Sprout tea harvesting machine and tea harvesting method based on multi modal recognition by using infrared ray spectrum
The sprout tea picking machine uses multimodal recognition to differentiate between old and new leaves, enhancing tea quality and efficiency by employing thermography, hyperspectral imaging, and laser cutting.
Patent Information
- Application Number
- JP2024073103
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-05-24
- Filing Date
- 2024-04-26
- Publication Date
- 2024-12-05
- Estimated Expiration
- 2044-04-26
AI Technical Summary
Existing tea picking machines fail to distinguish between old leaves and new buds, resulting in a mixture that affects tea quality and increases manual sorting costs, as they rely on visual recognition that is labor-intensive and inefficient for large areas.
A sprout tea picking machine using multimodal recognition by infrared spectrum, incorporating a fan assembly, slide rail assembly, recognition picking assembly, and manipulator, with thermography and hyperspectral imaging to differentiate between old and new leaves, and a laser generator for precise cutting.
Accurately distinguishes and picks new leaves, reducing manual sorting and improving tea quality by using thermography and hyperspectral imaging to detect cellulose content, while ensuring precise cutting with a laser generator.
Smart Images

Figure 2024169326000001_ABST
Abstract
Description
[Technical field]
[0001] The present invention relates to the technical field of tea plantation machinery, in particular to a tea bud plucking machine and tea plucking method based on multimodal recognition by infrared spectrum. [Background technology]
[0002] In recent years, China's tea industry has developed rapidly, and the cultivation area is constantly increasing. The contradiction between tea leaf picking and labor is a major challenge that exists in the current development of the tea industry. Tea leaf picking work has high requirements, and currently relies mainly on manual picking, but manual picking is time-consuming and labor-intensive, and cannot meet the needs of large-area tea leaf picking. When performing the picking work, the existing tea picking machine cannot distinguish old leaves and new shoots for all tea leaves in the visual area, and old leaves and new shoots are picked together, resulting in old leaves and new shoots mixed together, which affects the quality of the entire tea leaves, and then needs to be manually sorted and classified, which increases the cost input. Summary of the Invention [Problem to be solved by the invention]
[0003] The objective of the present invention is to provide a tea bud plucking machine and tea plucking method based on multimodal recognition by infrared spectrum, which can be used to distinguish old leaves from new buds and pick the new buds. [Means for solving the problem]
[0004] The technical solutions used by the present invention to solve the technical problems are as follows:The new tea bud picking machine based on multimodal recognition by infrared spectrum includes a frame, a fan assembly, a slide rail assembly, a recognition picking assembly and a manipulator, and the bottom of the frame has a traveling mechanism, the fan assembly includes a fan drive motor, a fan, a heating wire, a temperature sensor and a wind guide case, the fan and the fan drive motor are fixed to the top of the frame, the fan drive motor is used to drive and operate the fan, the wind guide case is located at the air outlet of the fan, the heating wire and the temperature sensor are located in the wind guide case, the heating wire is used to heat the air in the wind guide case to 40 degrees Celsius, and the temperature sensor is used to detect the air temperature at the outlet of the wind guide case, the slide rail assembly is located at the top of the frame, the recognition picking assembly includes a suction tube, a fixed frame and a manipulator, the fixed frame is located in a slide rail assembly and drives the vertical movement or / and lateral movement of the fixed frame through the slide rail assembly, the fixed frame is provided with a lead screw guide rail, a depth camera, The tea picking machine has a thermography and a hyperspectral imaging system, the thermography recognizes sprouts through a thermography image, and the hyperspectral imaging system is used to detect the cellulose content of the object recognized by the thermography to confirm whether it is a sprout or not, the depth camera recognizes and positions the sprouts recognized by the hyperspectral imaging system, and the intersection of the lines of sight of the thermography, the depth camera and the hyperspectral imaging system is a recognition area, one end of the suction tube is connected to a collection box located on a frame, and the other end of the suction tube is fixed to a positioning slider of a lead screw guide rail, and the other end of the suction tube moves vertically along with the positioning slider, the manipulator is fixed to the frame, and the manipulator has a laser generator, and the frame further has a control cabinet, and the control cabinet is signal-connected to the traveling mechanism, the fan assembly, the slide rail assembly, the recognition and picking assembly and the manipulator, and controls the operating state of the entire tea picking machine.
