Tea tip weight rapid measurement method for tea leaf picking robot

Abstract

The application discloses a kind of tea picking robot with tea tip weight fast measurement method, comprising the following steps: using binocular camera to collect tea leaf RGB image and depth image;The image collected is preprocessed;Feature point extraction is carried out to the image after preprocessing;Iterative segmentation is used to the image after feature extraction Segmentation processing;Contour processing is carried out to the segmented image, and the pixel number in tea bud area is calculated using pixel counting method, and then the tea bud area is obtained;Based on contour and pixel, tea bud thickness is obtained;By sample mean method, the average density of tea bud is obtained;Based on the tea bud area, thickness and average density, the weight of tea bud is obtained.The application can realize fast and high-precision tea bud weight estimation in complex environment.

Description

Technical Field

[0001] This invention belongs to the field of vision technology for tea picking robots, and in particular, it is a method for rapidly measuring the weight of tea tips using a tea picking robot. Background Technology

[0002] Tea is one of the oldest and most widely consumed beverages in the world. Tea picking has always been a labor-intensive task. Harvesting requires prior estimation of tea garden yield beforehand. Researching tea weight estimation techniques for harvesting can help predict tea garden yield in advance, enabling rational production planning and reducing manpower consumption. Currently, there is no technology, either domestically or internationally, for estimating tea yield based on image processing, which hinders resource allocation and management and can lead to significant economic losses. Therefore, there is an urgent need for a method to accurately and quickly calculate the weight of tender tea buds before the harvesting stage.

[0003] To address the aforementioned technical challenges, research has emerged that applies computer vision and image processing technologies to the tea industry. This typically involves tea identification and detection techniques based on machine feature recognition or deep learning, as detailed below:

[0004] Chinese patent document CN112633212A discloses a deep learning-based method for tea bud segmentation. This method constructs a deep learning network model and uses a trained, improved network model to segment, identify, and predict tea buds. While this method further improves the efficiency of tea bud segmentation, if the number of tea bud samples in the training data is unbalanced with other categories, the model may make poor predictions for the fewer categories. This could lead to the model failing to accurately identify tea buds or misidentifying samples from other categories as tea buds. Furthermore, segmenting tender buds using only this network model is relatively coarse, which is not conducive to subsequent contour processing of the bounding boxes.

[0005] Patent document CN114973233A discloses a method for detecting tea shoots based on feature fusion graph neural networks. This method overcomes the occlusion problem in tea shoot detection by using an optical trap point cloud acquisition system and improves detection accuracy by employing an improved graph neural network, thus achieving accurate identification of tea shoots. Although this method can effectively overcome the occlusion problem in tea shoot detection, its model, which relies heavily on manually labeled feature extraction, struggles to extract deep information between image pixels, making it difficult to transfer the model to other tea varieties and exhibiting weak generalization performance.

[0006] Patent document CN104050668A discloses a target recognition method for green tea buds based on binocular vision technology. This invention analyzes and processes images of green tea buds growing in a standard tea garden to identify the bud areas, controls corresponding actuators to perform harvesting activities through a corresponding program, and improves the target recognition algorithm. However, this invention still uses a relatively traditional RGB space to HSI space conversion algorithm. This algorithm has low segmentation accuracy and is difficult to cope with the complex real-world environment. Special weather and lighting conditions in the actual tea field environment can significantly affect the accuracy of segmentation and recognition.

[0007] While the existing literature mentioned above can all identify tea buds, and even has special methods to overcome the occlusion problem, they are generally difficult to transfer to a wide range of practical applications. The recognition accuracy is also not high, and no design method has been considered to obtain the actual yield of tea gardens through image processing to help better plan production activities. Summary of the Invention

[0008] The purpose of this invention is to address the problems existing in the prior art by providing a method for rapid measurement of tea tip weight using a tea picking robot.

[0009] The technical solution to achieve the purpose of this invention is: a method for rapid measurement of tea tip weight using a tea picking robot, the method comprising the following steps:

[0010] Step 1: Use a stereo camera to acquire RGB images and depth images of tea leaves;

[0011] Step 2: Preprocess the images acquired in Step 1;

[0012] Step 3: Extract feature points from the preprocessed image;

[0013] Step 4: Iterative segmentation is used to segment the image after feature extraction;

[0014] Step 5: Perform contour processing on the segmented image, and use the pixel counting method to calculate the number of pixels in the tea bud area, thereby obtaining the area of ​​the tea bud.

