Wood sorting device based on machine vision recognition

Abstract

The utility model discloses a wood sorting device based on machine vision identification, which comprises a vision identification platform mechanism, a conveying mechanism, a mechanical arm, a chassis mechanism and a working chamber, the vision identification platform mechanism, the conveying mechanism, the mechanical arm and the chassis mechanism are arranged in the working chamber, and the vision identification platform mechanism and the mechanical arm are arranged on the chassis mechanism; the visual recognition platform mechanism comprises a main stand column, a telescopic stand column, an overhanging beam, a camera and a lamp strip support, the telescopic stand column can stretch out and draw back along the main stand column, the telescopic stand column is connected with the overhanging beam, the overhanging beam is connected with the camera, and the lamp strip support is connected with the main stand column and the chassis mechanism. According to the utility model, the stability of visual detection is good, the whole visual identification platform mechanism is arranged in a working room, the illumination condition is not changed, unmanned conveying of black lamps is realized, the identification accuracy is greatly improved, and the influence of an external light source on identification is avoided; conveying is stable, efficiency is higher, the illumination range is more comprehensive, and the rejection rate is reduced.

Description

Technical Field

[0001] The utility model relates to a sorting device, in particular to a wood sorting device based on machine vision recognition. Background Art

[0002] There are various types of wood with different colors, and the uses of different materials vary greatly. Therefore, the classification of wood colors is particularly important. From the perspective of human eye perception, color is a physical quantity that includes three variables: hue, lightness, and saturation. It reaches the human eye after the interaction between light and the surrounding environment, and through a series of physical and chemical changes, it is converted into electrical pulses that the human eye can perceive, thus obtaining a conclusion about a certain color. Therefore, the formation of color in the human eye is a complex process of interaction between physics and psychology, which involves the propagation characteristics of light, the structure of the human eye, and the psychological perception of the human brain.

[0003] Currently, in the wood processing industry, workers often need to judge colors by naked eyes for classification. The method of manually classifying wood has a large workload, low efficiency, is prone to causing resource waste, and cannot meet the development trend of contemporary automation and intelligence as well as the increasing production demand. The cameras of existing sorting devices cannot adjust their positions, and changes in lighting conditions will result in inaccurate images obtained by the cameras, affecting the accurate classification of woods with similar colors. Content of the Utility Model

[0004] Object of the Invention: In order to overcome the deficiencies of the prior art, the object of the present utility model is to provide a wood sorting device based on machine vision recognition with stable transportation, high classification efficiency, and high sorting accuracy.

[0005] Technical Solution: A wood sorting device based on machine vision recognition described in the present utility model includes a vision recognition platform mechanism, a conveying mechanism, a robotic arm, a chassis mechanism, and a working chamber. The vision recognition platform mechanism, the conveying mechanism, the robotic arm, and the chassis mechanism are arranged in the working chamber. The vision recognition platform mechanism and the robotic arm are arranged on the chassis mechanism; the vision recognition platform mechanism includes a main column, a telescopic column, an extended cross beam, a camera, and a light strip bracket. The telescopic column can extend and retract along the main column, and can adjust the height of the recognition camera by extension and retraction to find the optimal shooting height. The telescopic column is connected to the extended cross beam, the extended cross beam is connected to the camera, and the light strip bracket is respectively connected to the main column and the chassis mechanism. The light strip bracket can provide good support, and a light strip is installed on the light strip bracket as a light source for illumination.

[0006] Further, the chassis mechanism includes a working platform, a conveyor belt cross beam, and a chassis support frame. The working platforms are symmetrically arranged along the conveying direction of the conveying mechanism. The working platforms are connected to the chassis support frame, and the conveying mechanism is connected to the working platforms through the conveyor belt cross beam.

[0007] Further, the chassis support frame includes a cross beam and a base column, and the cross beam and the base column are perpendicular to each other. To further increase the stability of the working platform, the chassis support frame further includes an inclined support beam. An inclined support beam is provided at the connection of the cross beam and the base column, which can effectively disperse the vibration and force from the motor, thereby reducing the impact force on the base. It provides good support for the general assembly and enhances stability. Preferably, reinforcing plates are provided on the outer surfaces of the connections of the inclined support beam with the cross beam and the base column.

[0008] Further, the chassis mechanism further includes a motor and a motor bracket, and the motor is connected to the conveyor belt cross beam through the motor bracket.

[0009] Further, the robotic arm is arranged on the surface of the working platform and at the starting end of the conveying mechanism.

[0010] Further, the conveying mechanism includes a plurality of conveyor belts. The conveyor belts are on the same horizontal plane and are used to convey the wood raw materials from the feeding position to the visual recognition platform mechanism.