[0005] Furthermore, the traveling mechanism includes a traveling bracket, the traveling bracket has a traveling wheel and a traveling drive motor for driving the traveling wheel to operate, and the front end of the traveling bracket has two laser radars distributed on the left and right and installed diagonally downward, and the laser radars are used for map creation, positioning and route planning.
[0006] Further, the frame has two motor mounting frames, each of the motor mounting frames has one fan drive motor, and each of the fan drive motors has one fan on its output shaft, and the air outlets of the two fans are connected to the upper end of the air guide case, and the lower end of the air guide case faces the recognition area.
[0007] Further, the slide rail assembly includes a vertical slide rail, a lateral slide rail, a vertical slider, and a lateral slider, the lateral slide rail is fixed to the top of the frame, the lateral slider is slidably installed on the lateral slide rail, the vertical slide rail is fixed to the lateral slider and installed perpendicular to the lateral slide rail, the vertical slider is slidably installed on the vertical slide rail, the lateral slide rail has a first drive motor for driving and moving the lateral slider, the vertical slide rail has a second drive motor for driving and moving the vertical slider, and the fixed frame is fixedly connected to the vertical slider.
[0008] Furthermore, one end of the suction tube is fixed to the positioning slider via a suction tube fastener.
[0009] Furthermore, the cross-sectional dimension of the upper end of the air guide case is larger than the cross-sectional dimension of the lower end of the air guide case.
[0010] The present invention provides a method for picking tea buds based on multimodal recognition by infrared spectrum, comprising: (1) A laser radar performs mapping, positioning and route planning, and the traveling mechanism moves the entire tea picking machine forward according to the planned route; (2) the fan assembly is operated to blow hot air at a constant temperature of 40 degrees Celsius onto the leaf blade at the tip of the tea plant bud through the air guide case, to blow off dirt on the leaf blade at the tip of the bud, and to recognize the new buds using thermography based on the heating temperature difference between the tip of the bud and the leaf blade; (3) Using a hyperspectral imaging system to detect the cellulose content of the sprouts recognized by thermography, confirming the recognition accuracy of the thermography based on the cellulose content, and performing step (4) after the sprouts are detected as sprouts by the hyperspectral imaging system; (4) recognizing and locating the sprouts using a depth camera; (5) placing the manipulator in position, aligning the laser generator with the picking point on the sprout, and then placing the suction tube in position; (6) The laser generator operates to cut the sprouts, and at the same time, the suction tube operates to suck the cut sprouts into the collection box, completing the picking of the sprouts, and then resetting the suction tube and the manipulator; (7) A method for picking new tea buds based on multimodal recognition using infrared spectra, the method including a step of continuing to perform steps (2) to (6) by causing the tea picking machine to continue moving forward using a traveling mechanism until all tea picking work is completed. Effect of the Invention
[0011] The beneficial effects of the present invention are as follows: (1) The air inside the fan is heated to a constant temperature of 40°C using an electric heating wire, and then blown onto the tea leaves on the tea plant. On the one hand, the tips of the buds and the blades of the tea plants are heated, creating a temperature difference, and on the other hand, the hot air is blown onto the tea leaves, shaking them and blowing off any dirt on them. (2) The present invention first recognizes the heated tea leaf sprouts by thermography, and then uses a hyperspectral imaging system to detect the cellulose content of the tea leaf sprouts, thereby recognizing and determining the sprouts through dual detection. (3) The present invention combines deep learning to recognize and position the picking point of fresh leaf sprouts, and then cuts them with a laser generator, so the cutting is accurate. (4) The present invention can move the recognition and picking assembly according to the height of the tea leaves on the tea plant, and can recognize and pick the fresh leaf sprouts of tea plants at different canopy layer heights. [Brief description of the drawings]
[0012] [Figure 1] FIG. [Diagram 2] FIG. [Diagram 3] FIG. 2 is a front view of the fan assembly. [Figure 4] FIG. 2 is a front view of the slide rail assembly. [Diagram 5] FIG. 2 is a three-dimensional view of the recognition picking assembly. [Figure 6] FIG. 1 is a schematic diagram of the stereo location of the thermography, hyperspectral imaging system and depth camera. [Figure 7] 1 is a schematic diagram of the process of plucking new shoots. [Figure 8] 2 is a schematic diagram of how new shoots are picked. [Figure 9] 3 is a schematic diagram of how new shoots are picked. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0013] As shown in Figures 1 to 9, the present invention includes a traveling mechanism 1, a frame 2, a fan assembly 3, a slide rail assembly 4, a recognition and picking assembly 5, a manipulator 6 and a control cabinet 61, and the present invention will be described in detail below with reference to the drawings.