[0015] Step 6: Calculate the thickness of the tea buds based on the contour and pixels;

[0016] Step 7: Calculate the average density of tea buds using the sample mean method;

[0017] Step 8: Calculate the weight of the tea buds based on their area, thickness, and average density.

[0018] Furthermore, step 4 employs iterative segmentation to segment the image after feature extraction, specifically including the following steps:

[0019] Step 4-1, let the minimum grayscale value T of the acquired tea leaf image be... min and maximum grayscale value T max Calculate the grayscale threshold T of the tea leaf image. k for:

[0020]

[0021] Step 4-2: Segment the tea image using the grayscale threshold, dividing the tea image into region C1 with a grayscale value greater than T0 and region C2 with a grayscale value less than T0. Then calculate the average grayscale values ​​of regions C1 and C2, which are μ1 and μ2, respectively.

[0022] Step 4-3, calculate the new grayscale threshold T k+1 :

[0023]

[0024] Based on the above process, the grayscale threshold of the image is repeatedly calculated until the interpolation of two adjacent thresholds is less than the preset value, and then the current grayscale threshold is taken as the optimal threshold.

[0025] Step 4-4: Perform image segmentation using the optimal threshold.

[0026] Furthermore, step 6, which involves determining the thickness of tea buds based on contours and pixels, specifically includes:

[0027] Step 6-1: Using the uniform segmentation method, select a point as a sampling point on each line segment of the outline of the tea bud area.

[0028] Step 6-2: Based on the parallax principle, calculate the distance from each sampling point to the back of the blade, and use it as the thickness value of that sampling point;

[0029] Step 6-3: Multiply the thickness value of each sampling point by the pixel size to obtain the actual thickness value;

[0030] Step 6-4: Statistically analyze the actual thickness values ​​of all sampling points and calculate the average thickness value, which is taken as the thickness of the tea buds.

[0031] Furthermore, step 7, which involves calculating the average density of tea buds using the sample mean method, specifically includes:

[0032] A large number of tea leaves were selected as samples and their density was measured. The formula for calculating the density of a single tea bud is as follows:

[0033] Density of a single tea bud = Weight of a single tea bud / (Area of ​​a tea bud × Thickness of a tea bud)

[0034] The average density of all tea bud samples is calculated as the average density of tea buds.

[0035] Furthermore, in step 8, the weight of the tea buds is calculated based on their area, thickness, and average density. The specific calculation formula is as follows:

[0036] Weight of a single tea bud = area of ​​tea bud × thickness of tea bud × average density of tea bud.

[0037] Furthermore, the method also includes:

[0038] Based on the weight of tea buds, the average yield of the tea garden is estimated as follows:

[0039] Randomly sample tea trees and determine the average number of their tea buds. Multiply this average number by the weight of the tea buds to obtain the average yield of the tea garden.

[0040] Compared with the prior art, the present invention has the following significant advantages:

[0041] (1) The improved SIFT algorithm was used to extract feature points from tea leaf images, which accelerated the recognition speed of the images and avoided the problems of slow analysis speed and poor analysis effect of traditional algorithms.

[0042] (2) The median filtering method is adopted, which continuously adjusts the position of the pixel value to make the pixel value move towards the peak point of the density estimation, so as to achieve a smoothing effect. It can effectively remove noise while preserving edge details.

[0043] (3) The Canny algorithm is used to perform edge detection with different parameters and thresholds for each region, which increases the accuracy of edge detection and makes it more robust.

[0044] (4) By using the parallax principle, the thickness of the bud is calculated based on the distance information between pixels, which solves the problem of calculating the thickness of tea buds with irregular shapes and small volume, and thus enables more accurate and faster calculation of tea weight.

[0045] The present invention will now be described in further detail with reference to the accompanying drawings. Attached Figure Description

[0046] Figure 1 This is a flowchart of a method for rapidly measuring the weight of tea tips using a tea-picking robot, as described in one embodiment.

[0047] Figure 2 This is a flowchart illustrating the iterative segmentation and contour extraction process of tea bud images using a tea-picking robot in one embodiment.

