Tea leaf picking method with recognition function for tea leaf planting

Abstract

The invention relates to the technical field of picking, in particular to a tea leaf picking method with a recognition function for tea leaf planting, which comprises the following steps: a deep learning algorithm is combined with an image recognition technology to process tea leaf images in real time and recognize and position tea leaf tender shoots, the recognition accuracy is not lower than 98%, and the processing speed is not lower than 10 frames per second; the tea leaf picking method with the recognition function for tea leaf planting has the beneficial effects that a deep learning algorithm is combined with an image recognition technology, tea leaf images are processed in real time, tea leaf tender shoots are recognized and positioned, the recognition accuracy is not lower than 98%, and the picking accuracy and efficiency are effectively improved.

Description

Technical Field

[0001] The invention relates to the field of tea picking, in particular to a tea picking method with a tea planting identification function. Background Art

[0002] Tea is an important economic crop, and its picking process has traditionally relied on manual labor. However, with the expansion of tea planting areas and the increase in labor costs, traditional manual picking methods have been unable to meet the needs of efficient and accurate picking. In order to improve the efficiency and quality of tea picking, a variety of automated and intelligent tea picking methods have emerged in recent years.

[0003] Most existing automated tea picking methods use robotic arms and image recognition technology to pick tea leaves according to preset picking paths and fixed cutting parameters. However, these methods have some problems, such as low recognition accuracy, inaccurate control of cutting force and angle, lack of autonomous navigation and obstacle avoidance functions, etc., which makes it difficult to further improve picking efficiency and quality.

[0004] Therefore, a tea picking method with higher recognition accuracy, more precise cutting control, autonomous navigation and obstacle avoidance capabilities is needed to meet the actual needs of the tea planting industry. Summary of the invention

[0005] The object of the present invention is to provide a tea picking method with tea planting identification function to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solution: a tea picking method with tea planting identification function, the method comprising the following steps: Use deep learning algorithms combined with image recognition technology to process tea images in real time, identify and locate tea buds, with an identification accuracy of no less than 98% and a processing speed of no less than 10 frames per second; According to the size, color and position information of the buds, the cutting force and angle of the picking mechanical module are automatically adjusted through a closed-loop control system. The cutting force error does not exceed ±2 grams, and the cutting angle error does not exceed ±2 degrees. By using unmanned driving technology, combined with high-precision maps and real-time positioning systems, the picking system can achieve autonomous navigation and obstacle avoidance in the tea garden. The driving speed is controlled at 0.8-1.2 meters per second, and the positioning accuracy is not less than ±5 centimeters.

[0007] Preferably, the bud identification comprises the following steps: Collect high-definition images in the tea garden, with an image resolution of no less than 20 million pixels and a collection frequency of no less than 20 Hz; Preprocess the image, including denoising, contrast enhancement and color correction, with the preprocessing time not exceeding 0.1 seconds; The feature vector of the sprout is extracted from the preprocessed image. The feature vector includes color histogram, texture feature and shape contour, and the feature extraction accuracy is not less than 95%.

[0008] Preferably, the picking machine module further comprises the following steps: Use high-precision servo motor to drive the cutting device to achieve precise control of cutting force, with a cutting force range of 10-15 Newtons; The end effector uses double negative pressure adsorption technology, and the negative pressure value is controlled between -60 and -80 kPa, ensuring that the tender buds are not damaged during the picking process while improving the picking efficiency.

[0009] Preferably, the end effector also includes intelligent recognition and adjustment functions, which automatically adjusts the adsorption position and strength according to the actual situation of the sprouts to ensure the integrity and quality of the sprouts.

[0010] Preferably, it also includes a driving system, comprising the following steps: The environmental perception system that combines lidar and cameras can sense obstacles and terrain changes in the tea garden in real time, and the obstacle avoidance distance is controlled between 10-15 cm. The positioning technology combining differential GPS and inertial navigation system is adopted to realize real-time high-precision positioning of the picking system, with a positioning accuracy of no less than ±3 cm.

