A fruit quality detection method and system based on fruit appearance parameters

Abstract

The present invention relates to the technical field of dimension measurement, and provides a fruit quality detection method and system based on fruit appearance parameters, including: obtaining a plurality of radius representative points and their circumferential structure points for the nectarine structure distribution; obtaining the nectarine contour probability according to the circumferential structure points; obtaining the corrected local outlier factor according to the radius of the radius representative points and the nectarine contour probability; obtaining the size representative degree according to the nectarine contour probability and the corrected local outlier factor; obtaining the overall radius of the nectarine according to the radius of the radius representative points and the size representative degree; and dividing the quality grade of the nectarine according to the overall radius of the nectarine to obtain the nectarine quality detection result. The present invention obtains the accurate overall radius of the nectarine, divides the quality grade of the nectarine, and obtains a more accurate nectarine quality detection result.

Description

Technical Field

[0001] The present invention relates to the technical field of dimension measurement, and particularly relates to a method and system for detecting fruit quality based on fruit appearance parameters. Background Art

[0002] In the detection of fruit quality, appearance parameters are important indicators for evaluating the quality of fruits. Generally, the analysis of appearance characteristics such as fruit color, shape, size, and texture can effectively evaluate the quality and market adaptability of fruits.

[0003] During the detection process of the appearance quality of nectarines, in order to shorten the overall detection time, maintain high production efficiency, and reduce manual operations, a batch of nectarines needs to be subjected to appearance detection at one time. As an important indicator, the size of nectarines directly affects the richness of its flesh and water content. Larger nectarines usually have richer pulp, higher sugar content, and better taste; while smaller fruits may be relatively shriveled and have a lighter taste. In addition, the appearance size of nectarines is also closely related to factors such as their growth environment and nutrient supply, reflecting the level of planting management. This makes it very intuitive and effective to conduct preliminary screening through size detection.

[0004] When detecting the size of nectarines, since the edge contour of nectarines is approximately circular, it is necessary to obtain the edge radius of the approximate circle of the edge contour of nectarines, so as to estimate the size of nectarines. Currently, the method of Hough circle detection is often used to extract the radius of nectarines. However, the representativeness of the extracted nectarine radius in the size of a batch of nectarines is different, that is, the circular contour presented by the nectarine radius is not necessarily the true contour of the nectarine, thereby reducing the accuracy of the fruit quality detection results. Summary of the Invention

[0005] The present invention provides a method and system for detecting fruit quality based on fruit appearance parameters to solve the problem that the representativeness of the nectarine radius extracted by the existing Hough circle detection is different, thereby reducing the accuracy of the fruit quality detection results. The specific technical solutions adopted are as follows:

[0006] The present invention proposes a method for detecting fruit quality based on fruit appearance parameters, and the method includes the following steps:

[0007] Obtain the nectarine structure distribution; obtain a number of radius representative points from the nectarine structure distribution; the radius representative points characterize a number of circular contours in the nectarine structure distribution, the radius representative points represent the position and radius size of the circular contours, and the circular contours are possible contour shapes of nectarines;

[0008] Obtain a number of circumferential structure points for each radius representative point; according to the color change and difference of the structure points near the corresponding position of each circumferential structure point in the nectarine structure distribution, obtain the nectarine contour probability of each radius representative point;

[0009] According to the radius of each radius representative point and the nectarine contour probability, the corrected local outlier factor of each radius representative point is obtained;

[0010] According to the nectarine contour probability and the corrected local outlier factor of each radius representative point, the degree of size representation of each radius representative point is obtained; according to the radii and the degrees of size representation of all radius representative points, the overall radius of the nectarine is obtained;

[0011] According to the overall radius of the nectarine, the quality grade of the nectarine is divided to obtain the nectarine quality detection result.

