A method for measuring the appearance and dimensions of sweet cherries based on image recognition

Abstract

The present application discloses a method for measuring the external dimensions of sweet cherries based on image recognition, which relates to the field of measurement technology, including: obtaining a top view and a side view of the sweet cherry through a sweet cherry external image acquisition device; preprocessing the top view and the side view to remove useless information and noise interference; determining a calibration template of the sweet cherry in the top view and the side view, the calibration template being used to determine pixel equivalents; combining the pixel equivalents with the two points in the top view and the side view that are the farthest apart in the horizontal, vertical and vertical directions of the sweet cherry contour to determine the horizontal diameter, longitudinal length and thickness information of the sweet cherry. The accurate horizontal diameter, longitudinal length and thickness information of the sweet cherry can be determined based on the number of pixels between the sweet cherry in the horizontal, vertical and vertical directions in the top view and the side view combined with the pixel equivalents determined by the calibration template. In this way, the size of the sweet cherry can be accurately measured regardless of its size, which has far-reaching significance for sweet cherry breeding.

Description

Technical Field

[0001] The present application relates to the field of measurement technology, and in particular to a method for measuring the external dimensions of sweet cherries based on image recognition. Background Art

[0002] In the sweet cherry breeding work, it is necessary to collect phenotypes of a large number of sweet cherry varieties. Among them, the size index of sweet cherry is its important phenotypic information. The size index mainly includes three parameters: horizontal diameter, vertical diameter and thickness.

[0003] At present, the measurement of the size of sweet cherries is mainly carried out manually, such as Chinese patent CN216523571U, a portable cherry diameter multi-size measuring device, which is provided with an upper cover plate and a lower cover plate of exactly the same size and shape, the upper cover plate and the lower cover plate are arranged in parallel, and a number of teardrop-shaped measuring plates are arranged between the upper cover plate and the lower cover plate, and the measuring plates are provided with measuring holes, wherein different measuring holes have different sizes. Another example is Chinese patent CN203595479U, a cherry specification card, comprising: a card body, on which two rows of through holes for measuring the diameter of the cherry to be measured are opened, and the diameter of each row of through holes decreases successively. By opening through holes that can measure different cherry diameters on the card body, the user can directly pinch the cherry stem with his fingers and vertically insert the cherry into the through hole that is roughly closest to the diameter of the cherry to be measured, so that the fruit specification size can be quickly determined.

[0004] Although the above two patents can both measure the size of cherries to a certain extent, since the number of measuring holes in the above two measuring devices is limited and the measuring dimensions are fixed, they can only measure whether the size is larger or smaller than the through holes, and cannot accurately obtain the size of the cherries. Summary of the invention

[0005] In order to solve the above technical problems, this application proposes the following technical solutions:

[0006] In a first aspect, an embodiment of the present application provides a method for measuring the external dimensions of sweet cherries based on image recognition, comprising:

[0007] A top view and a side view of the sweet cherry are obtained by a sweet cherry appearance image obtaining device;

[0008] Preprocessing the top view and the side view to remove useless information and noise interference;

[0009] Determine a calibration template of the sweet cherry in the top view and the side view, wherein the calibration template is used to determine a pixel equivalent;

[0010] The horizontal diameter, longitudinal length and thickness information of the sweet cherry are determined by combining the pixel equivalent with the two points of the sweet cherry outline in the top view and the side view that are farthest apart in the horizontal, longitudinal and vertical directions.

[0011] In a possible implementation, the sweet cherry appearance image acquisition device includes: a main body support, the main body support is a semi-closed structure, a transparent V-shaped clamp is arranged in the main body support, the transparent V-shaped clamp is inclined on both sides and smooth in the middle, the middle smooth section of the transparent V-shaped clamp is provided with a sweet cherry fixing mechanism, the top and one side of the main body support are respectively provided with a first camera and a second camera, the center line of the shooting end of the first camera passes through the center of the cherry fixing position of the sweet cherry fixing mechanism, and the center line of the shooting end of the second camera passes through one side of the transparent V-shaped clamp and intersects vertically with the center line of the shooting end of the first camera.

