiwi fruit images and corresponding parameter data uploaded by users; The initial inspection and evaluation module is configured to calculate an image quality score for the kiwifruit image based on clarity, exposure, and edge feature values, perform integrity detection on the parameter data based on the ratio of the number of uploaded fields to the expected number of fields, and determine whether to input the kiwifruit image and parameter data into the feature comparison module based on the image quality score and integrity detection results; The feature comparison module is configured to extract the texture of the kiwifruit image and the features of the fruit crown area, construct an image feature path map based on a deep image trajectory fitting algorithm, calculate the matching degree between the image feature path and the standard reference map, and determine whether to perform environmental and parameter anomaly analysis based on the matching degree; The anomaly analysis module is configured to perform environmental and parameter anomaly analysis, including: calculating the degree of deviation of the fruit diameter parameter from the standard variety data; retrieving the temperature records before and after the kiwi fruit image was taken, and determining whether there is an environmental disturbance based on the temperature change rate; The uploading module is configured to determine whether the uploading standard is met based on the degree of deviation and whether there is environmental disturbance, and to upload the kiwi fruit image and parameter data to the database.

2. The digital management system for kiwifruit resource collection information according to claim 1, characterized in that: When calculating image quality scores based on sharpness, exposure, and edge feature values, these include: After graying the kiwifruit image, an edge response image is extracted by 3*3 Laplace operator convolution, and the variance of the edge response image is calculated. The variance of the edge response image and a preset variance value are used as the clarity; the preset variance value is the average variance of the edge response image of the historical kiwifruit image; Set a high brightness interval and a low brightness interval, and calculate the sum of the proportions of pixels in the kiwi fruit image located in the high brightness interval and the low brightness interval, where the sum of the proportions is the exposure; After extracting the edges in the kiwifruit image based on the Canny edge detection algorithm, the ratio of the number of edge points in the kiwifruit target area to the number of valid pixels in the kiwifruit image is calculated, and the ratio is the edge feature value.

3. The digital management system for kiwifruit resource collection information according to claim 2, characterized in that: When calculating image quality scores based on sharpness, exposure, and edge feature values, it also includes: The image quality score is the sum of the products of the clarity, exposure and edge feature values ​​and their corresponding weight coefficients.

4. The digital management system for kiwifruit resource collection information according to claim 3, characterized in that: The integrity detection includes: Match the parameter data with the field template, and calculate the ratio of the number of fields in the parameter data to the expected number of fields in the field template, which is recorded as the field ratio; Set format rules for each field of parameter data and use regular expressions to verify whether there are any missing fields; When the field ratio is 0.8-1.2 and there is no missing field, the integrity test result is determined to be passed.

5. The digital management system for kiwifruit resource collection information according to claim 4, characterized in that: When the image quality score is greater than or equal to the preset image score upper threshold and the integrity test result passes, the kiwi fruit image and parameter data are input into the upload module and output; When the image quality score is less than the preset image score upper threshold, greater than or equal to the preset image score lower threshold, and the integrity test result passes, the kiwi fruit image and parameter data are input into the feature comparison module; When the image quality score is lower than the preset image score lower threshold, or the integrity test result fails, the kiwi fruit image and parameter data will not be uploaded.

6. The digital management system for kiwifruit resource collection information according to claim 5, characterized in that: When calculating the matching degree based on the image feature route and the standard reference image, it includes: The target area of ​​kiwifruit images is segmented, and image blocks of the fruit crown area, surrounding texture area and background boundary area are extracted, and feature point detection and directional gradient calculation are performed respectively. Based on the SIFT algorithm, the feature descriptor of local key points is extracted for each image block, and a set of image partition feature vectors is constructed; Perform the same processing on the corresponding area in the standard reference image to obtain a set of standard image feature vectors; Each feature vector set of the kiwifruit image is spliced ​​according to the physical space order of the image blocks to form a two-dimensional feature trajectory matrix, and Bezier curve interpolation fitting is used to generate a continuous image feature path; Calculate the DTW distance between the feature path of the kiwi image and the feature path of the standard image; Map the DTW distance value to a matching degree in the range [0,1].

7. The digital management system for kiwifruit resource collection information according to claim 6, characterized in that: When determining whether to perform an environment and parameter anomaly analysis based on the matching degree, it includes: Divide the matching degree into low matching degree interval, medium matching degree interval and high matching degree interval from small to large; When the matching degree is in the high matching degree range, the kiwi fruit image and the corresponding parameter data are transmitted to the upload module; When the matching degree is in the medium matching degree range, the kiwi fruit image and the corresponding parameter data are transmitted to the abnormality analysis module; When the matching degree is in a low matching degree range, the kiwi fruit image and the corresponding parameter data are not transmitted to the abnormality analysis module or the upload module.

8. The digital management system for kiwifruit resource collection information according to claim 7, characterized in that: The degree of deviation is the ratio of the fruit diameter in the kiwifruit image to the fruit diameter in the standard variety data; when determining whether there is environmental disturbance, the temperature record is obtained, the retrieval length of the temperature record is proportional to the current maximum temperature difference at the shooting time, and the standard deviation of all temperature values ​​in the temperature record is calculated. When the standard deviation is greater than or equal to the reference standard deviation, it is determined that there is environmental disturbance.

9. The digital management system for kiwifruit resource collection information according to claim 8, characterized in that: When the degree of deviation is less than a preset deviation threshold and there is no environmental disturbance, the kiwi fruit image and parameter data are input into an uploading module; When the degree of deviation is greater than or equal to the deviation threshold, or when there is environmental disturbance, the kiwi fruit image and parameter data are returned.

10. A digital management method for kiwifruit resource collection information, characterized in that: The system according to any one of claims 1 to 9 comprises: Receive kiwi fruit images and corresponding parameter data uploaded by users; Calculate the image quality score of kiwifruit images based on clarity, exposure, and edge feature values. Perform integrity checks on parameter data based on the ratio of the number of uploaded fields to the expected number of fields. Determine whether to construct an image feature roadmap based on the image quality score and integrity test results. When it is necessary to construct an image feature roadmap, the texture of the kiwifruit image and the features of the fruit crown area are extracted, and the image feature roadmap is constructed based on the deep image trajectory fitting algorithm. The matching degree between the image feature roadmap and the standard reference image is calculated, and the decision on whether to perform environmental and parameter anomaly analysis is made based on the matching degree. When conducting environmental and parameter anomaly analysis, it includes: comparing the degree of deviation of the fruit diameter parameter with the standard variety data; retrieving the temperature records before and after the kiwifruit image was taken, and judging whether there is environmental disturbance based on the temperature change rate; Based on the degree of deviation and whether there is environmental disturbance, determine whether the upload criteria are met and upload the kiwifruit image and parameter data to the database.

Images