[0011] Preferably, the method further comprises data processing and storage steps, wherein: Use high-performance embedded processors to process and analyze the bud identification results, picking location and time information in real time, with a processing speed of no less than 1GB / s; The processed data is stored in a solid-state drive with a storage capacity of no less than 1TB and a data storage speed of no less than 200MB / s; The stored data is deeply mined and analyzed to optimize the picking strategy, with the analysis time not exceeding 30 seconds.

[0012] Preferably, the data analysis step further comprises: According to the growth cycle of tea buds and historical data of picking, the picking quantity and quality in the next week are predicted; According to the terrain and climate data of the tea garden, the picking path and time are optimized to improve the picking efficiency and quality.

[0013] Preferably, it also includes a remote monitoring and control step, wherein: Remote real-time monitoring of the picking system is achieved through 4G / 5G networks, including the accuracy of bud identification, cutting force, driving speed and location information; Remote control and parameter adjustment of the picking system can be achieved through mobile phone APP or web page, with the control accuracy not less than ±1%.

[0014] Preferably, it also includes intelligent maintenance and fault warning steps, wherein: Use sensors to monitor the status of key components of the picking system in real time, such as motor temperature, hydraulic system pressure, and battery power; When an abnormal state is detected, a fault warning is automatically issued and fault location and solutions are provided; Based on maintenance plans and historical data, it can intelligently predict and remind users to perform regular maintenance and care.

[0015] Preferably, the method further comprises the step of intelligently classifying and storing the picked tea buds, wherein: Use machine vision technology to intelligently classify the picked tea buds, with a classification accuracy of no less than 99% and a classification speed of no less than 1,000 buds per minute; The sorted tea buds are stored in an intelligent cold storage with a temperature controlled between 0-4°C and a humidity controlled between 65%-70% to maintain the freshness and quality of the tea buds.

[0016] Compared with the prior art, the present invention has the following beneficial effects: The tea picking method with tea planting identification function proposed by the present invention processes tea images in real time through deep learning algorithm combined with image recognition technology, identifies and locates tea buds, and the recognition accuracy rate is not less than 98%, which effectively improves the accuracy and efficiency of picking; utilizes unmanned driving technology, combined with high-precision maps and real-time positioning systems, to achieve autonomous navigation and obstacle avoidance of the picking system in the tea garden, control the driving speed at 0.8-1.2 meters per second, and the positioning accuracy is not less than ±5 centimeters, which improves the flexibility and safety of picking; uses machine vision technology to intelligently classify the picked tea buds, with a classification accuracy rate of not less than 99% and a classification speed of not less than 1000 buds per minute, and stores the classified buds in an intelligent cold storage, thereby maintaining the freshness and quality of the buds and providing better raw materials for subsequent tea processing. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 The figure is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0017] In order to make the purpose and technical solution of the present invention clearly and completely described, and the advantages more clearly understood, the embodiments of the present invention are further described in detail with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are part of the embodiments of the present invention, rather than all of the embodiments, and are only used to explain the embodiments of the present invention, and are not used to limit the embodiments of the present invention. All other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0018] For example, see Figure 1 The present invention provides a technical solution: a tea picking method with tea planting identification function, the specific steps are as follows: Bud recognition: Using deep learning algorithms combined with image recognition technology, tea images are processed in real time to identify and locate tea buds, with an accuracy rate of 98.5% and a processing speed of 12 frames per second. High-definition images of tea gardens are collected with an image resolution of 22 million pixels and an acquisition frequency of 25Hz. The images are preprocessed (denoising, contrast enhancement, and color correction) with a preprocessing time of 0.08 seconds. Feature vectors of buds are extracted from the preprocessed images with a feature extraction accuracy of 96%.

[0019] Picking mechanical module: A high-precision servo motor is used to drive the cutting device, and the cutting force is controlled at 12 Newtons with an error of ±1 gram. The cutting angle is automatically adjusted according to the position information of the buds with an error of ±1 degree. The end effector adopts double negative pressure adsorption technology with a negative pressure value of -70 kPa to ensure that the buds are not damaged.