[0012] In a possible implementation, the sweet cherry fixing mechanism includes a first baffle and a second baffle, and the first baffle and the second baffle are rotatably connected by a forward and reverse screw rod. When the forward and reverse screw rod rotates counterclockwise, the first baffle and the second baffle move outward at the same time, and the distance between the first baffle and the second baffle increases. When the forward and reverse screw rod rotates clockwise, the first baffle and the second baffle move inward at the same time, and the distance between the first baffle and the second baffle decreases.

[0013] In a possible implementation, a first clamp link and a second clamp link are arranged in parallel in the main body bracket, and both ends of the first clamp link and the second clamp link are respectively fixedly connected to the two sides of the main body bracket, and a first LED lamp and a second LED lamp are respectively arranged on the inner walls of the two sides of the main body bracket.

[0014] In a possible implementation, a sensing device is provided on one side of the transparent V-shaped clamp, and the sensing device includes a first fixed rod, a first end of the first fixed rod is fixedly connected to the bottom plate of the main bracket, a second end of the first fixed rod is fixedly provided with a fixed support, a sensor support connecting rod is provided on the fixed support, one end of the sensor support is sleeved on the sensor support connecting rod, and a photoelectric sensor is fixedly provided on the sensor support.

[0015] In a possible implementation, the step of acquiring a top view and a side view of the sweet cherry by using a sweet cherry appearance image acquisition device includes:

[0016] When the user places the sweet cherries in the cherry fixing position of the transparent V-shaped clamp, the first baffle plate and the second baffle plate are adjusted by rotating the forward and reverse screw rods to just clamp the sweet cherries;

[0017] After the photoelectric sensor senses the presence of cherries in the transparent V-shaped clamp, it controls the first LED light and the second LED light on both sides of the main support to turn on respectively, and then starts the first camera and the second camera to collect images to obtain the top view and side view of the sweet cherries.

[0018] In a possible implementation, preprocessing the top view and the side view to remove useless information and noise interference includes:

[0019] First, the image is grayed out. The image graying process uses the weighted average method. The formula is as follows: , where Gray(i,j) is the gray value at the coordinate (i,j) after transformation, R(i,j) is the red component of the color image at the coordinate (i,j) before transformation, G(i,j) is the green component of the color image at the coordinate (i,j) before transformation, and B(i,j) is the blue component of the color image at the coordinate (i,j) before transformation;

[0020] Then the median filtering method is used to remove the salt and pepper noise generated during the transmission process;

[0021] Finally, the closing operation in the geometric morphology method is used to remove the tiny impurities in the image.

[0022] In a possible implementation, the finally using a closing operation in a geometric morphology method to remove tiny impurities in the image includes:

[0023] First, perform the dilation operation: ;

[0024] Then perform the erosion operation: ;

[0025] Finally, the closed operation formula is: ;

[0026] Among them, A and B are two image sets, x is the distance B moves, is the translation of set B, is a mapping of set B.

[0027] In a possible implementation, determining a calibration template of the sweet cherry in the top view and the side view, wherein the calibration template is used to determine a pixel equivalent, includes:

[0028] Determine the ROI region of interest of the sweet cherry from the top view and the side view respectively;

[0029] The side view uses the same calibration template to calibrate the ROI area;

[0030] The top view is matched with different matching templates in the ROI area according to different sizes of the sweet cherries to obtain geometric matching scores of the sweet cherries, and the geometric matching template with the highest score is obtained by comparison according to the geometric matching scores and used as the calibration template for calibration;

[0031] The distance between two points in pixel coordinates is determined by the number of pixels between the two points. The pixel equivalent, that is, the distance of a unit pixel in the real world, is determined through calibration, thereby establishing a connection between the number of pixels between the two points and the distance between the two points in the real world.

[0032] In a possible implementation, the top view is matched with different matching templates in the ROI area according to different sizes of the sweet cherries to obtain a geometric matching score of the sweet cherries, and the geometric matching template with the highest score is obtained by comparison according to the geometric matching scores as a calibration template for calibration, including: establishing a conversion between real-world coordinates and pixel coordinates by the following formula:

[0033]

[0034]

[0035] Where: ( , ) is the world coordinate, ( ) is the image coordinate, ( , ) is the pixel coordinate, is the rotation matrix, is the offset vector, is the camera intrinsic parameter matrix, is the perspective photography matrix, is the scale factor.