[0012] Furthermore, the specific method for obtaining the nectarine contour probability of each radius representative point according to the color change and difference of the structural points near the corresponding position of each circumferential structural point in the nectarine structure distribution includes:

[0013] For any circumferential structural point, obtain the characteristic change direction of this circumferential structural point; obtain several color structural points of the nectarine structure distribution in the RGB space, and record the color structural point corresponding to this circumferential structural point in the RGB space as the color reference point of this circumferential structural point. Denote the absolute value of the difference between the G-channel values of the two color structural points adjacent before and after this color reference point in the characteristic change direction of this circumferential structural point as the green mutation degree of this circumferential structural point; denote the absolute value of the difference between the R-channel values of the two color structural points adjacent before and after this color reference point in the characteristic change direction of this circumferential structural point as the red mutation degree of this circumferential structural point;

[0014] According to the difference in the red and green mutation degrees of the circumferential structural points, the nectarine contour probability of each radius representative point is obtained.

[0015] Furthermore, the specific method for obtaining the nectarine contour probability of each radius representative point according to the difference in the red and green mutation degrees of the circumferential structural points includes:

[0016]

[0017] In the formula, is the nectarine contour probability of the th radius representative point; is the number of circumferential structural points of the th radius representative point; is the red mutation degree of the th circumferential structural point of the th radius representative point; is the green mutation degree of the th circumferential structural point of the th radius representative point; represents a hyperparameter; Represents a linear normalization function.

[0018] Furthermore, obtaining the modified local outlier factor for each radius representative point according to the radius and the nectarine contour probability of each radius representative point includes the following specific method:

[0019] Using the LOF outlier detection algorithm to perform outlier detection on the radii of all radius representative points. During the detection process, the calculation method of the modified local outlier factor of the th radius representative point is:

[0020]

[0021] In the formula, is the modified local outlier factor of the th radius representative point at the distance neighborhood; is the initial parameter of the LOF outlier detection algorithm; represents is a radius representative point within the th radius representative point's distance neighborhood; is the nectarine contour probability of the radius representative point ; is the nectarine contour probability of the radius representative point ; -local reachability density, specifically obtained by the LOF outlier detection algorithm; is the nectarine contour probability of the th radius representative point; is the th radius representative point's -local reachability density; is the weight normalization function.

[0022] Furthermore, obtaining the size representativeness degree of each radius representative point according to the nectarine contour probability and the modified local outlier factor of each radius representative point includes the following specific method:

[0023]

[0024] In the formula, is the size representativeness degree of the th radius representative point; is the nectarine contour probability of the th radius representative point; is the modified local outlier factor of the th radius representative point at the distance neighborhood; is the linear normalization function.

[0025] Further, obtaining the overall radius of the nectarine according to the radii and size representation degrees of all radius representative points includes the following specific method:

[0026]

[0027] In the formula, represents the overall radius of the nectarine; is the number of radius representative points; is the th size representation degree of the radius representative point; is the th radius of the radius representative point; is the weight normalization function.

[0028] Further, obtaining several radius representative points from the nectarine structure distribution includes the following specific method:

[0029] Extract several nectarine contours from the nectarine structure distribution, input the several nectarine contours into the Hough circle transform, and obtain several curves in the Hough space;

[0030] Mark the intersection points of more than curves in the Hough space as a radius representative point, where is the preset intersection threshold; the abscissa of the radius representative point in the Hough space is the abscissa of the center of the circle, the ordinate is the ordinate of the center of the circle, and the vertical coordinate is the radius.

[0031] Further, obtaining several circumferential structure points for each radius representative point includes the following specific method:

[0032] For any curve passing through any radius representative point in the Hough space, mark the several structure points corresponding to the curve in the nectarine structure distribution as the several circumferential structure points of the radius representative point.

[0033] Further, classifying the nectarines according to the overall radius of the nectarine to obtain the nectarine quality detection result includes the following specific method:

[0034] Classify the nectarines with as special grade; classify the nectarines with as first grade; classify the nectarines with as second grade; where is the preset special grade standard size, is the preset first grade standard size, and and is the overall radius of the nectarine.

[0035] The present invention also provides a fruit quality detection system based on fruit appearance parameters. The system includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the steps of the above method are implemented.