[0048] Figure 3 One embodiment uses a grayscale image of a tea-picking robot, wherein... Figure 3 (a) in the figure is a scatter plot of the H factor in the HIS space; Figure 3 (b) in the image is the image with the H factor extracted and grayscale processed.

[0049] Figure 4 One embodiment uses a tea-picking robot to capture original images of tea leaves, in which... Figure 4 (a) in the image is an RGB image; Figure 4 (b) in the image is a depth image.

[0050] Figure 5 This is a schematic diagram of image preprocessing used by a tea-picking robot in one embodiment. Figure 5 (a) in the image is the image before correction; Figure 5 (b) in the image is the corrected image; Figure 5 (c) in the image is a binary image.

[0051] Figure 6 This is a schematic diagram illustrating image segmentation and contour extraction used in one embodiment of a tea-picking robot. Figure 6 (a) in the diagram represents image segmentation; Figure 6 (b) in the image is the edge image extracted by the Canny operator. Detailed Implementation

[0052] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0053] It should be noted that if the embodiments of the present invention involve descriptions such as "first" and "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" and "second" may explicitly or implicitly include at least one of those features. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by the present invention.

[0054] This invention is based on machine vision and combines image processing technology, machine learning technology and deep learning technology to accurately identify the weight of tender tea shoots in complex environments.

[0055] In one embodiment, combined Figure 1 and Figure 2 A method for rapidly measuring the weight of tea tips using a tea-picking robot is provided, the method comprising the following steps:

[0056] Step 1: Use a stereo camera to acquire RGB images and depth images of tea leaves, such as... Figure 4 As shown;

[0057] Here, the camera first takes a picture of the reference surface, fixing the distance between the camera and the reference. The camera then captures the original image of the bud. The camera position is stabilized, with the distance between the camera and the tea leaves equal to the distance between the camera and the reference, and the digital camera is set to autofocus mode to ensure image sharpness. The resulting original image is shown below. Figure 3 As shown.

[0058] Step 2 involves preprocessing the images acquired in Step 1, specifically including:

[0059] (1) Grayscale conversion

[0060] For color images, first convert them to grayscale. For example... Figure 4 As shown, grayscale images contain only brightness information and not color information, which helps simplify image processing and reduce data volume. Based on the human eye's sensitivity to different colors, the RGB channel values ​​are weighted and averaged to obtain the grayscale value.

[0061] (2) Geometric correction

[0062] When capturing images of plant leaves using a digital camera, the images may exhibit tilt distortion. Geometric correction is necessary to ensure that the leaf images accurately represent the true information.

[0063] (3) Binarization

[0064] (4) Median filtering for noise reduction

[0065] To eliminate noise in an image, median filtering is used. This method involves sorting the neighboring pixels of a given pixel in ascending order of grayscale level, and then taking the median value as the new grayscale value for that pixel. This approach effectively eliminates noise while preserving the integrity of the image's edge information.

[0066] Preprocessing diagram as shown Figure 5 As shown.

[0067] Step 3: Extract feature points from the preprocessed image;

[0068] Here, the SIFT (Scale Invariant Feature Transform) algorithm is preferably used. The following are the steps for extracting feature points from the preprocessed leaf image using the SIFT algorithm:

[0069] (1) The SIFT algorithm finds potential feature points by detecting local extrema at different scales. First, it defines an image scale function I(x,y), performs convolution operation between the original image I(x,y) and a variable-scale 2D Gaussian function G(x,y,σ), and then subtracts the Gaussian smoothed images at adjacent scales to obtain the difference Gaussian pyramid.

[0070] The Gaussian kernel function is defined as:

[0071]

[0072] (2) Around the detected extreme points, the SIFT algorithm uses a difference Gaussian function to pinpoint the precise location of keypoints. Feature descriptors are generated by calculating the gradient magnitude m(x, y) and direction θ(x, y) of all pixels within the Gaussian window.

[0073]

[0074] We obtain key information for each feature point, which is independent of scale.

[0075] (3) The extracted feature points and their descriptors are used for matching between different images. SIFT is used for efficient matching by calculating two feature vectors a(x) 11 x 12 ,.........,x 1n ) and b(x 21 x 22 ,.........,x 2n Using the Euclidean distance between the two images, we can match them:

[0076]

[0077] (4) Visualize the feature points on the original image to observe their location and distribution.

