Computer vision-based agalloch grading device and method

Abstract

The application discloses a kind of based on computer vision's eaglewood grading processing device and method, including the image data of the cross section of two sides of the eaglewood section to be processed is collected, based on image processing technology obtains the eaglewood area above clamping component, the maximum value of the height of two sides is taken as cutting parameter, and the overall cutting above the height is carried out to the eaglewood section raw material;Rotating processing piece, repeating the above operation, realize first grade rough machining;In the second stage, based on computer vision technology detects processing piece, utilizes the hooking operation component perpendicular to eaglewood section and carries out translation operation, and hooks silk to white wood area, after the current angle is completed, according to the fixed angle value set, processing piece is rotated, and the above operation is repeated until the processing of eaglewood section is completed.The application comprehensively utilizes computer vision, automation and mechanical design, and carries out the automatic grading processing of the eaglewood section to be processed, and can effectively improve the processing efficiency of eaglewood section cutting and hooking.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of automatic processing device and method of eaglewood, and particularly relates to a computer vision-based eaglewood grading processing device and method. BACKGROUND

[0002] Eaglewood is a kind of wood rich in resin produced by Aquilaria Lam. and Gyrinops Gaertn. plants due to natural or artificial factors.

[0003] Eaglewood primary processing is a core problem in the secondary industry of eaglewood, and how to reduce the processing links, improve the efficiency and reduce the cost is a fundamental problem to be solved. Therefore, the research and development of automatic and intelligent equipment for eaglewood primary processing is very important. At present, with the continuous expansion of artificial planting of eaglewood trees, a large number of artificial eaglewood trees will need to be processed in the future. However, the primary processing of eaglewood still remains in the stage of manual processing, which is high in cost, low in efficiency and high in danger. In addition to simple tools or equipment such as electric saw, table saw, semi-automatic cutting knife, chopping knife, hook knife and shaving knife, there is a lack of efficient intelligent and automatic processing equipment.

[0004] In recent years, with the prominence of primary processing issues, there has been a significant increase in the filing of patents for agarwood processing equipment aimed at improving efficiency and reducing costs. However, despite the existence of some related utility model and invention patents, they are difficult to implement and promote due to various reasons. For example, patent CN117207294A discloses an automatic agarwood hooking processing equipment, which uses a three-axis linkage sliding table, a depth camera, a detection unit, a main control unit, multiple servo motors, a shovel knife, a thimble, etc. It uses a deep learning method to automatically identify the image of the agarwood material to be processed, which requires manual annotation of the position coordinate information of the white wood area after image acquisition. At the same time, the CNN parameters need to be constantly adjusted during the training model process. The overall design is relatively complex, and it has a high dependence on training data, which will result in high human and material costs in practice. Moreover, this method does not perform cutting during the rough processing of raw materials, but directly performs hooking, which will result in low work efficiency. Another example is patent CN213732315U, which discloses an agarwood and huangxiangmu processing equipment that realizes sawing, breaking, and fine planing of wood segments. However, it still requires manual identification of agarwood and white wood, and cannot achieve intelligent identification and automatic breaking and cutting. CN220362714U discloses a cutting device for agarwood processing, CN220661192U discloses an agarwood raw material cutting device, and CN218534921U discloses an agarwood processing cutting device. All of these are suitable for cutting larger agarwood, focusing on solving the stability of clamps, the safety of processing, and the cleaning of blades. However, they cannot achieve further cutting, hooking, and separating white wood and agarwood layers, and are even more difficult to automate and intelligentize. SUMMARY

[0005] The purpose of the present application is to provide an agarwood grading and processing device based on computer vision.

[0006] The purpose of the present application is also to provide an agarwood grading and processing method based on the above-mentioned device.