[0036] In the embodiment of the present application, the top view and side view of the sweet cherry are obtained by an image acquisition device, and then the accurate horizontal diameter, longitudinal diameter and thickness information of the sweet cherry can be determined according to the number of pixels in the horizontal, longitudinal and vertical directions of the sweet cherry in the top view and the side view combined with the pixels determined by the calibration template. In this way, the size of the sweet cherry can be accurately measured regardless of the size of the sweet cherry, which has far-reaching significance for sweet cherry breeding. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A schematic diagram of a process for measuring the external dimensions of sweet cherries based on image recognition provided in an embodiment of the present application;

[0038] Figure 2 A schematic diagram of the structure of a device for acquiring a sweet cherry appearance image provided in an embodiment of the present application;

[0039] Figure 3A schematic diagram of the structure of a transparent V-shaped clamp provided in an embodiment of the present application;

[0040] Figure 4 A side view of a device for acquiring a sweet cherry appearance image provided in an embodiment of the present application;

[0041] Figure 5 A schematic diagram of the structure of the sensing device provided in the embodiment of the present application;

[0042] Figure 6 The geometric matching results of the sweet cherry image and other fruit images provided in the embodiments of the present application;

[0043] Figure 7 Schematic diagram of using different calibration templates for top views of sweet cherries of different sizes provided in the embodiment of the present application;

[0044] Figure 1-Figure 7 In the symbol, it is represented as:

[0045] 1-main body bracket, 2-transparent V-shaped clamp, 3-first camera, 4-second camera, 5-first baffle, 6-second baffle, 7-positive and negative screw rods, 8-first clamp connecting rod, 9-second clamp connecting rod, 10-first LED light, 11-second LED light, 12-sensing device, 13-first fixing rod, 14-fixing support, 15-sensor support connecting rod, 16-sensor support, 17-photoelectric sensor. DETAILED DESCRIPTION

[0046] The present solution is described below in conjunction with the accompanying drawings and specific implementation methods.

[0047] See also Figure 1 The method for measuring the external dimensions of sweet cherries based on image recognition provided in the embodiment of the present application includes:

[0048] S101, obtaining a top view and a side view of the sweet cherry by a sweet cherry appearance image obtaining device.

[0049] See also Figure 2, the sweet cherry appearance image acquisition device for acquiring the top view and side view of sweet cherries in the embodiment of the present application comprises: a main body support 1, the main body support 1 is a semi-enclosed structure, a transparent V-shaped clamp 2 is arranged in the main body support 1, the transparent V-shaped clamp 2 is inclined on both sides and smooth in the middle, a sweet cherry fixing mechanism is arranged in the middle smooth section of the transparent V-shaped clamp 2, a first camera 3 and a second camera 4 are arranged on the top and one side of the main body support 1 respectively, the shooting end center line of the first camera 3 passes through the center of the cherry fixing position of the sweet cherry fixing mechanism, and the shooting end center line of the second camera 4 passes through one side of the transparent V-shaped clamp 2 and intersects perpendicularly with the shooting end center line of the first camera 3. In this embodiment, the first camera 3 and the second camera 4 are both high-definition distortion-free cameras with wireless transmission function, and the sweet cherry top view taken by the first camera 3 and the sweet cherry side view taken by the second camera 4 are wirelessly transmitted to the processor.

[0050] See also Figure 3 The sweet cherry fixing mechanism includes a first baffle 5 and a second baffle 6, which are rotatably connected by a positive and negative screw 7. When the positive and negative screw 7 rotates counterclockwise, the first baffle 5 and the second baffle 6 move outward at the same time, and the distance between the first baffle 5 and the second baffle 6 increases. When the positive and negative screw 7 rotates clockwise, the first baffle 5 and the second baffle 6 move inward at the same time, and the distance between the first baffle 5 and the second baffle 6 decreases. When placing a larger sweet cherry, the two sides of the sweet cherry just form a triangular structure with the inclined sides of the transparent V-shaped clamp 2, so that the sweet cherry can be placed stably. When placing a smaller sweet cherry, the position of the first baffle 5 and the second baffle 6 can be adjusted by rotating the positive and negative screw 7, thereby fixing the smaller sweet cherry.