[0078] Step 4: Iterative segmentation is used to segment the image after feature extraction;

[0079] Preferably, this invention employs iterative segmentation to segment the image after feature extraction, and the segmentation result is as follows: Figure 6 As shown in (a), the specific steps include:

[0080] Step 4-1, let the minimum grayscale value T of the acquired tea leaf image be... min and maximum grayscale value T max Calculate the grayscale threshold T of the tea leaf image. k for:

[0081]

[0082] Step 4-2: Segment the tea image using the grayscale threshold, dividing the tea image into region C1 with a grayscale value greater than T0 and region C2 with a grayscale value less than T0. Then calculate the average grayscale values ​​of regions C1 and C2, which are μ1 and μ2, respectively.

[0083] Step 4-3, calculate the new grayscale threshold T k+1 :

[0084]

[0085] Based on the above process, the grayscale threshold of the image is repeatedly calculated until the interpolation of two adjacent thresholds is less than the preset value, and then the current grayscale threshold is taken as the optimal threshold.

[0086] Step 4-4: Perform image segmentation using the optimal threshold.

[0087] Step 5: Perform contour processing on the segmented image, and use the pixel counting method to calculate the number of pixels in the tea bud area, thereby obtaining the area of ​​the tea bud.

[0088] Preferably, the Canny algorithm is used for contour processing of the image, and the result is as follows: Figure 6 As shown in (b) above, the specific steps are as follows:

[0089] (1) A Gaussian filter is used to perform convolution filtering on the input image to filter out noise and reduce its impact on gradient calculation. The Gaussian filter is shown below:

[0090]

[0091] (2) The first-order difference operator is used to calculate the gradient magnitude components in the vertical and horizontal directions to obtain the magnitude M and direction θ of the image gradient:

[0092]

[0093]

[0094] (3) Apply non-maximum suppression to the gradient magnitude using the gradient direction. Traverse each pixel in the gradient magnitude image M[i,j] and use interpolation to calculate the gradient magnitude between two adjacent pixels in the gradient direction of the current pixel. If the gradient magnitude of the current pixel is greater than or equal to these two values, the pixel is considered to be on the edge of the image and may be an edge point; otherwise, the pixel is considered not to be an edge point, and the image edge is thinned to one pixel to obtain the non-maximum suppressed image.

[0095] (4) Perform double threshold detection and edge connection on the image, and use a fill scan algorithm to fill the leaf and obtain the number of pixels filled. Apply two high and low thresholds T1 and T2 to the non-maximum suppression image to obtain edge images E1 and E2, respectively. Among them, E1 is a strong edge point that may have discontinuities; E2 is a weak edge point. Track the edges in E1. When an edge reaches an endpoint, search for edge points in the 8-neighborhood of the corresponding position in image E2 to connect the discontinuities in the strong edge E1. Gradually search and track edges to connect the discontinuous edges in the high threshold image. When a complete edge is obtained, use a fill scan algorithm (such as FloodFill) to fill the inside of the leaf. Select a seed point (usually at a certain position inside the leaf), and then start scanning and filling from that point. The filling will continue to expand until a strong edge or image boundary is encountered, thereby obtaining the number of pixels filled in the leaf. Obtain the edge image with extracted contours, such as Figure 3 As shown.

[0096] Here, the area of ​​the tea buds is calculated as follows:

[0097] The method for calculating the area of ​​a tea leaf involves determining the actual area represented by each pixel in the image by using the actual area of ​​a known reference object. The total number of pixels in the image is then calculated by counting the number of pixels filling the leaf. This total number of pixels can be obtained by determining the image resolution, which can usually be found in the image's metadata or obtained from the image's source. For example... Figure 6 The resolution of (a) in the image is 1280*960. Given the actual area of ​​the reference object, then...

[0098]

[0099] The area of ​​the photograph can be measured manually or by a program.

[0100] The actual area represented by each pixel in the image = the actual area of ​​the reference object / the number of pixels in the reference object. This is achieved by taking images of tea leaves under identical conditions.

[0101]

[0102] Step 6: Calculate the thickness of the tea buds based on the contour and pixels. The specific process includes:

[0103] Step 6-1: Using the uniform segmentation method, select a point as a sampling point on each line segment of the outline of the tea bud area.