[0007] The first object of the present application can be achieved by the following technical solution: a computer vision-based agarwood grading processing device, comprising a base provided with an embedded computing system for image processing and motor driving, a lower support provided in the middle of the base, left and right supports provided on the two sides of the base, and an upper support provided above the left and right supports, wherein the lower support is provided with a lower support motor and a lower support sliding rail, the lower support sliding rail is provided with a cutting telescopic mechanism movable on the lower support sliding rail, the top of the cutting telescopic mechanism is provided with a cutting component, the left and right supports are respectively provided with clamping components for clamping the to-be-processed agarwood segments in the opposite positions, the clamping component of the left support is provided above with a first industrial camera for shooting the first cross section of the to-be-processed agarwood segment, and the clamping component of the right support is provided above at a corresponding position with a second industrial camera for shooting the second cross section of the to-be-processed agarwood segment, the upper support is provided with an upper support motor and an upper support sliding rail, the upper support sliding rail is provided with a hook knife telescopic mechanism movable on the upper support sliding rail, the top of the hook knife telescopic mechanism is provided with a hook knife working component, and the upper support is provided with a third industrial camera opposite the position of the to-be-processed agarwood segment.

[0008] Optionally, the cutting telescopic mechanism is a cutting telescopic rod.

[0009] Optionally, the hook knife telescopic mechanism is a hook knife telescopic rod.

[0010] Optionally, the cutting component is a gear cutting tool.

[0011] Optionally, the hook knife working component comprises a hook knife connecting rod and a hook knife head.

[0012] Optionally, the width of the hook knife head of the hook knife working component is approximately 2-5 mm.

[0013] The second object of the present application can be achieved by the following technical solution: an agarwood grading processing method based on the above device, comprising the following steps:

[0014] S1, using the camera of the first industrial camera and the second industrial camera to collect image information of the cross sections of the to-be-processed agarwood segments;

[0015] S2, using the embedded computing system in the base to calculate the image information obtained in S1 based on image processing technology to obtain cutting parameters;

[0016] S3, transmitting the cutting parameters to the cutting component to drive the cutting component to cut the to-be-processed agarwood segment;

[0017] S4, rotating the to-be-processed agarwood segment by a fixed angle by using the clamping component, and repeating steps S1-S3 to realize the first-stage rough machining.

[0018] S5, the camera of the third industrial camera with the upper support collects image data of the material after rough machining, and the hooking operation area Q is marked according to the width of the hooking operation part w ;

[0019] S6, the hooking operation area Q is identified based on computer vision technology w ; b According to the white wood area Q b , the hooking operation parameters are transmitted to the hooking operation part;

[0020] S7, according to the preset hooking sinking parameter g, the hooking position is lowered from the top position of the white wood area Q b , and the white wood area Q b is operated by translation to realize hooking of the white wood area;

[0021] S8, after the hooking operation is performed, the step S7 is repeated until the white wood area is not detected in the current hooking operation area;

[0022] S9, the workpiece is rotated by a fixed angle, and the steps S5-S8 are repeated to realize the second level of fine machining.

[0023] In the above-mentioned agalloch classification processing method:

[0024] Optionally, the image processing and motor driving in steps S1-S9 are controlled by an embedded computing system in the base, the embedded computing system is a lightweight embedded system with ARM architecture, and an image processing module for processing image information of the two side cross sections of the agalloch wood segment to be processed is carried. The operation parameters required in the calculation process are transmitted to the motor for controlling each operation part (such as the lower support sliding rail, the upper support sliding rail, the cutting telescopic mechanism, the hooking telescopic mechanism, etc.) through a bus, and each operation part is driven to work.

[0025] Optionally, the embedded computing system in the base is used in step S2 to calculate the image information obtained in S1 based on image processing technology to obtain the cutting parameters, and the specific steps include:

[0026] S21, the image of one side cross section of the agalloch wood segment to be processed is processed to obtain the distance L1 between the highest part of the agalloch and the clamping part;

[0027] S21, the image of the other side cross section of the agalloch wood segment to be processed is processed to obtain the distance L2 between the highest part of the agalloch and the clamping part;

[0028] S23, the cutting parameters are calculated: L = Max (L1, L2);

[0029] The image processing technology mentioned therein includes identifying the agarwood part in the image information based on the threshold segmentation method in the CV image processing library, then using the edge detection method to calculate the edge position of the agarwood part, and calculating the cutting parameters according to the edge position parameters.