[0014] As shown in Figures 1 to 6, the new bud tea picking machine based on multimodal recognition by infrared spectrum includes a frame 2, a fan assembly 3, a slide rail assembly 4, a recognition picking assembly 5 and a manipulator 6, and the bottom of the frame 2 has a traveling mechanism 1, and the installation of the traveling mechanism 1 realizes the movement of the entire tea picking machine in the tea garden. As shown in Figure 2, the traveling mechanism 1 includes a traveling bracket 11, which has a traveling wheel 12 and a traveling drive motor 14 that drives and operates the traveling wheel 12, and the front end of the traveling bracket 11 has two laser radars 13 distributed on the left and right and installed diagonally downward, and the laser radar 13 is used for map creation, positioning and route planning, and the traveling wheel 12 moves forward according to the route planned by the laser radar 13.
[0015] As shown in Figures 1 and 3, the fan assembly 3 includes a fan driving motor 32, a fan 31, a heating wire 34, a temperature sensor 35 and an air guide case 36, the fan 31 and the fan driving motor 32 are fixed to the top of the frame 2, the fan driving motor 32 is used to drive the fan 31 to operate, the air guide case 36 is located at the air outlet of the fan 31, the heating wire 34 and the temperature sensor 35 are located in the air guide case 36, the heating wire 34 is used to heat the air in the air guide case 36 to 40 degrees Celsius, the temperature sensor 35 is used to detect the air temperature at the outlet of the air guide case 36, and when the air temperature at the outlet of the air guide case 36 is lower than 40 degrees Celsius, the power of the heating wire 34 is increased to raise the air temperature at the outlet of the air guide case 36, ensuring that the temperature of the hot air blown from the air guide case 36 to the tea leaves is 40 degrees Celsius. When the air temperature at the outlet of the air guide case 36 is higher than 40 degrees Celsius, the power of the heating wire 34 is reduced to reduce the air temperature at the outlet of the air guide case 36 until the air temperature at the outlet of the air guide case 36 reaches 40 degrees Celsius. The temperature sensor 35 and the heating wire 34 work together to stabilize the air temperature at the outlet of the air guide case 36 at 40 degrees Celsius. To facilitate assembly, the frame 2 has two motor mounting frames 33, each of which has one fan driving motor 32, each of which has one fan 31 on its output shaft, and the air outlets of the two fans 31 are connected to the upper end of the air guide case 36. In this way, the air guide case 36 is located between the two fans 31, and when the fan 31 operates, it blows air into the air guide case 36. The fan assembly 3 serves to blow out air at a constant temperature of 40 degrees Celsius.