[0036] The beneficial effects of the present invention are as follows: For nectarines of the same batch, several radius representative points are extracted through their possible circular contours. Some of the radius representative points may be caused by the red-green boundary line on the surface of the nectarine. The present invention constructs the green mutation degree and red mutation degree of the circular structure points of the radius representative points through the changes in the R-channel value and G-channel value at the position of the red-green boundary line, and judges the possibility that the radius representative point corresponds to the true contour of the nectarine. Since there are many structure points reflecting the contour in the nectarine structure distribution, the circular contour constructed by the radius representative points may not be the nectarine contour. The present invention calculates the corrected local outlier factor of each radius representative point by combining the property that the sizes of nectarines in the same batch are similar and the nectarine contour probability of the radius representative point, and judges the abnormal situation of the radius corresponding to each radius representative point. By utilizing the different representativeness of the radius representative points in the sizes of nectarines in the batch, the present invention obtains the size representativeness degree of each radius representative point through the nectarine contour probability and the corrected local outlier factor of the radius representative point, and then obtains the overall radius of the nectarine, so as to detect the quality of the whole batch of nectarines. The present invention divides the quality grades of nectarines by obtaining the accurate overall radius of the nectarines, and obtains a more accurate fruit quality detection result. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or in the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0038] Figure 1 It is a schematic flowchart of a fruit quality detection method based on fruit appearance parameters provided by an embodiment of the present invention;

[0039] Figure 2 It is an RGB channel separation diagram of the structure distribution of a nectarine provided by this embodiment. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0040] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0041] Please refer to Figure 1 , which shows a flowchart of a fruit quality detection method based on fruit appearance parameters provided by an embodiment of the present invention. The method includes the following steps:

[0042] Step S001: Obtain the nectarine structure distribution; obtain several radius representative points for the nectarine structure distribution.

[0043] It should be noted that the purpose of this embodiment is to analyze the circular contour of nectarines in the same batch. Since the collected nectarine structure distribution contains a circular contour composed of many structure points, it is necessary to screen the circular contour that can truly reflect the nectarine contour. Therefore, it is necessary to first extract several circular contours from the nectarine structure distribution through Hough circle detection.

[0044] Specifically, a light-colored conveyor belt is arranged on the fruit quality detection production line, and a batch of nectarines to be detected are laid flat on the conveyor belt. During the laying process, avoid overlapping of nectarines. An industrial camera is arranged directly above the conveyor belt, so that the lens of the industrial camera is aligned with the conveyor belt. During the conveying process of the conveyor belt, use the industrial camera to collect nectarine images as the nectarine structure distribution; and extract several nectarine contours from the nectarine structure distribution, and input the several nectarine contours into the Hough circle transformation to obtain several curves in the Hough space;

[0045] The intersection points of more than curves in the Hough space are recorded as a radius representative point, where is a preset intersection threshold, and this embodiment uses as an example for description; the abscissa of the radius representative point in the Hough space is the abscissa of the center of the circle, the ordinate is the ordinate of the center of the circle, and the vertical coordinate is the radius; for any curve passing through any radius representative point in the Hough space, the several structure points corresponding to the curve in the nectarine structure distribution are recorded as the several circumferential structure points of the radius representative point.

[0046] Optionally, in an embodiment of the present invention, for the extraction of the nectarine contour from the nectarine structure distribution, in this embodiment, the nectarine structure distribution is processed, that is, the nectarine image is grayscaled to obtain a nectarine grayscale image; since the Hough circle detection is sensitive to noise, the nectarine grayscale image is median filtered to obtain a nectarine filtered image; the nectarine filtered image is subjected to Canny edge detection to obtain a number of edge lines of the nectarine filtered image, and the edge lines are the number of nectarine contours in the nectarine structure distribution; among them, grayscaling, median filtering, Canny edge detection, and Hough circle transformation are well-known technologies, and the specific methods will not be introduced here.

[0047] It should be noted that since the nectarine contour usually presents a nearly circular shape, most of the structural points on the nectarine contour will transform into curves in the Hough space that intersect at one point. Because there are a large number of structural points on the nectarine contour, there are also many curves that intersect at one point when the structural points on the real nectarine contour are transformed into the Hough space.

[0048] Step S002, obtain a number of circumferential structural points for each radius representative point; according to the color change and difference of the structural points near the corresponding position of each circumferential structural point in the nectarine structure distribution, obtain the nectarine contour probability of each radius representative point.