[0051] See also Figure 4 A first clamp link 8 and a second clamp link 9 are arranged in parallel in the main body bracket 1. The two ends of the first clamp link 8 and the second clamp link 9 are fixedly connected to the two sides of the main body bracket 1 respectively. A first LED lamp 10 and a second LED lamp 11 are arranged on the inner walls of the two sides of the main body bracket 1 respectively. The arrangement of the two LED lamps ensures good lighting conditions when taking images.

[0052] The transparent V-shaped clamp 2 is provided with a sensing device 12 on one side. Figure 5The sensing device 12 includes a first fixing rod 13, a first end of which is fixedly connected to the bottom plate of the main bracket 1, a second end of which is fixedly provided with a fixing support 14, a sensor support connecting rod 15 is provided on the fixing support 14, one end of a sensor support 16 is sleeved on the sensor support connecting rod 15, and a photoelectric sensor 17 is fixedly provided on the sensor support 16.

[0053] When the user places the sweet cherry in the cherry fixing position of the transparent V-shaped clamp 2, the first baffle 5 and the second baffle 6 are adjusted by rotating the positive and negative screw rods 7 to clamp the sweet cherry. After the photoelectric sensor 17 senses the presence of the cherry in the transparent V-shaped clamp 2, the first LED light 10 and the second LED light 11 on both sides of the main frame 1 are controlled to turn on, and then the first camera 3 and the second camera 4 are activated to collect images to obtain the top view and the side view of the sweet cherry.

[0054] There are a large number of fruits with similar sizes to sweet cherries, but sweet cherries have their own inherent characteristics. Most sweet cherries are heart-shaped, that is, the fruit is concave at the petiole, wide at both ends, and flat or pointed at the top. Based on this geometric shape, after the image is collected in this embodiment, geometric matching is performed on the collected sweet cherry top view, and the geometric matching includes two stages: learning and matching.

[0055] In the learning stage, the system learns and saves the geometric feature information of the sweet cherry template image by manually setting the sweet cherry template image. In the matching stage, the system extracts the geometric feature information of the collected image and matches it with the geometric feature information of the sweet cherry template image, and assigns a score according to the similarity between the geometric feature information of the collected image and the geometric feature information of the sweet cherry template image, thereby determining whether the collected image is a sweet cherry image.

[0056] See also Figure 6 In this embodiment, two sweet cherries are placed at different angles, and cherry tomatoes and sweet grapes are placed in sequence. The captured images are obtained through geometric matching. Figure 6 (a) and Figure 6 Middle (b) shows two sweet cherries placed at different angles. Figure 6 (c) in the middle is cherry tomatoes. Figure 6 In the middle (d) is sweet grape. Figure 6 (a) and Figure 6 (b) all match successfully, however Figure 6 (c) and Figure 6 There is no match in (d). There is basically no feature similar to the sweet cherry template image in the cherry tomato. Although there is some similarity in the angles of the sweet grapes, there is no significant feature of the heart-shaped fruit and the concave petiole, so the match also fails.

[0057] S102, pre-processing the top view and the side view to remove useless information and noise interference.

[0058] In this embodiment, in order to remove useless information and noise interference in the top view and side view of the sweet cherry obtained by shooting, image graying, image median filtering and image morphology are used for processing.

[0059] Specifically, the image is first grayed out, and the image graying process adopts the weighted average method, and the formula is as follows: , where Gray(i,j) is the gray value at coordinate (i,j) after transformation, R(i,j) is the red component of the color image at coordinate (i,j) before transformation, G(i,j) is the green component of the color image at coordinate (i,j) before transformation, and B(i,j) is the blue component of the color image at coordinate (i,j) before transformation. Then the median filtering method is used to remove the salt and pepper noise generated during the transmission process, and finally the closing operation in the geometric morphology method is used to remove the fine impurities in the image.

[0060] In this embodiment, a closing operation in a geometric morphology method is used to remove fine impurities in the image, including: firstly performing a dilation operation: ; Then perform the erosion operation: ;Finally, perform the closed operation formula: ; Where A and B are two image sets, x is the distance B moves, is the translation of set B; is a mapping of set B.

[0061] S103, determining a calibration template of the sweet cherry in the top view and the side view, wherein the calibration template is used to determine a pixel equivalent.