[0016] As shown in FIG. 1, the slide rail assembly 4 is located at the top of the frame 2, and as shown in FIG. 5 and FIG. 6, the recognition and picking assembly 5 includes a suction tube 51, a fixed frame 53 and a manipulator 6, and the fixed frame 53 is located on the slide rail assembly 4 and drives the vertical movement and / or lateral movement of the fixed frame 53 via the slide rail assembly 4. As shown in Fig. 4, the slide rail assembly 4 includes a vertical slide rail 42, a horizontal slide rail 45, a vertical slider 43, and a horizontal slider 44, the horizontal slide rail 45 is fixed to the top of the frame 2, the horizontal slider 44 is slidably installed on the horizontal slide rail 45, the vertical slide rail 42 is fixed to the horizontal slider 44 and installed perpendicular to the horizontal slide rail 45, the vertical slider 43 is slidably installed on the vertical slide rail 42, the horizontal slide rail 45 has a first drive motor 46 for driving and moving the horizontal slider 44, and the vertical slide rail 42 has a second drive motor 41 for driving and moving the vertical slider 43. When the first drive motor 46 operates, the vertical slide rail 42 moves laterally in a horizontal plane together with the horizontal slide block 44. When the second drive motor 41 operates, the vertical slider 43 moves vertically on the vertical slide rail 42. The fixed frame 53 is fixedly connected to the vertical slider 43, thereby realizing the vertical and / or lateral movement of the fixed frame 53 under the action of the slide rail assembly 4.
[0017] As shown in FIG. 5, the fixed frame 53 has a lead screw guide rail 54, a depth camera 56, a thermograph 57 and a hyperspectral imaging system 58, the thermograph 57 recognizes the sprouts through the thermograph image, the hyperspectral imaging system 58 is used to detect the cellulose content of the object recognized by the thermograph 57 to confirm whether the object recognized by the thermograph 57 is a sprout, and the depth camera 56 recognizes and positions the sprouts identified by the hyperspectral imaging system 58. As shown in FIG. 5, the depth camera 56, the thermograph 57 and the hyperspectral imaging system 58 are fixed to the bottom of the fixed frame 53, and the depth camera 56, the thermograph 57 and the hyperspectral imaging system 58 are installed sequentially. The intersection of the lines of sight of the depth camera 56, the thermograph 57 and the hyperspectral imaging system 58 is the recognition area, and the lower end of the wind guide case 36 faces the recognition area. The cross-sectional dimension of the upper end of the wind guide case 36 is larger than that of the lower end of the wind guide case 36, thus increasing the pressure of the air discharged from the lower end of the wind guide case 36. The lead screw guide rail 54 is located between the depth camera 56, the thermography 57 and the hyperspectral imaging system 58 and is installed vertically, and the top of the lead screw guide rail 54 is fixedly connected to the fixed frame 53. One end of the suction tube 51 is connected to the collection box 62 located on the frame 2, and the other end of the suction tube 51 is fixed to the positioning slider of the lead screw guide rail 54, and for ease of assembly, the other end of the suction tube 51 is fixed to the positioning slider via the suction tube fastener 55, so that the other end of the suction tube 51 can move vertically along with the positioning slider. The fixed frame 53 has a servo 52, which is used to drive and move a positioning slider on a lead screw guide rail 54, thereby driving and moving the other end of the suction tube 51 vertically so that the other end of the suction tube 51 enters the recognition area.The manipulator 6 is fixed to the frame 2, and the manipulator 6 has a laser generator 59, which can move under the driving of the manipulator 6 until the laser generator 59 reaches a suitable position to perform the operations of cutting and picking. The frame 2 further has a control cabinet 61, which is signal-connected to the traveling mechanism 1, the fan assembly 3, the slide rail assembly 4, the recognition picking assembly 5 and the manipulator 6, and controls the operating state of the entire tea picking machine.
[0018] The present invention provides a method for picking tea buds based on multimodal recognition by infrared spectrum, comprising: (1) A step in which the laser radar 13 performs mapping, positioning and route planning, and the traveling mechanism moves the entire tea picking machine forward according to the planned route; (2) the fan assembly is operated to blow hot air at a constant temperature of 40 degrees Celsius onto the leaf blade at the tip of the tea plant bud through the air guide case, to blow off dirt on the leaf blade at the tip of the bud, and to recognize the new buds using thermography based on the heating temperature difference between the tip of the bud and the leaf blade; (3) As shown in FIG. 7, using a hyperspectral imaging system to detect the cellulose content of the sprouts recognized by the thermography 57, confirming the recognition accuracy of the thermography based on the cellulose content, and performing step (4) after the sprouts are detected as sprouts by the hyperspectral imaging system 58; (4) recognizing and locating the sprouts using a depth camera 56; (5) the control cabinet 61 sends a signal to the manipulator 6 and the servo 52 to place the manipulator 6 in a predetermined position and align the laser generator 59 with the picking point on the sprout, and then the servo 52 operates to place the suction tube 51 in a predetermined position; (6) As shown in FIG. 8, the laser generator 59 operates to cut the sprouts, and at the same time, as shown in FIG. 9, the suction tube 51 operates to suck the cut sprouts into the collection box 62, completing the picking of the sprouts 7, and then resetting the suction tube 51 and the manipulator 6; (7) A method for picking new tea buds based on multimodal recognition using infrared spectra, the method including a step of continuing to perform steps (2) to (6) by causing the tea picking machine to continue moving forward using a traveling mechanism until all tea picking work is completed.