[0030] Optionally, in step S3, the cutting parameters are transmitted to the cutting component, driving the cutting component to cut the agarwood segment to be processed. Specific steps include:

[0031] S31. Based on the cutting parameter L, use the lower support motor to drive the height of the cutting telescopic mechanism;

[0032] S32. Based on the adjusted height, the lower support motor drives the cutting telescopic mechanism to move horizontally via the lower support slide rail, while the cutting component rotates to complete the cutting of the agarwood segment.

[0033] Optionally, in steps S4 and S8, the angle at which the agarwood segment to be processed is rotated and fixed using a clamping component is 40 to 60°.

[0034] Optionally, in step S5, the camera of the third industrial camera on the upper bracket is used to acquire image data of the material after rough processing, and the hook knife working area Q is marked according to the width of the hook knife working component. w Specifically, this includes: using the camera of a third industrial camera mounted on the upper bracket to capture a top-view image of the material after rough machining; and marking the hook-cutting working area Q according to the width of the hook-cutting working component. w The width of the hook-shaped working component is between 2 and 5 mm.

[0035] Optionally, in step S6, the white wood region Q is identified based on computer vision technology. b According to the white wood region Q b The hook cutter operation parameters are transmitted to the hook cutter operation component, specifically including:

[0036] S61. Identify the hook knife working area Q using computer vision technology. w The white wood area in the image is marked as white wood area Q. b ;

[0037] S62. Based on the identified white wood area Q b Calculate the hook cutter's operating parameters and transmit them to the hook cutter's operating component;

[0038] The computer vision technology mentioned includes using a convolutional neural network method to extract the hook knife working area Q. w Agarwood region and white wood region Q bimage features, mainly color features and texture features, and further trains and learns through a lightweight convolutional neural network model EfficientNet with an attention mechanism to construct a binary classification model of the agarwood region and the white wood region Q b in the hooking operation region Q w . b .

[0039] Optionally, according to the hooking requirement, the hooking position is lowered from the top position of the white wood region Q b according to the preset hooking sinking parameter g, and the white wood region Q b is subjected to a translation operation to realize hooking of the white wood region, and the specific steps include:

[0040] S71, according to the obtained hooking operation parameter, the hooking operation component is aligned with the edge position on the right side of the white wood region Q b using the upper support motor and the upper support sliding rail in the vertical direction;

[0041] S72, the hooking operation component is lowered to the current highest position of the agarwood section by using the hooking extension mechanism, and is further lowered according to the hooking sinking parameter g;

[0042] S73, the white wood region Q b is subjected to a hooking operation by using the hooking operation component;

[0043] S74, when there are multiple white wood regions Q b in step S52, after the hooking of each white wood region is completed, the hooking operation component is raised to the current highest position of the agarwood section, and is then translated to the edge on the right side of the next white wood region Q b for lowering and hooking operation.

[0044] Optionally, the hooking sinking parameter g is 0.05-0.1 mm.

[0045] The present application has the following advantages: the device and method of the present application comprehensively utilize computer vision, automation and mechanical design, and can realize automatic classification processing of the agarwood section raw material to be processed, and can effectively improve the processing efficiency of the agarwood section cutting and hooking. BRIEF DESCRIPTION OF DRAWINGS

[0046] The specific embodiments of the present application will be described in detail below with reference to the accompanying drawings:

[0047] Figure 1 is a front view of the agarwood classification processing device based on computer vision in embodiment 1 of the present application;

[0048] Figure 2Cutting part used for the first stage rough processing in embodiment 1 of the present application

[0049] Figure 3 Cross section of one side of the lignum wood section raw material to be processed in embodiment 2 of the present application

[0050] Figure 4 Cross section image of one side of the lignum wood section raw material to be processed in embodiment 2 of the present application after processing by the computer vision recognition

[0051] Figure 5 Cross section image of the lignum wood section raw material after cutting in embodiment 2 of the present application according to step S3

[0052] Figure 6 Cross section image of the lignum wood section raw material after cutting in embodiment 2 of the present application according to step S4 Figure 4

[0053] Figure 7 Schematic diagram of the hooking operation part in embodiment 1 of the present application

[0054] Figure 8 Hooking operation area processing based on computer vision recognition in embodiment 2 of the present application DETAILED DESCRIPTION

[0055] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of protection of the present application.