[0011] Further, the robotic arm is a six-axis robotic arm. Compared with a three-coordinate robotic arm, it is more suitable for larger-sized wood raw materials. The robotic arm is a pneumatic fixture and is precisely controlled through a controller. When the wood raw materials are conveyed to a certain area, the wood raw materials are grabbed and placed at a designated position for stacking.

[0012] Further, the visual recognition platform mechanism, the conveying mechanism, and the robotic arm are all connected to the controller.

[0013] Further, positioning sensors are provided on the conveying mechanism to ensure the stability of the position of the wood raw materials during the recognition process.

[0014] Working principle: After the wood is conveyed by the conveyor belt, its color characteristics are automatically captured by the camera. The visual recognition platform mechanism analyzes its color characteristics and converts the obtained optical signal into a digital signal. Then, the manipulator and the robotic arm, according to the corresponding digital signal, when the corresponding wood is conveyed to a specific area, grab it and move it to a designated position for stacking, thereby realizing the automatic recognition and sorting functions of the wood color depth.

[0015] Sorting method:

[0016] Step 1, the wood is placed on the conveyor belt through the feeding device and conveyed by the conveyor belt to the lower part of the camera of the visual recognition platform mechanism. The light belt on the light belt bracket is used as a light source for illumination and is automatically captured and photographed by the recognition camera.

[0017] In Step 2, after the camera automatically captures and shoots, the color characteristics of the wood are transmitted to the controller. The controller first preprocesses the image, such as grayscale conversion and median filtering, and then trains and predicts the solid wood floor image based on the BP (Back Propagation) neural network algorithm to identify the floor color. Different digital signals are used to represent the floor raw materials of different depths. For example, different depths of the floor are represented by three-bit binary numbers, where dark color is 100, medium color is 010, and light color is 001.

[0018] In Step 3, the controller transmits the digital signal to the robotic arm. After identifying the signal, the robotic arm grabs the floor raw material at the specified position and then places it into a specific bin for palletizing according to its color depth.

[0019] Beneficial effects: Compared with the prior art, the utility model has the following characteristics:

[0020] 1. Good stability of visual detection. The entire visual recognition platform mechanism is placed in the working room to ensure that the lighting conditions do not change, realizing unmanned conveying in the dark, greatly improving the recognition accuracy, avoiding the influence of external light sources on recognition, and there is a detection port in the independent working space for facilitating the detection of recognition accuracy;

[0021] 2. The conveyor belt is used to transport the wood raw materials, and more than one piece of wood can be transported simultaneously for the robotic arms on both sides to grab, sort, and palletize. The transportation is stable and the efficiency is higher;

[0022] 3. The light strip bracket is set. For large raw materials such as wooden boards with a length of more than 1m, the wood grain has a great interference with visual recognition. It can ensure that the entire floor raw material is illuminated, and the light source is provided by the light strip, and the lighting range is more comprehensive;

[0023] 4. When the wood is transported to the visual recognition platform mechanism, the camera automatically captures and shoots the image information. The visual recognition algorithm classifies the wood color into three color categories of deep, medium, and light. This process does not require manual intervention, which is beneficial to improving the sorting accuracy. Subsequently, the robotic arm will grab the wood raw material according to the recognition result and move it to the specified position for palletizing, which is beneficial to reducing the rejection rate;

[0024] 5. By introducing advanced automation and intelligent technologies, the whole process of automated operation is realized, which not only significantly improves the production efficiency, but also greatly saves the production cost. By reducing manual operations, the possibility of human errors is reduced, and at the same time, the accuracy and consistency of the operation are improved. The application of the intelligent system makes the whole sorting process more accurate and efficient, and can quickly respond to the needs of recognizing and classifying different wood color categories;

[0025] 6. The six-axis robotic arm palletizing ensures that the wood is stacked in a completely regular manner and in a larger volume, and can sort larger-sized boards, which not only meets the requirements of the production process for efficiency and accuracy, but also provides great convenience for the subsequent processing and transportation links;

[0026] 7. Using the computer vision recognition system as the host computer, the recognition signal is sent to the controller, and the entire production line is integrally linked through the controller's IO communication. The control system is simple and reliable, and the human-machine interaction is convenient. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 is a schematic structural view of the present utility model;

[0028] Figure 2 is a rear view of the present utility model;

[0029] Figure 3 is a schematic structural view inside the working chamber 5 of the present utility model;

[0030] Figure 4 is a schematic structural view of the robotic arm 3 of the present utility model;

[0031] Figure 5 is a schematic structural view of the vision recognition platform mechanism 1 of the present utility model;