[0019] The beneficial effects of the present invention are as follows: The air in the fan is heated to a constant temperature of 40°C by an electric heating wire, and then blown onto the tea leaves of the tea tree, so that on the one hand, the tip of the bud and the leaf blade of the tea tree are heated to form a temperature difference, and on the other hand, hot air is blown onto the tea leaves to shake the tea leaves and blow off the dirt on the tea leaves. The present invention first recognizes the buds of the heated tea leaves by thermography, and then uses a hyperspectral imaging system to detect the cellulose content of the tea leaf buds, and recognizes and determines the buds by double detection. The present invention combines deep learning to recognize and position the picking point of the fresh leaf buds, and cuts them with a laser generator, and the cutting is accurate. The present invention can move the recognition and picking assembly according to the height of the tea tree leaves, and recognize and pick the fresh leaf buds of tea trees at different crown layer heights. [Explanation of symbols]
[0020] In the figure: 1 traveling mechanism, 11 traveling bracket, 12 traveling wheel, 13 laser radar, 14 traveling drive motor, 2 frame, 3 fan assembly, 31 fan, 32 fan drive motor, 33 motor mounting frame, 34 heating wire, 35 temperature sensor, 36 air guide case, 4 slide rail assembly, 46 first drive motor, 42 vertical slide rail, 43 vertical slider, 44 lateral slider, 45 lateral slide rail, 41 second drive motor, 5 recognition picking assembly, 51 suction tube, 52 servo, 53 fixed frame, 54 lead screw guide rail, 55 suction tube locking device, 56 depth camera, 57 thermography, 58 hyperspectral imaging system, 59 laser generator, 6 manipulator, 61 control cabinet, 62 collection box, 7 sprout.
Claims
1. A new tea bud picking machine based on multimodal recognition by infrared spectrum, comprising a frame, a fan assembly, a slide rail assembly, a recognition picking assembly and a manipulator; The bottom of the frame has a running mechanism, the fan assembly includes a fan drive motor, a fan, a heating wire, a temperature sensor, and an air guide case, the fan and the fan drive motor are fixed to a top of a frame, the fan drive motor is used to drive and operate the fan, the air guide case is located at an air outlet of the fan, the heating wire and the temperature sensor are located in the air guide case, the heating wire is used to heat the air in the air guide case, and the temperature sensor is used to detect the air temperature at the outlet of the air guide case, The slide rail assembly is located at the top of the frame, and the recognition and picking assembly includes a suction tube, a fixed frame, and a manipulator. The fixed frame is located on the slide rail assembly and drives the vertical and / or lateral movement of the fixed frame via the slide rail assembly. The fixed frame has a lead screw guide rail, a depth camera, a thermography, and a hyperspectral imaging system. The thermography recognizes sprouts through a thermography image, and the hyperspectral imaging system is used to detect the cellulose content of the object recognized by the thermography to confirm whether it is a sprout. The depth camera recognizes and positions the sprouts recognized by the hyperspectral imaging system. The intersection of the lines of sight of the thermography, the depth camera, and the hyperspectral imaging system is a recognition area. One end of the suction tube is connected to a collection box located on the frame, and the other end of the suction tube is fixed to a positioning slider of the lead screw guide rail, and the other end of the suction tube moves vertically along with the positioning slider. The manipulator is fixed to a frame, the manipulator has a laser generator, and the frame further has a control cabinet, the control cabinet is signal-connected to the traveling mechanism, the fan assembly, the slide rail assembly, the recognition picking assembly and the manipulator, and controls the operating state of the entire tea picking machine.