[0056] As Figure 1 ​As shown, the computer vision-based agarwood grading device provided in the embodiment includes a base 1 provided with an embedded computing system for image processing and motor driving, a lower support 2 provided in the middle of the base 1, left and right supports 3 and 4 provided on the two sides of the base 1, and an upper support 5 provided above the left and right supports 3 and 4. The lower support 2 is provided with a lower support motor and a lower support sliding rail (conventional design, not shown in the figure), and the lower support sliding rail is provided with a cutting telescopic mechanism 6 movable on the lower support sliding rail. The cutting telescopic mechanism 6 is provided with a cutting component 7 at the top. The left and right supports 3 and 4 are respectively provided with clamping components 8 for clamping the to-be-processed agarwood segments 14 at the opposite positions. The clamping component 8 of the left support 3 is provided with a first industrial camera 9 above for shooting the first cross section of the to-be-processed agarwood segment. The clamping component 8 of the right support 4 is provided with a second industrial camera 10 at the corresponding position above for shooting the second cross section of the to-be-processed agarwood segment. The upper support 5 is provided with an upper support motor and an upper support sliding rail (conventional design, not shown in the figure), and the upper support sliding rail is provided with a hook knife telescopic mechanism 11 movable on the upper support sliding rail. The hook knife telescopic mechanism 11 is provided with a hook knife working component 12 at the top. The upper support 5 is provided with a third industrial camera 13 opposite to the position of the to-be-processed agarwood segment.

[0057] Wherein:

[0058] The embedded computing system is a conventional ARM architecture lightweight embedded system, which is provided with an image processing module for image processing of the cross sections of the two sides of the to-be-processed agarwood segment raw material. The operation parameters required in the calculation process are transmitted to the motors for controlling various operation components through a bus, so as to drive the operation components to work.

[0059] The cutting telescopic mechanism 6 is a cutting telescopic rod.

[0060] The hook knife telescopic mechanism 11 is a hook knife telescopic rod.

[0061] The cutting component 7 is a gear cutting tool, and its structure is shown in Figure 2 .

[0062] The structure of the hook knife working component 12 is shown in Figure 7 , which includes a hook knife connecting rod 121 and a hook knife head 122.

[0063] The width of the hook knife head 122 of the hook knife working component 12 is about 2-5 mm.

[0064] Embodiment 2

[0065] The agarwood grading method based on the device in Embodiment 1 includes the following steps:

[0066] S1. Industrial cameras, specifically the first industrial camera 9 and the second industrial camera 10, acquire image information of the cross-sections on both sides of a fixed-length (e.g., 100mm) piece of agarwood to be processed. Figure 3 As shown;

[0067] S2. Using the embedded computing system in base 1, the image information obtained in S1 is calculated based on image processing technology to obtain cutting parameters;

[0068] S3. The cutting parameters are transmitted to the cutting component 7, and the cutting component 7 is driven to cut the agarwood section raw material.

[0069] S4. Use the clamping component 8 to rotate and fix the workpiece by an angle, and repeat steps S1-S3 to achieve the first stage of rough machining.

[0070] S5. Obtain image data of the rough-processed material from the industrial camera 13 of the upper bracket 5, and mark the hook knife working area Q according to the width of the hook knife working component. w ;

[0071] S6. Identifying the hook cutter's working area Q based on computer vision technology. w White wood area Q in b According to the white wood region Q b Transmit the hook cutter's operating parameters to the hook cutter component;

[0072] S7. Based on the hook wire requirements and according to the preset hook knife sinking parameter g, from the white wood area Q... b The top position will lower the hook knife position, targeting the white wood area Q. b A translation operation is used to create the lace pattern on the white wood area;

[0073] S8. After performing the hooking operation, repeat step S7 until no white wood area can be detected in the current hooking knife working area.

[0074] S9. Rotate the workpiece to a fixed angle and repeat steps S5-S8 to achieve the second level of fine machining.

[0075] In Example 1 and the method of this example, the image processing and motor drive are both controlled by the embedded computing system in the base.

[0076] The image processing and motor drive described in steps S1-S9 are controlled by the embedded computing system in the base. The embedded computing system is a lightweight embedded system using a conventional ARM architecture and is equipped with an image processing module. The operation parameters required during the calculation process are transmitted to the motors controlling each operation component through the bus, thereby driving each operation component to perform the operation.