[0049] It should be noted that during the growth of nectarines, the surface of the fruit is irradiated by sunlight, and the ultraviolet and visible light in the sunlight have a direct impact on the pigment synthesis of the fruit peel. The surface of a mature nectarine usually gradually changes from green to red. However, during the growth of nectarines, some areas on the surface of the nectarine may be blocked by leaves, etc., reducing the synthesis of the fruit peel pigment in the blocked area, resulting in the color of the blocked area remaining green instead of red. When the nectarine is mature and harvested, there will be an obvious red-green boundary on the surface of the nectarine, and the boundary may also present an arc shape. During the process of obtaining the radius representative points using the Hough space, since the red-green boundary is relatively obvious, the red-green boundary may also be judged as a radius representative point. Therefore, it is necessary to distinguish the radius representative points caused by the red-green boundary on the surface of the nectarine.

[0050] It should be further noted that the red-green boundary is not obvious in the RGB channels of the nectarine structure distribution. The boundary effect of the red-green boundary in the R channel is not obvious, that is, the difference in the R channel values of the pixel points on both sides of the red-green boundary is not large, while the boundary effect in the G channel is more obvious, that is, the difference in the G channel values of the pixel points on both sides of the red-green boundary is relatively large. The separation effect of the red-green boundary in the RGB channels of the nectarine structure distribution is as Figure 2 shown. Therefore, the radius representative points caused by the red-green boundary are distinguished accordingly.

[0051] Specifically, for any curve passing through any radius representative point in the Hough space, the pixel points corresponding to the curve in the nectarine filtered image are recorded as a circumferential structure point of the radius representative point;

[0052] For any circumferential structure point, obtain the characteristic change direction of the circumferential structure point; obtain several color structure points of the nectarine structure distributed in the RGB space, and record the color structure point corresponding to the circumferential structure point in the RGB space as the color reference point of the circumferential structure point. Denote the absolute value of the difference between the G-channel values of the two adjacent color structure points of the color reference point in the characteristic change direction of the circumferential structure point as the green mutation degree of the circumferential structure point; denote the absolute value of the difference between the R-channel values of the two adjacent color structure points of the color reference point in the characteristic change direction of the circumferential structure point as the red mutation degree of the circumferential structure point.

[0053] Optionally, in an embodiment of the present invention, the Sobel operator is used to obtain the gradient direction of the circumferential structure point and used as the characteristic change direction of the circumferential structure point. Among them, using the Sobel operator to obtain the gradient direction of the pixel point is a well-known technology, and the specific method will not be introduced here; and several color structure points of the nectarine structure distributed in the RGB space are, in this embodiment, the RGB image of the nectarine image and the RGB three-channel values of each pixel point therein.

[0054] Further, the calculation method of the nectarine contour probability of the th radius representative point is:

[0055]

[0056] In the formula, is the nectarine contour probability of the th radius representative point; is the number of circumferential structure points of the th radius representative point; is the red mutation degree of the th circumferential structure point of the th radius representative point; is the green mutation degree of the th circumferential structure point of the th radius representative point; represents a hyperparameter; represents a linear normalization function, and the normalization object is of all radius representative points.

[0057] Step S003: Obtain the corrected local outlier factor of each radius representative point according to the radius and the nectarine contour probability of each radius representative point.

[0058] It should be noted that since the resolution of the nectarine structure distribution collected by the industrial camera is relatively high and the number of edge pixel points obtained by edge detection is large, there are a large number of curves in the Hough space, resulting in the possibility that the radius representative points formed by the intersection of the curves may not be the nectarine contour. Since the same variety and consistent management measures are adopted during the planting of the nectarines in the same batch, the sizes of the nectarines in the same batch are relatively similar, which is manifested as the vertical axis coordinates of the radius representative points being relatively concentrated in the Hough space, that is, the radii are relatively similar.