[0032] Figure 6 is a schematic structural view of the chassis mechanism 4 of the present utility model;

[0033] Figure 7 is a partial enlarged view at the motor 44 of the present utility model;

[0034] Figure 8 is a partial enlarged view at the chassis support frame 43 of the present utility model. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0035] As Figures 1 to 4 , the overall dimensions of the wood sorting device based on machine vision recognition are 6050mm × 800mm × 2400mm, and it includes a vision recognition platform mechanism 1, a conveying mechanism 2, a robotic arm 3, a chassis mechanism 4, and a working chamber 5. The vision recognition platform mechanism 1, the conveying mechanism 2, the robotic arm 3, and the chassis mechanism 4 are all arranged inside the working chamber 5. The chassis mechanism 4 is provided with the vision recognition platform mechanism 1 and the robotic arm 3. The vision recognition platform mechanism 1 is provided with a detection port. The conveying mechanism 2 includes a number of conveyor belts, and the conveyor belts are on the same horizontal plane. The conveying mechanism 2 is provided with a positioning sensor. The robotic arm 3 is a six-axis manipulator, and the robotic arm 3 is connected by bolts and installed on the chassis mechanism 4 with screws, and is located on both sides of the conveying mechanism 2. The vision recognition platform mechanism 1, the conveying mechanism 2, and the robotic arm 3 are all electrically connected to the controller, and the sorting and comparison program adopts the existing color sorting program.

[0036] The computer acts as the host computer, and the camera and the transmission mechanism are controlled by the industrial robot (controller) through IO communication. Taking a certain model of FANUC industrial robot as an example, the CRMA15 / 16 communication method is used to complete the communication between the robot and the peripheral equipment in the form of hardware wiring, and the IO allocation is used to complete the association between the logical signal and the physical address. Among them, the input digital signals DI read by the robot include: 2 conveyor belt travel switch signals (recording the wooden board transmission position: camera photo position, industrial robot grab position), 3 wooden board color signals (dark, medium, light), and several safety signals (emergency stop signal light); the output digital signals DO sent by the robot include: 1 camera photo signal, 1 conveyor belt start and stop signal, and several auxiliary equipment control signals (added according to the auxiliary functions of the production line).

[0037] like Figure 5 The visual recognition platform mechanism 1 includes a main column 11, a telescopic column 12, an outwardly extending crossbeam 13, a camera 14, and a light strip bracket 15. The telescopic column 12 can be telescoped along the main column 11, the telescopic column 12 is fixedly connected to the outwardly extending crossbeam 13, the camera 14 is installed on the outwardly extending crossbeam 13, and the light strip bracket 15 is fixedly connected to the bottom of the main column 11 and the chassis mechanism 4 respectively to ensure the stability of the mechanism during the telescopic process. The telescopic column 12 and the main column 11 are standard aluminum alloy profiles, and there are standard slide grooves on the profiles, and T-nuts or bolts can be installed in the slide grooves. The telescopic column 12 and the main column 11 are connected by aluminum alloy angle pieces, and the connecting pieces are standard T-bolts and nuts. The T-nuts and bolts used for connection can be manually loosened to adjust the relative positions of the telescopic column 12 and the main column 11, thereby realizing the height adjustment of the camera 14 in the visual recognition platform mechanism 1, so as to facilitate the adjustment of the camera 14 shooting height after changing cameras with different focal lengths. The two main columns 11 are used as guide devices to ensure the directionality and stability of the telescopic column 12 during movement, and to prevent it from deviating from a predetermined track or shaking.

[0038] like Figures 6 to 8, the chassis mechanism 4 includes a working platform 41, a conveyor beam 42, a chassis support frame 43, a motor 44 and a motor support 45. The working platforms 41 are symmetrically arranged along the conveying direction of the conveying mechanism 2, and the working platforms 41 are fixedly connected to the chassis support frame 43. The conveying mechanism 2 is arranged in the middle of the working platform 41 through the conveyor beam 42. The chassis support frame 43 includes a cross beam 431, a base column 432, an inclined support beam 433 and a reinforcing plate 434. The cross beam 431 and the base column 432 are perpendicular to each other. An inclined support beam 433 is provided at the connection of the cross beam 431 and the base column 432, and a reinforcing plate is provided on the outer surface of the connection. The robotic arm 3 is provided on the surface of the working platform 41 and at the starting end of the conveying mechanism 2. The motor 44 is fixed to the motor support 45 by screws, and the motor support 45 is fixedly connected to the conveyor beam 42 by bolts to provide power for the conveying mechanism 2. The motor 44 drives the intermediate gear set through the driving gear, and finally drives the end gear to rotate, thereby driving the conveyor belt to rotate. Power and torque are transmitted between the gears through the meshing of the teeth to achieve the change of direction and the adjustment of speed. The driving gear is installed on the output shaft of the motor 44, the axis of the intermediate gear is parallel to the rotating shaft of the conveyor belt and meshes with the driving gear. The end gear is directly installed on the rotating shaft of the conveyor belt and meshes with the last intermediate gear.