[0077] In step S2, the embedded computing system in the base is used to calculate the image information based on image processing technology to obtain the cutting parameters. The specific steps include:

[0078] S21, such as Figure 4 As shown, the cross-sectional image of one side of the agarwood section to be processed is processed to obtain the distance L1 between the highest point of the agarwood and the clamping component;

[0079] S21. Process the cross-sectional image of the other side of the agarwood section to be processed to obtain the distance L2 between the highest point of the agarwood and the clamping component;

[0080] S23. Calculate the cutting parameters: L = Max(L1, L2);

[0081] The image processing technology mentioned therein includes identifying the agarwood part in the image information based on the threshold segmentation method in the CV image processing library, then using the edge detection method to calculate the edge position of the agarwood part, and calculating the cutting parameters according to the edge position parameters.

[0082] In step S3, the cutting parameters are transmitted to the cutting component to cut the agarwood segment to be processed. The specific steps include:

[0083] S31. Based on the cutting parameter L, use the lower support motor to drive the height of the cutting telescopic rod;

[0084] S32. Based on the adjusted height, the lower support motor drives the cutting telescopic rod to move horizontally via the lower support slide rail, while the cutting component rotates simultaneously to complete the cutting of the agarwood segment. The cross-section of one side after cutting is as follows: Figure 5 As shown.

[0085] In steps S4 and S8, the agarwood segment to be processed is rotated and fixed by the clamping component at an angle of 40° to 60°. In this embodiment, 60° is recommended.

[0086] According to step S4 Figure 4 The cross-sectional image of the agarwood log after it has been rotated once and then cut is shown below. Figure 6 As shown.

[0087] In step S5, the camera of the third industrial camera on the upper bracket is used to acquire image data of the material after rough machining, and the hook knife working area Q is marked according to the width of the hook knife working part. w Specifically, this includes: using the camera of a third industrial camera mounted on the upper bracket to capture a top-view image of the material after rough machining; and marking the hook-cutting working area Q according to the width of the hook-cutting working component. w The width of the hook blade's working part is between 2 and 5 mm; mark the hook blade's working area as follows: Figure 8 The dashed box part Q w .

[0088] In step S6, the white wood region Q is identified based on computer vision technology. b According to the white wood region Q b The hook cutter operation parameters are transmitted to the hook cutter operation component, specifically including:

[0089] S61. Identify the hook knife working area Q using computer vision technology. w The white wood area in the image is marked as white wood area Q. b In reality, there may be multiple white wood areas;

[0090] S62. Based on the identified white wood area Q b Calculate the hook cutter's operating parameters and transmit them to the hook cutter's operating component;

[0091] The computer vision technology mentioned includes using a convolutional neural network method to extract the hook knife working area Q. w Central Agarwood Region ( Figure 8 Based on the differences in color and texture between the agarwood region and the white wood region (the shaded area), color and texture features have the highest weights among the extracted image features. Furthermore, a lightweight convolutional neural network model, EfficientNet, incorporating an attention mechanism, is used to train and learn a binary classification model for the agarwood region and the white wood region Qb, thus identifying the hook-knife operation area Q. w White wood area Q in b .

[0092] Among them, the white wood area Q b The hook cutter operating area Q w The agarwood region in Central Africa, the agarwood region is Figure 8 The shaded area.

[0093] Hook cutter working area Q w The width is one of the operating parameters of the hook cutter, that is, the width of the hook cutter's operating component, which is roughly between 2-5mm.

[0094] In step S7, according to the hook wire requirements, the hook knife is lowered from the white wood area Q according to the preset hook knife sinking parameter g. b The top position will lower the hook knife position, targeting the white wood area Q. b The process of using a translation operation to create wire mesh on the white wood area includes the following steps:

[0095] S71. Based on the obtained hook knife operating parameters, use the upper bracket motor and slide rail to align the hook knife operating component vertically with the white wood area Q. b The right edge position;

[0096] S72, the hook knife operation part is lowered to the current highest position of the agarwood section by using the hook knife telescopic rod, and is further lowered according to the hook knife sinking parameter g (such as 0.1 mm);

[0097] S73, according to Figure 8 The hook knife operation part is lowered to the current highest position of the agarwood section by using the hook knife telescopic rod, and is further lowered according to the hook knife sinking parameter g (such as 0.1 mm);

[0098] S74, if there are multiple white wood regions in step S52, after the hooking of each white wood region is completed, the hook knife operation part is raised to the current highest position of the agarwood section, and is then moved to the right edge of the next white wood region to be lowered and to be hooked.