[0059] Specifically, the radii of all radius representative points are subjected to outlier detection using the LOF outlier detection algorithm. During the detection process, since some radius representative points belong to the red-green boundary, it is necessary to reduce the weights of these radius representative points during outlier detection. The calculation method of the corrected local outlier factor of the th radius representative point is as follows:

[0060]

[0061] In the formula, is the corrected local outlier factor of the th radius representative point at the distance neighborhood; is the initial parameter of the LOF outlier detection algorithm; represents is a radius representative point within the th radius representative point's distance neighborhood; is the nectarine contour probability of the radius representative point ; is the -local reachability density of the radius representative point , which is specifically obtained by the LOF outlier detection algorithm; is the nectarine contour probability of the th radius representative point; is the -local reachability density of the th radius representative point; is the weight normalization function, and the normalization object is the nectarine contour probabilities of all radius representative points within the th radius representative point's distance neighborhood.

[0062] It should be noted that the larger the value of , the more abnormal the radius of the th radius representative point, that is, the

[0063] Step S004: Obtain the degree of size representation of each radius representative point based on the nectarine contour probability and the corrected local outlier factor of each radius representative point; obtain the overall radius of the nectarine based on the radii and degrees of size representation of all radius representative points.

[0064] It should be noted that among all the radius representative points, the radii of some radius representative points are not the radii of the real nectarine contour edge. Therefore, when obtaining the radius sizes of the whole batch of nectarines, the weights of these radius representative points need to be reduced.

[0065] It should be further noted that when analyzing the weights of the radius representative points, mainly reduce the weights of the radius representative points with lower nectarine contour probabilities and the weights of the radius representative points with abnormal radii of the radius representative points.

[0066] Specifically, the calculation method of the degree of size representation of the th radius representative point is:

[0067]

[0068] In the formula, is the degree of size representation of the th radius representative point; is the nectarine contour probability of the th radius representative point; is the corrected local outlier factor of the th radius representative point at the distance neighborhood; is a linear normalization function, and the normalization object is the corrected local outlier factors of all radius representative points at the distance neighborhood.

[0069] It should be noted that the radius representative point with a larger degree of size representation can better represent the size of the whole batch of nectarines. Therefore, calculate the overall radius of the nectarine according to the degree of size representation of the radius representative points.

[0070] Specifically, the calculation method of the overall radius of the nectarine is:

[0071]

[0072] In the formula, represents the overall radius of the nectarine; is the number of radius representative points; is the th radius representative point's degree of size representation; is the th radius representative point's radius; is a weight normalization function, and the normalization object is the degree of size representation of all radius representative points.

[0073] Step S005: Classify the nectarines according to the overall radius of the nectarines to obtain the nectarine quality inspection results.

[0074] It should be noted that the size of nectarines is generally described by the diameter. Nectarines with a larger diameter have rich pulp and a higher sugar content. Therefore, the quality grades of nectarines are classified according to the size of the nectarines.

[0075] Specifically, nectarines with

[0076] are classified as special grade; nectarines with

[0077] are classified as first grade; nectarines with

[0078] Among them, is the preset special grade standard size, is the preset first grade standard size, and is the overall radius of the nectarine; This embodiment is described by taking this as an example.

[0079] It should be noted that so far, by classifying each batch of nectarines according to the overall radius of the nectarines, the inspection of the quality of nectarines is realized.

[0080] Another embodiment of the present invention provides a fruit quality inspection system based on fruit appearance parameters. The system includes a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, the above method steps S001 to S005 are realized.