[0020] Autonomous navigation and obstacle avoidance: Using unmanned driving technology, combined with high-precision maps and real-time positioning systems, the picking system can achieve autonomous navigation in the tea garden, with a driving speed of 1 meter per second and a positioning accuracy of ±4 centimeters. Using an environmental perception system combining laser radar and cameras, the obstacle avoidance distance is 12 centimeters.

[0021] Embodiment 2, based on embodiment 1, proposes a tea picking method with tea planting identification function, and the specific steps are as follows: Bud recognition: recognition accuracy 99%, processing speed 15 frames / second. Image resolution 25 megapixels, acquisition frequency 22Hz. Preprocessing time 0.09 seconds, feature extraction accuracy 97%.

[0022] Picking mechanical module: cutting force range 10-14 Newtons, error ±1.5 grams. Cutting angle error ±1.5 degrees. The end effector has intelligent recognition and adjustment functions, automatically adjusting the adsorption position and strength according to the actual situation of the buds.

[0023] Data processing and storage: Using high-performance embedded processors, the processing speed is 1.2GB / s. The storage capacity is 1.5TB, and the data storage speed is 250MB / s. The stored data is deeply mined and analyzed, and the analysis time is 25 seconds.

[0024] Embodiment 3, based on embodiment 2, proposes a tea picking method with tea planting identification function, and the specific steps are as follows: Bud recognition: recognition accuracy 98%, processing speed 10 frames / second. Image resolution 20 megapixels, acquisition frequency 20Hz. Preprocessing time 0.1 seconds, feature extraction accuracy 95%.

[0025] Autonomous navigation and obstacle avoidance: Driving speed 0.9 m / s, positioning accuracy ±3 cm. Obstacle avoidance distance 10 cm, using positioning technology combining differential GPS and inertial navigation system.

[0026] Remote monitoring and control: Remote real-time monitoring is achieved through the 5G network, and the monitoring content includes the accuracy of sprout recognition, cutting force, driving speed and location information, etc. Remote control and parameter adjustment are achieved through the mobile phone APP, with a control accuracy of ±0.8%.

[0027] Intelligent maintenance and fault warning: Real-time monitoring of key component status such as motor temperature, hydraulic system pressure and battery power. Automatically issue fault warnings and provide fault location and solutions.

[0028] Embodiment 4, based on embodiment 3, proposes a tea picking method with tea planting identification function, and the specific steps are as follows: Bud recognition: recognition accuracy 98.2%, processing speed 11 frames / second. Image resolution 21 megapixels, acquisition frequency 23Hz. Preprocessing time 0.095 seconds, feature extraction accuracy 95.5%.

[0029] Picking and storage: Cutting force range is 11-15 Newtons, with an error of ±2 grams. Cutting angle error is ±2 degrees. Machine vision technology is used to intelligently classify the picked tea buds, with a classification accuracy of 99.5% and a classification speed of 1,200 buds per minute. The classified tea buds are stored in an intelligent cold storage with a temperature controlled at 2°C and a humidity controlled at 68%.

[0030] Data analysis and optimization: Based on the growth cycle of tea buds and historical data of picking, the picking quantity and quality in the next week are predicted. Based on the terrain and climate data of the tea garden, the picking path and picking time are optimized.

[0031] The comparative data table of Examples 1 to 4 and the prior art is as follows:

[0032] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A tea picking method with tea planting identification function, characterized in that: The method comprises the following steps: Use deep learning algorithms combined with image recognition technology to process tea images in real time, identify and locate tea buds, with an identification accuracy of no less than 98% and a processing speed of no less than 10 frames per second; According to the size, color and position information of the buds, the cutting force and angle of the picking mechanical module are automatically adjusted through a closed-loop control system. The cutting force error does not exceed ±2 grams, and the cutting angle error does not exceed ±2 degrees. By using unmanned driving technology, combined with high-precision maps and real-time positioning systems, the picking system can achieve autonomous navigation and obstacle avoidance in the tea garden. The driving speed is controlled at 0.8-1.2 meters per second, and the positioning accuracy is not less than ±5 centimeters.