[0099] In summary, the method of the present application comprises collecting image data of the cross sections of both sides of the fixed-length agarwood section to be processed, obtaining the agarwood region above the fixed axis based on computer vision technology, and calculating the height of the highest position of the region; taking the maximum value of the heights of both sides as a cutting parameter, and cutting the agarwood section above the height by using a rotary cutter according to the parameter; rotating the processed piece according to the set fixed angle value, and repeating the above operations to achieve the first-stage rough machining; in the second stage, a hook perpendicular to the rough-machined agarwood section is used for translation operation, and hooking operation is performed to achieve more fine machining, the hook is gradually sunk according to the fixed step length set in the vertical direction, and the processed piece is detected based on computer vision technology, the white wood is hooked, and the part reaching the dark color threshold value is considered as agarwood, then the hooking of the region is stopped, after the current angle is completed, the processed piece is rotated according to the set fixed angle value, and the above operations are repeated until the processing of the agarwood section is completed.

[0100] Although the embodiments of the present application have been shown and described, it can be understood by those skilled in the art that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirits of the present application, the scope of the present application is defined by the appended claims and their equivalents, and all should be included in the protection scope of the present application.

Claims

1. A computer vision-based agarwood grading device and method, characterized in that, The computer vision-based agarwood grading device comprises a base provided with an embedded computing system for image processing and motor driving, a lower support in the middle of the base, left and right supports opposite to the two sides of the base, and an upper support above the left and right supports. The lower support is provided with a lower support motor and a lower support sliding rail, and the lower support sliding rail is provided with a cutting telescopic mechanism movable thereon. The cutting telescopic mechanism is provided with a cutting component at the top. The left and right supports are respectively provided with clamping components for clamping the agarwood segments to be processed at the opposite positions. The clamping component of the left support is provided with a first industrial camera above for shooting the first cross section of the agarwood segment to be processed. The clamping component of the right support is provided with a second industrial camera above at the corresponding position for shooting the second cross section of the agarwood segment to be processed. The upper support is provided with an upper support motor and an upper support sliding rail, and the upper support sliding rail is provided with a hook knife telescopic mechanism movable thereon. The hook knife telescopic mechanism is provided with a hook knife working component at the top. The upper support is provided with a third industrial camera opposite to the position of the agarwood segment to be processed. The computer vision-based agarwood grading method comprises the following steps: S1, using the camera of the first and second industrial cameras to collect the image information of the two cross sections of the agarwood segment to be processed; S2, using the embedded computing system in the base to calculate the image information obtained in S1 based on image processing technology to obtain cutting parameters; S3, transmitting the cutting parameters to the cutting component to drive the cutting component to cut the agarwood segment to be processed; S4, rotating the agarwood segment to be processed by a fixed angle by using the clamping component, and repeating steps S1-S3 to realize the first-stage rough machining; S5, the camera of the third industrial camera with the upper support collects image data of the material after rough machining, and marks the hooking operation area Q according to the width of the hooking operation part w ; S6. Identifying the hook cutter's working area Q based on computer vision technology. w White wood area Q in b According to the white wood region Q b Transmit the hook cutter operation parameters to the hook cutter operation component; S7、According to the hooking requirement, the hooking cutter position is lowered from the white wood area Q b top position according to the preset hooking cutter sinking parameter g, and the hooking cutter position is lowered to the white wood area Q b The translation operation is adopted to realize the hooking of the white wood area. S8, after performing the hooking operation, repeating step S7 until no white wood area is detected in the current hook knife working area; S9, rotating the processed piece by a fixed angle, repeating steps S5-S8 to realize the second-stage fine machining.