[0039] The sorting method of the wood sorting device based on machine vision recognition in this embodiment is as follows:

[0040] (1) The controller controls the conveying mechanism 2 to transport the wooden board under the camera 14. When the travel switch at the photographing position of the camera 14 is triggered, after the controller obtains the input signal, it controls the conveying mechanism 2 to stop and controls the camera 14 to find the optimal shooting height for taking pictures. By adjusting the lighting angle of the light strip installed on the light strip bracket 15 and adjusting the height of the telescopic column 12, the distance between the camera 14 and the wooden board is changed, so that the wooden board can be fully photographed by the camera 14. At the same time, the focal length and brightness of the camera 14 are adjusted to make the photo clear.

[0041] (2) The photo taken by the camera 14 is stored in the storage position set by the computer. The visual recognition program on the computer reads the photo at the specified storage position and makes color recognition, and sends the digital signals representing different colors to the controller.

[0042] (3) After the controller obtains the color signal, it controls the conveying mechanism 2 to start and sends the wooden board to the grasping position, that is, transports it near the robotic arm 3 for sorting. When the travel switch at the grasping position is triggered, after the controller obtains the input signal, it controls the conveying mechanism 2 to stop, and grasps and places the wooden board at the specified position for palletizing and warehousing according to the color signal.

[0043] (4) After completing one sorting, the cycle is repeated in sequence.

[0044] The grasping position of the wooden board is fixed, and it only needs to place the grasped wooden board into different bins according to different color signals sent by the computer.

Claims

1. A wood sorting device based on machine vision recognition, characterized in that: It includes a visual recognition platform mechanism (1), a conveying mechanism (2), a robotic arm (3), a chassis mechanism (4), and a working chamber (5). The visual recognition platform mechanism (1), the conveying mechanism (2), the robotic arm (3), and the chassis mechanism (4) are arranged inside the working chamber (5). The visual recognition platform mechanism (1) and the robotic arm (3) are arranged on the chassis mechanism (4). The visual recognition platform mechanism (1) includes a main column (11), a telescopic column (12), an extended cross beam (13), a camera (14), and a light strip bracket (15). The telescopic column (12) can telescopically move along the main column (11). The telescopic column (12) is connected to the extended cross beam (13). The extended cross beam (13) is connected to the camera (14). The light strip bracket (15) is respectively connected to the main column (11) and the chassis mechanism (4).

2. The wood sorting device based on machine vision recognition according to claim 1, characterized in that: The chassis mechanism (4) includes a working platform (41), a conveyor belt cross beam (42), and a chassis support frame (43). The working platforms (41) are symmetrically arranged along the conveying direction of the conveying mechanism (2). The working platform (41) is connected to the chassis support frame (43). The conveying mechanism (2) is connected to the working platform (41) through the conveyor belt cross beam (42).

3. The wood sorting device based on machine vision recognition according to claim 2, characterized in that: The chassis support frame (43) includes a cross beam (431) and a base column (432). The cross beam (431) and the base column (432) are perpendicular to each other.

4. The wood sorting device based on machine vision recognition according to claim 3, wherein: The chassis support frame (43) further includes an inclined support beam (433). The inclined support beam (433) is arranged at the connection of the cross beam (431) and the base column (432).

5. A wood sorting device based on machine vision recognition according to claim 2, characterized in that: The chassis mechanism (4) further includes a motor (44) and a motor support (45). The motor (44) is connected to the conveyor belt cross beam (42) through the motor support (45).

6. The wood sorting device based on machine vision recognition according to claim 2, wherein: The robotic arm (3) is arranged on the surface of the working platform (41) and at the starting end of the conveying mechanism (2).

7. A wood sorting device based on machine vision recognition according to claim 1, characterized in that: The conveying mechanism (2) includes a number of conveyor belts, and the conveyor belts are on the same horizontal plane.

8. A wood sorting device based on machine vision recognition according to claim 1, characterized in that: The robotic arm (3) is a six-axis robotic arm.

9. The wood sorting device based on machine vision recognition according to claim 1, characterized in that: The visual recognition platform mechanism (1), the conveying mechanism (2), and the robotic arm (3) are all connected to a controller.

10. The wood sorting device based on machine vision recognition according to claim 1, characterized in that: Position sensors are arranged on the conveying mechanism (2).

Images