[0062] Due to the structural characteristics of the transparent V-shaped clamp, when sweet cherries of different sizes are placed on the clamp, the distance from the center plane of the side view sweet cherries to the camera is consistent, but the distance from the center plane of the top view to the camera will be affected by the size of the sweet cherries. According to the principle of perspective, it will cause the phenomenon of large objects appearing closer and small objects appearing farther away. Therefore, the side view of the sweet cherry can be calibrated with the same calibration template, while the top view of the sweet cherry needs to set different calibration templates according to the distance from the center plane to the camera.

[0063] In this embodiment, the mature and high-precision YOLO detection module is used to detect the position information of the sweet cherry in the image, and the ROI area is delineated based on this information. Specifically, the ROI area of ​​interest of the sweet cherry is determined from the top view and the side view respectively, and the ROI area is calibrated using the same calibration template in the side view. Figure 7The top view is matched with different matching templates in the ROI area according to the different sizes of the sweet cherry to obtain the geometric matching score of the sweet cherry, and the geometric matching template with the highest score is compared according to the geometric matching score to be calibrated as the calibration template. After obtaining the pixel coordinates, the distance between the two specific points can be determined by the number of pixels between the two points. The system determines the pixel equivalent through calibration, that is, the distance of the unit pixel in the real world, so as to establish a connection between the number of pixels between the two points and the distance between the two points in the real world.

[0064] In this implementation, the calibration template is determined by establishing the conversion between real world coordinates and pixel coordinates through the following formula:

[0065]

[0066]

[0067] Where: ( , ) is the world coordinate, ( ) is the image coordinate, ( , ) is the pixel coordinate, is the rotation matrix, is the offset vector, is the camera intrinsic parameter matrix, is the perspective photography matrix, is the scale factor.

[0068] S104, combining the pixel equivalent with the two points of the sweet cherry outline in the top view and the side view that are farthest apart in the horizontal, longitudinal and vertical directions to determine the transverse diameter, longitudinal length and thickness information of the sweet cherry.

[0069] The outline of the cherry at this viewing angle is obtained through the brightness information in the top view. Due to the fixation of the transparent V-shaped clamp, the posture of each cherry is determined and the same. The two points with the farthest distance between the cherry outline in this direction are determined by horizontal scanning to obtain the two points containing the horizontal diameter size information of the cherry. The number of pixels between the two points is calculated, and the horizontal diameter size information can be obtained through pixel equivalent. The outline of the cherry at this viewing angle is obtained through the brightness information in the side view. The posture of each cherry is determined and the same. The two points with the farthest distance between the cherry outline in these two directions are determined by horizontal and vertical scanning respectively. These two groups of points contain longitudinal diameter information and thickness information. The longitudinal diameter and thickness of the cherry are obtained by calibrating the determined pixel equivalent. The horizontal diameter, longitudinal diameter and thickness information of the cherry obtained by the image processing and measurement system will be saved in an Excel spreadsheet for the analysis of subsequent breeding work.

[0070] In the embodiments of the present application, "at least one" refers to one or more, and "plurality" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent the existence of A alone, the existence of A and B at the same time, and the existence of B alone. Among them, A and B can be singular or plural. The character " / " generally indicates that the previous and subsequent associated objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c can be represented by: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or multiple.