[0081] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A fruit quality detection method based on fruit appearance parameters, characterized in that: The method comprises the following steps: Obtaining a nectarine structure distribution; obtaining a number of radius representative points for the nectarine structure distribution; the radius representative points represent a number of circular contours in the nectarine structure distribution, the radius representative points represent the position and radius size of the circular contour, and the circular contour is a possible contour shape of the nectarine; Obtain several circumferential structure points of each radius representative point; obtain the nectarine outline probability of each radius representative point according to the color change and difference of the structure points near the corresponding position of each circumferential structure point in the nectarine structure distribution; According to the radius of each radius representative point and the probability of the nectarine contour, the modified local anomaly factor of each radius representative point is obtained; According to the nectarine contour probability of each radius representative point and the corrected local anomaly factor, the size representative degree of each radius representative point is obtained; according to the radius and size representative degree of all radius representative points, the overall radius of the nectarine is obtained; The quality grades of nectarines are classified according to the overall radius of nectarines, and the quality test results of nectarines are obtained; The method of obtaining the nectarine contour probability of each radius representative point according to the color change and difference of the structure points near the corresponding position of each circumferential structure point in the nectarine structure distribution includes: For any circumferential structure point, obtain the characteristic change direction of the circumferential structure point; obtain a number of color structure points of the nectarine structure distribution in the RGB space, record the color structure point corresponding to the circumferential structure point in the RGB space as the color reference point of the circumferential structure point, and record the absolute value of the difference between the G channel values ​​of the two color structure points adjacent to each other in the characteristic change direction of the circumferential structure point as the green mutation degree of the circumferential structure point; record the absolute value of the difference between the R channel values ​​of the two color structure points adjacent to each other in the characteristic change direction of the circumferential structure point as the red mutation degree of the circumferential structure point; According to the difference in red-green mutation degree of the circular structure points, the probability of the nectarine outline at each radius representative point is obtained; The specific method of obtaining the nectarine contour probability of each radius representative point according to the difference in red-green mutation degree of the circumferential structure points is as follows: In the formula, For the The radius represents the contour probability of the nectarine at each point; For the The number of circular structure points represented by each radius; For the The radius represents the point The red mutation degree of each circular structure point; For the The radius represents the point The degree of green mutation of each circular structure point; represents a hyperparameter; represents the linear normalization function; The method of obtaining the corrected local anomaly factor of each radius representative point according to the radius of each radius representative point and the nectarine contour probability includes the following specific methods: The radius of all radius representative points is used for anomaly detection using the LOF anomaly detection algorithm. During the detection process, The calculation method of the corrected local anomaly factor of the representative point of radius is: In the formula, For the The radius represents the point Corrected local anomaly factor when away from the neighborhood; is the initial parameter of the LOF anomaly detection algorithm; express For the The radius represents the point A radius representative point within the distance neighborhood; The radius represents the point The probability of the nectarine outline; The radius represents the point of - Local reachability density, obtained by the LOF anomaly detection algorithm; For the The radius represents the contour probability of the nectarine at each point; For the The radius represents the point - local reachability density; is the weight normalization function; The specific method of obtaining the size representative degree of each radius representative point according to the nectarine contour probability of each radius representative point and the corrected local anomaly factor is as follows: In the formula, For the The radius represents the size of the point, representing the degree; For the The radius represents the contour probability of the nectarine at each point; For the The radius represents the point Corrected local anomaly factor when away from the neighborhood; is a linear normalization function; The method of obtaining the overall radius of the nectarine according to the radius and size representative degree of all radius representative points includes: In the formula, represents the overall radius of the nectarine; is the number of points representing the radius; For the The radius represents the size of the point, representing the degree; For the The radius represents the radius of the point; is the weight normalization function.

2. A method for detecting fruit quality based on fruit appearance parameters according to claim 1, characterized in that: The specific method of obtaining a plurality of radius representative points for the nectarine structure distribution includes: Extracting a number of nectarine contours from the nectarine structure distribution, inputting the several nectarine contours into the Hough circle transform, and obtaining a number of curves in the Hough space; The Hough space exceeds The intersection of the curves is recorded as a radius representative point, where is a preset intersection threshold; the abscissa of the radius representative point in the Hough space is the abscissa of the center of the circle, the ordinate is the ordinate of the center of the circle, and the ordinate is the radius.

3. A method for detecting fruit quality based on fruit appearance parameters according to claim 2, characterized in that: The specific method of obtaining a number of circular structure points of each radius representative point includes: In Hough space, for any curve passing through any radius representative point, several structural points corresponding to the curve in the nectarine structure distribution are recorded as several circular structural points of the radius representative point.

4. The method for detecting fruit quality based on fruit appearance parameters according to claim 1, characterized in that: The method of classifying the quality of nectarines according to the overall radius of the nectarines to obtain the nectarine quality test results includes the following specific methods: Will of nectarines are classified as special grade; The nectarines are classified into the first grade; Nectarines are divided into two levels; among them, For the preset super standard size, is a preset first-level standard size, and , is the overall radius of the nectarine.

5. A fruit quality detection system based on fruit appearance parameters, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of a fruit quality detection method based on fruit appearance parameters as described in any one of claims 1 to 4 are implemented.

Images