2. The method of classifying agarwood according to claim 1, wherein, The image processing and motor driving in steps S1-S9 are controlled by the embedded computing system in the base. The embedded computing system is a lightweight embedded system with ARM architecture, which is loaded with an image processing module for processing the image information of the two cross sections of the agarwood segment to be processed. The required working parameters in the calculation process are transmitted to the motors controlling the various working components through a bus to drive the working components to work.

3. The method of classifying agarwood according to claim 2, wherein, In step S2, the embedded computing system in the base is used to calculate the image information obtained in S1 based on image processing technology to obtain cutting parameters. The specific steps include: S21, processing the image of one side cross section of the agarwood segment to be processed to obtain the distance L1 between the highest part of the agarwood and the clamping component; S21, processing the image of the other side cross section of the agarwood segment to be processed to obtain the distance L2 between the highest part of the agarwood and the clamping component; S23, calculating the cutting parameters: L = Max (L1, L2); The image processing technology comprises: identifying the agarwood part in the image information based on a threshold segmentation method in a CV image processing library, and calculating the edge position of the agarwood part by using an edge detection method, and calculating the cutting parameters according to the edge position parameters.

4. The method of classifying agarwood according to claim 3, wherein, The cutting parameters are transmitted to the cutting component in the step S3 to drive the cutting component to cut the agarwood section to be processed, and the specific steps comprise: S31, driving the height of the cutting telescopic mechanism by the lower support motor according to the cutting parameter L; S32, driving the cutting telescopic mechanism to translate by the lower support motor through the lower support slide rail according to the adjusted height, and simultaneously rotating the cutting component to complete the cutting of the agarwood section.

5. The method of grading agarwood based on computer vision as claimed in claim 1, wherein, The angle of the agarwood section to be processed is fixed by the clamping component in the steps S4 and S8, and the fixed angle is 40-60°.

6. The method of grading agarwood based on computer vision as claimed in claim 1, wherein, The camera of the third industrial camera of the upper support in the step S5 collects image data of the material after rough machining, and marks a hooking operation area Q according to the width of the hooking operation part w , specifically comprising: collecting an overhead view image of the material after rough machining by the camera of the third industrial camera of the upper support, and marking a hooking operation area Q according to the width of the hooking operation part w , the width of the hooking operation part is between 2-5mm.

7. The method of grading agarwood based on computer vision as claimed in claim 1, wherein, The step S6 identifies the white wood region Q in the hooking and cutting operation region Q based on computer vision technology w b According to the white wood region Q b , the hooking and cutting operation parameters are transmitted to the hooking and cutting operation component, specifically including:​ S61, identify the hooking and cutting operation area Q according to computer vision technology w The white wood area in the image is marked as white wood area Q b ; S62、According to the identified white wood area Q b Calculate the hooking operation parameters and transmit the hooking operation parameters to the hooking operation component; Wherein the computer vision technology includes extracting the hooking operation area Q w using a convolutional neural network method b The image features of the eaglewood area and the white wood area Q b are mainly color features and texture features. On this basis, a lightweight convolutional neural network model EfficientNet with attention mechanism is trained and learned to construct a binary classification model of the eaglewood area and the white wood area Q w , so as to identify the white wood area Q b in the hooking operation area Q .

8. The method of grading agarwood based on computer vision as claimed in claim 1, wherein, According to the gulleting requirement, the gullet position is lowered from the top position of the white wood area Q b according to the preset gullet sinking parameter g in step S7 b The gulleting of the white wood area is realized by the translation operation, and the specific steps include: S71、According to the obtained hooking operation parameters, the upper support motor and the upper support sliding rail are used to align the hooking operation component in the vertical direction to the white wood area Q b Right edge position; S72, lowering the hook knife operation component to the current highest position of the agarwood section by the hook knife telescopic mechanism, and further lowering according to the hook knife sinking parameter g; S73, adopt hooking operation component to white wood area Q b Carry out hooking operation; S74, when there are multiple white wood regions Q in step S52 b After the hooking of each white wood region is completed, the hooking tool part is raised to the current highest position of the rosewood section, and then moved to the right edge of the next white wood region Q b to perform the lowering and hooking operations.

9. The method of classifying agarwood according to claim 8, wherein, The hook knife sinking parameter g is 0.05-0.1 mm.

Images