[0071] The above is only a specific implementation of the present application. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. The protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for measuring the external dimensions of sweet cherries based on image recognition, characterized in that: include: A top view and a side view of the sweet cherry are obtained by a sweet cherry appearance image obtaining device; The sweet cherry appearance image acquisition device comprises: a main body support, the main body support is a semi-enclosed structure, a transparent V-shaped clamp is arranged in the main body support, the transparent V-shaped clamp is inclined at both sides and smooth in the middle, a sweet cherry fixing mechanism is arranged in the middle smooth section of the transparent V-shaped clamp, a first camera and a second camera are arranged on the top and one side of the main body support respectively, the shooting end center line of the first camera passes through the center of the cherry fixing position of the sweet cherry fixing mechanism, and the shooting end center line of the second camera passes through one side of the transparent V-shaped clamp and intersects perpendicularly with the shooting end center line of the first camera; The sweet cherry fixing mechanism comprises a first baffle and a second baffle, wherein the first baffle and the second baffle are rotatably connected via a forward and reverse screw rod, and when the forward and reverse screw rod rotates counterclockwise, the first baffle and the second baffle move outward at the same time, and the distance between the first baffle and the second baffle increases; when the forward and reverse screw rod rotates clockwise, the first baffle and the second baffle move inward at the same time, and the distance between the first baffle and the second baffle decreases; A first clamp connecting rod and a second clamp connecting rod are arranged in parallel in the main body bracket, and two ends of the first clamp connecting rod and the second clamp connecting rod are respectively fixedly connected to two sides of the main body bracket, and a first LED lamp and a second LED lamp are respectively arranged on the inner walls of both sides of the main body bracket; A sensing device is provided on one side of the transparent V-shaped clamp, and the sensing device includes a first fixing rod, a first end of the first fixing rod is fixedly connected to the bottom plate of the main frame, a second end of the first fixing rod is fixedly provided with a fixing support, a sensor support connecting rod is provided on the fixing support, one end of the sensor support is sleeved on the sensor support connecting rod, and a photoelectric sensor is fixedly provided on the sensor support; The method of obtaining a top view and a side view of the sweet cherry by using a sweet cherry appearance image obtaining device comprises: When the user places the sweet cherries in the cherry fixing position of the transparent V-shaped clamp, the first baffle plate and the second baffle plate are adjusted by rotating the forward and reverse screw rods to just clamp the sweet cherries; After the photoelectric sensor senses that there are cherries in the transparent V-shaped clamp, the first LED light and the second LED light on both sides of the main support are controlled to turn on respectively, and then the first camera and the second camera are activated to collect images to obtain a top view and a side view of the sweet cherries; Preprocessing the top view and the side view to remove useless information and noise interference; Determining a calibration template of the sweet cherry in the top view and the side view, wherein the calibration template is used to determine a pixel equivalent, including: Determine the ROI region of interest of the sweet cherry from the top view and the side view respectively; The side view uses the same calibration template to calibrate the ROI area; The top view is matched with different matching templates in the ROI area according to different sizes of the sweet cherries to obtain geometric matching scores of the sweet cherries, and the geometric matching template with the highest score is obtained by comparison according to the geometric matching scores and used as the calibration template for calibration; Determine the distance between two points in pixel coordinates by using the number of pixels between the two points. The pixel equivalent, that is, the distance of a unit pixel in the real world, is determined by calibration, thereby establishing a connection between the number of pixels between the two points and the distance between the two points in the real world. The horizontal diameter, longitudinal length and thickness information of the sweet cherry are determined by combining the pixel equivalent with the two points of the sweet cherry outline in the top view and the side view that are farthest apart in the horizontal, longitudinal and vertical directions.

2. The method for measuring the external dimensions of sweet cherries based on image recognition according to claim 1, characterized in that: The preprocessing of the top view and the side view to remove useless information and noise interference includes: First, the image is grayed out. The image graying process uses the weighted average method. The formula is as follows: , where Gray(i,j) is the gray value at the coordinate (i,j) after transformation, R(i,j) is the red component of the color image at the coordinate (i,j) before transformation, G(i,j) is the green component of the color image at the coordinate (i,j) before transformation, and B(i,j) is the blue component of the color image at the coordinate (i,j) before transformation; Then the median filtering method is used to remove the salt and pepper noise generated during the transmission process; Finally, the closing operation in the geometric morphology method is used to remove the tiny impurities in the image.

3. The method for measuring the external dimensions of sweet cherries based on image recognition according to claim 2, characterized in that: The method of finally using the closing operation in the geometric morphology method to remove tiny impurities in the image includes: First, perform the dilation operation: ; Then perform the erosion operation: ; Finally, the closed operation formula is: ; Among them, A and B are two image sets, x is the distance B moves, is the translation of set B, is a mapping of set B.

4. The method for measuring the external dimensions of sweet cherries based on image recognition according to claim 1, characterized in that: The top view matches different matching templates in the ROI area according to different sizes of the sweet cherries to obtain geometric matching scores of the sweet cherries, and the geometric matching template with the highest score is obtained by comparison according to the geometric matching scores as a calibration template for calibration, including: establishing a conversion between real-world coordinates and pixel coordinates by the following formula: Where: ( , ) is the world coordinate, ( ) is the image coordinate, ( , ) is the pixel coordinate, is the rotation matrix, is the offset vector, is the camera intrinsic parameter matrix, is the perspective photography matrix, is the scale factor.

Images