[0104] Step 6-2: Based on the parallax principle, calculate the distance from each sampling point to the back of the blade, and use it as the thickness value of that sampling point;

[0105] Step 6-3: Multiply the thickness value of each sampling point by the pixel size to obtain the actual thickness value;

[0106] Step 6-4: Statistically analyze the actual thickness values ​​of all sampling points and calculate the average thickness value, which is taken as the thickness of the tea buds.

[0107] Step 7: Calculate the average density of tea buds using the sample mean method; specifically including:

[0108] A large number of tea leaves were selected as samples and their density was measured. The formula for calculating the density of a single tea bud is as follows:

[0109] Density of a single tea bud = Weight of a single tea bud / (Area of ​​a tea bud × Thickness of a tea bud)

[0110] The average density of all tea bud samples is calculated as the average density of tea buds.

[0111] Step 8: Based on the area, thickness, and average density of the tea buds, calculate the weight of the tea buds using the following formula:

[0112] Weight of a single tea bud = area of ​​tea bud × thickness of tea bud × average density of tea bud.

[0113] Furthermore, in one embodiment, the method for rapid measurement of tea tip weight using the tea-picking robot further includes:

[0114] Based on the weight of tea buds, the average yield of the tea garden is estimated as follows:

[0115] Randomly sample tea trees and determine the average number of their tea buds. Multiply this average number by the weight of the tea buds to obtain the average yield of the tea garden.

[0116] In one embodiment, a rapid weight measurement system for tea tips using a tea-picking robot is provided, the system comprising:

[0117] The first module is used to acquire RGB and depth images of tea leaves using a binocular camera.

[0118] The second module is used for preprocessing the acquired images;

[0119] The third module is used to extract feature points from the preprocessed image;

[0120] The fourth module is used to segment the image after feature extraction using iterative segmentation.

[0121] The fifth module is used to perform contour processing on the segmented image, and uses the pixel counting method to calculate the number of pixels in the tea bud area, thereby obtaining the area of ​​the tea bud.

[0122] The sixth module is used to determine the thickness of tea buds based on contours and pixels;

[0123] The seventh module is used to calculate the average density of tea buds using the sample mean method.

[0124] The eighth module is used to calculate the weight of the tea buds based on the area, thickness, and average density of the tea buds.

[0125] Specific limitations regarding the rapid weight measurement system for tea buds in tea-picking robots can be found in the above-mentioned limitations on the rapid weight measurement method for tea buds in tea-picking robots, and will not be repeated here. Each module in the aforementioned rapid weight measurement system for tea buds in tea-picking robots can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0126] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps:

[0127] Step 1: Use a stereo camera to acquire RGB images and depth images of tea leaves;

[0128] Step 2: Preprocess the images acquired in Step 1;

[0129] Step 3: Extract feature points from the preprocessed image;

[0130] Step 4: Iterative segmentation is used to segment the image after feature extraction;

[0131] Step 5: Perform contour processing on the segmented image, and use the pixel counting method to calculate the number of pixels in the tea bud area, thereby obtaining the area of ​​the tea bud.

[0132] Step 6: Calculate the thickness of the tea buds based on the contour and pixels;

[0133] Step 7: Calculate the average density of tea buds using the sample mean method;

[0134] Step 8: Calculate the weight of the tea buds based on their area, thickness, and average density.

[0135] For specific limitations on each step, please refer to the limitations on the method for rapid measurement of tea tip weight using tea picking robots mentioned above, which will not be repeated here.

[0136] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0137] Step 1: Use a stereo camera to acquire RGB images and depth images of tea leaves;

[0138] Step 2: Preprocess the images acquired in Step 1;

[0139] Step 3: Extract feature points from the preprocessed image;

[0140] Step 4: Iterative segmentation is used to segment the image after feature extraction;

[0141] Step 5: Perform contour processing on the segmented image, and use the pixel counting method to calculate the number of pixels in the tea bud area, thereby obtaining the area of ​​the tea bud.

[0142] Step 6: Calculate the thickness of the tea buds based on the contour and pixels;

[0143] Step 7: Calculate the average density of tea buds using the sample mean method;

[0144] Step 8: Calculate the weight of the tea buds based on their area, thickness, and average density.