2. A tea picking method with tea planting identification function according to claim 1, characterized in that: Shoot identification involves the following steps: Collect high-definition images in the tea garden, with an image resolution of no less than 20 million pixels and a collection frequency of no less than 20 Hz; Preprocess the image, including denoising, contrast enhancement and color correction, with the preprocessing time not exceeding 0.1 seconds; The feature vector of the sprout is extracted from the preprocessed image. The feature vector includes color histogram, texture feature and shape contour, and the feature extraction accuracy is not less than 95%.

3. The tea picking method with tea planting identification function according to claim 1, characterized in that: The picking machine module also includes the following steps: Use high-precision servo motor to drive the cutting device to achieve precise control of cutting force, with a cutting force range of 10-15 Newtons; The end effector uses double negative pressure adsorption technology, and the negative pressure value is controlled between -60 and -80 kPa, ensuring that the tender buds are not damaged during the picking process while improving the picking efficiency.

4. The tea picking method with tea planting identification function according to claim 3 is characterized in that: The end effector also includes intelligent recognition and adjustment functions, which automatically adjusts the adsorption position and strength according to the actual situation of the sprouts to ensure the integrity and quality of the sprouts.

5. The tea picking method with tea planting identification function according to claim 1, characterized in that: Also included is a driving system, comprising the following steps: The environmental perception system that combines lidar and cameras can sense obstacles and terrain changes in the tea garden in real time, and the obstacle avoidance distance is controlled between 10-15 cm. The positioning technology combining differential GPS and inertial navigation system is adopted to realize real-time high-precision positioning of the picking system, with a positioning accuracy of no less than ±3 cm.

6. The tea picking method with tea planting identification function according to claim 1, characterized in that: It also includes data processing and storage steps, wherein: Use high-performance embedded processors to process and analyze the bud identification results, picking location and time information in real time, with a processing speed of no less than 1GB / s; The processed data is stored in a solid-state drive with a storage capacity of no less than 1TB and a data storage speed of no less than 200MB / s; The stored data is deeply mined and analyzed to optimize the picking strategy, with the analysis time not exceeding 30 seconds.

7. A tea picking method with tea planting identification function according to claim 6, characterized in that: The data analysis steps also include: According to the growth cycle of tea buds and historical data of picking, the picking quantity and quality in the next week are predicted; According to the terrain and climate data of the tea garden, the picking path and time are optimized to improve the picking efficiency and quality.

8. The tea picking method with tea planting identification function according to claim 1, characterized in that: It also includes remote monitoring and control steps, wherein: Remote real-time monitoring of the picking system is achieved through 4G / 5G networks, including the accuracy of bud identification, cutting force, driving speed and location information; Remote control and parameter adjustment of the picking system can be achieved through mobile phone APP or web page, with the control accuracy not less than ±1%.

9. The tea picking method with tea planting identification function according to claim 1, characterized in that: It also includes intelligent maintenance and fault warning steps, including: Use sensors to monitor the status of key components of the picking system in real time, such as motor temperature, hydraulic system pressure, and battery power; When an abnormal state is detected, a fault warning is automatically issued and fault location and solutions are provided; Based on maintenance plans and historical data, it can intelligently predict and remind users to perform regular maintenance and care.

10. The tea picking method with tea planting identification function according to claim 1, characterized in that: The method also includes the steps of intelligently classifying and storing the picked tea buds, wherein: Use machine vision technology to intelligently classify the picked tea buds, with a classification accuracy of no less than 99% and a classification speed of no less than 1,000 buds per minute; The sorted tea buds are stored in an intelligent cold storage with a temperature controlled between 0-4°C and a humidity controlled between 65%-70% to maintain the freshness and quality of the tea buds.

Images