2. The tea bud picking machine based on multimodal recognition by infrared spectrum as described in claim 1, characterized in that the traveling mechanism includes a traveling bracket, the traveling bracket has a traveling wheel and a traveling drive motor for driving the traveling wheel to operate, and the front end of the traveling bracket has two laser radars distributed on the left and right and installed diagonally downward, and the laser radars are used for map creation, positioning and route planning.
3. The tea bud picking machine based on multi-modal recognition by infrared spectrum, as described in claim 1, characterized in that the frame has two motor mounting frames, each of the motor mounting frames has one fan driving motor, each of the fan driving motors has one fan on its output shaft, the air outlets of the two fans are connected to the upper end of the air guide case, and the lower end of the air guide case faces the recognition area.
4. 2. The tea bud picking machine based on multimodal recognition by infrared spectrum according to claim 1, wherein the slide rail assembly includes a vertical slide rail, a lateral slide rail, a vertical slider, and a lateral slider, the lateral slide rail is fixed to a top of a frame, the lateral slider is slidably installed on the lateral slide rail, the vertical slide rail is fixed to the lateral slider and installed perpendicular to the lateral slide rail, the vertical slide rail is slidably installed on the vertical slide rail, the lateral slide rail has a first drive motor for driving and moving the lateral slider, the vertical slide rail has a second drive motor for driving and moving the vertical slider, and the fixed frame is fixedly connected to the vertical slider.
5. The tea bud picking machine based on multimodal recognition by infrared spectrum as claimed in claim 1, characterized in that one end of the suction tube is fixed to the positioning slider via a suction tube fastener.
6. The tea bud picking machine based on multi-modal recognition by infrared spectrum as claimed in claim 1, characterized in that the cross-sectional dimension of the upper end of the air guide case is larger than the cross-sectional dimension of the lower end of the air guide case.
7. (1) A laser radar performs mapping, positioning and route planning, and the whole tea picking machine is advanced according to the planned route by a traveling mechanism; (2) the fan assembly is operated to blow hot air at a constant temperature of 40 degrees Celsius onto the leaf blade at the tip of the tea plant bud through the air guide case, to blow off dirt on the leaf blade at the tip of the bud, and to recognize the new buds by using thermography based on the heating temperature difference between the tip of the bud and the leaf blade; (3) using a hyperspectral imaging system to detect the cellulose content of the sprouts recognized by thermography, and confirming the recognition accuracy of the thermography based on the cellulose content; and performing step (4) after the sprouts are detected as sprouts by the hyperspectral imaging system; (4) recognizing and locating the sprouts using a depth camera; (5) placing the manipulator in position, aligning the laser generator with the picking point on the sprout, and then placing the suction tube in position; (6) the laser generator operates to cut the sprouts, and at the same time, the suction tube operates to suck the cut sprouts into the collection box, completing the picking of the sprouts, and then resetting the suction tube and the manipulator; (7) a step of continuing to perform steps (2) to (6) of the tea picking machine continuing to move forward by the traveling mechanism until all tea picking work is completed.
Citation Information
Patent Citations
Tea tree tender shoot identification method based on convolutional neural network
CN112131982A
Deep learning-based tea leaf bud tip detection method for tea-leaf picker
CN113674226A
Tea picking method based on image and point cloud data processing
CN114190166A
Picking and sorting integrated harvesting equipment for famous high-quality tea
CN118661551A
Method for gripping fruit vegetable such as harvester for fruit vegetables
JP1994141649A
Cited By
Fruit and vegetable picking robot system based on data enhancement and identification precision optimization
CN120167229A
Picking path generation method and device of tea leaf picking robot
CN120663312A
Tea-leaf picker with posture regulation and control function and picking method
CN121014378A
Integrated harvesting device for picking and sorting famous tea
US20260053093A1