[0145] For specific limitations on each step, please refer to the limitations on the method for rapid measurement of tea tip weight using tea picking robots mentioned above, which will not be repeated here.

[0146] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention without departing from its spirit and scope should be included within the protection scope of the present invention.

Claims

1. A method for rapid measurement of tea tip weight using a tea-picking robot, characterized in that, The method includes the following steps: Step 1: Use a stereo camera to acquire RGB images and depth images of tea leaves; Step 2: Preprocess the images acquired in Step 1; Step 3: Extract feature points from the preprocessed image; Step 4: Iterative segmentation is used to segment the image after feature extraction; Step 5: Perform contour processing on the segmented image, and use the pixel counting method to calculate the number of pixels in the tea bud area, thereby obtaining the area of ​​the tea bud. Step 6: Calculate the thickness of the tea buds based on the contour and pixels; Step 7: Calculate the average density of tea buds using the sample mean method; Step 8: Calculate the weight of the tea buds based on their area, thickness, and average density. Step 6, which involves determining the thickness of tea buds based on contours and pixels, specifically includes: Step 6-1: Using the uniform segmentation method, select a point as a sampling point on each line segment of the outline of the tea bud area; Step 6-2: Based on the parallax principle, calculate the distance from each sampling point to the back of the blade, and use it as the thickness value of that sampling point; Step 6-3: Multiply the thickness value of each sampling point by the pixel size to obtain the actual thickness value; Step 6-4: Statistically analyze the actual thickness values ​​of all sampling points and calculate the average thickness value, which is taken as the thickness of the tea buds.

2. The method for rapid measurement of tea tip weight using a tea-picking robot according to claim 1, characterized in that, The preprocessing described in step 2 includes grayscale conversion, geometric correction, and median filtering for noise reduction.

3. The method for rapid measurement of tea tip weight using a tea-picking robot according to claim 2, characterized in that, Step 3 specifically uses the SIFT feature extraction method to extract feature points.

4. The method for rapid measurement of tea tip weight using a tea-picking robot according to claim 3, characterized in that, Step 4 uses iterative segmentation to segment the image after feature extraction, specifically including the following steps: Step 4-1, let T be the minimum grayscale value of the acquired tea leaf image. min and maximum grayscale value T max Calculate the grayscale threshold T of the tea leaf image. k for: Step 4-2: Segment the tea image using the grayscale threshold, dividing the tea image into region C1 with a grayscale value greater than T0 and region C2 with a grayscale value less than T0. Then calculate the average grayscale values ​​of regions C1 and C2, which are μ1 and μ2, respectively. Step 4-3, calculate the new grayscale threshold T k+1 : Based on the above process, the grayscale threshold of the image is repeatedly calculated until the interpolation of two adjacent thresholds is less than the preset value, and then the current grayscale threshold is taken as the optimal threshold. Step 4-4: Perform image segmentation using the optimal threshold.

5. The method for rapid measurement of tea tip weight using a tea-picking robot according to claim 4, characterized in that, Step 5 specifically utilizes the Canny algorithm to perform contour processing on the image.

6. The method for rapid measurement of tea tip weight using a tea-picking robot according to claim 5, characterized in that, Step 7, which involves calculating the average density of tea buds using the sample mean method, specifically includes: A large number of tea leaves were selected as samples and their density was measured. The formula for calculating the density of a single tea bud is as follows: Density of a single tea bud = Weight of a single tea bud / (Area of ​​a tea bud × Thickness of a tea bud) The average density of all tea bud samples is calculated as the average density of tea buds.

7. The method for rapid measurement of tea tip weight using a tea-picking robot according to claim 6, characterized in that, In step 8, the weight of the tea buds is calculated based on their area, thickness, and average density. The specific calculation formula is as follows: Weight of a single tea bud = area of ​​tea bud × thickness of tea bud × average density of tea bud.

8. The method for rapid measurement of tea tip weight using a tea-picking robot according to claim 7, characterized in that, The method also includes: Based on the weight of tea buds, the average yield of the tea garden is estimated as follows: Randomly sample tea trees and determine the average number of their tea buds. Multiply this average number by the weight of the tea buds to obtain the average yield of the tea garden.

Images