阅读说明：本技术 一种遥控器零件自动装配方法和系统 (Automatic assembly method and system for remote controller parts ) 是由 徐胜军 詹博涵 韩九强 孟月波 刘光辉 于 2021-09-03 设计创作，主要内容包括：本发明公开了一种遥控器零件自动装配方法和系统,属于人工智能与装配机器人技术领域,采集遥控器不同零件图像作为训练样本,训练后得到网络模型。启动传送带并运行程序,首先判断传送带上是否有零件,若有零件则将采集的图像输入到已训练的网络中,输出可以得到零件的类别等信息,计算可以得到零件的抓取点和角度信息。随着传送带的运动会导致遥控器零件出现旋转,在抓取遥控器零件前需要对旋转角度进行计算,以保证每个遥控器零件可以精准地到达安装位置。将抓取的待装配零件按照装配位置和装配顺序进行安装,直至遥控器安装完成。对于遥控器零件的识别准确率高、识别速度快、鲁棒性高,且可以快速完成装配。(The invention discloses a method and a system for automatically assembling remote controller parts, which belong to the technical field of artificial intelligence and assembly robots. Starting the conveyor belt and running a program, firstly judging whether parts exist on the conveyor belt, inputting the acquired images into a trained network if the parts exist, outputting information such as the types of the parts and the like, and calculating the information of the grabbing points and the angles of the parts. Along with the motion of conveyer belt can lead to the remote controller part to appear rotating, need calculate rotation angle before snatching the remote controller part to guarantee that every remote controller part can arrive the mounted position accurately. And installing the grabbed parts to be assembled according to the assembling positions and the assembling sequence until the remote controller is installed. The remote controller part recognition accuracy is high, the recognition speed is high, the robustness is high, and the assembly can be completed quickly.)

1. An automatic assembly method for remote controller parts is characterized by comprising the following steps:

step 1) acquiring image information of all parts of a remote controller, taking the image information of all parts as training samples, and obtaining a part network model of the remote controller after training;

step 2) collecting image information above the conveyor belt, judging whether parts exist on the conveyor belt or not based on the image information above the conveyor belt,

if no part exists on the conveyor belt, continuing waiting;

if the parts exist on the conveyor belt, further acquiring image information of the parts on the conveyor belt, and combining a part network model of a remote controller to obtain the part types, size information, position information and rotation angles on the conveyor belt;

and 3) grabbing the parts on the transmission belt based on the part types, the size information, the position information and the grabbing point information on the transmission belt, and installing the grabbed parts to be assembled according to the assembly positions and the assembly sequence by combining the preset assembly sequence and the assembly position of each part until the installation of the remote controller is finished.

2. The automatic remote controller part assembling method according to claim 1, wherein the training in step 1) is performed using an instance division network.

3. The automatic assembling method of parts for remote controller according to claim 1, wherein the specific operation of judging whether there is a part on the conveyor belt in step 2) is:

collecting a visual image above the conveyor belt, and converting the collected visual image into a gray scale image;

carrying out binarization processing on the gray level image according to the threshold value characteristics of the surface color of the conveyor belt, wherein the gray level image is a background when the threshold value characteristics of the surface color of the conveyor belt are reached, and otherwise the gray level image is a foreground;

calculating the maximum area of the foreground connected region;

if the area of the connected region is larger than or equal to a preset lowest threshold value, an object is considered to be on the conveyor belt, and at the moment, the frame of image is transmitted into a pre-trained network for recognition; if the connected area is smaller than the preset minimum threshold value, the conveyor belt is considered to be without objects.

4. The automatic assembly method of remote controller parts according to claim 1, wherein in step 3), the specific operation of acquiring the rotation angle in step 3) is:

the method comprises the steps of collecting visual images above parts on a conveyor belt, extracting rectangular outlines based on the part images so as to obtain a minimum external rectangle of the parts, and calculating the rotation angle of the parts of the remote controller according to four vertexes of the minimum external rectangle.

5. An automatic remote control part assembling system, comprising:

the image acquisition module is used for acquiring image information of all parts of the remote controller;

the image information processing module is interacted with the image acquisition module and is used for processing the image information of all parts of the remote controller, and the image information of all parts of the remote controller is used as a training sample for training to obtain a part network model of the remote controller;

the conveyor belt part information acquisition module is used for acquiring image information above the conveyor belt, judging whether parts exist on the conveyor belt or not based on the image information above the conveyor belt, and continuing to wait if no parts exist on the conveyor belt; if the conveyor belt is provided with the part, the part is the part to be assembled, and image information of the part to be assembled on the conveyor belt is further acquired;

the conveyor belt part identification module is respectively interacted with the image information processing module and the conveyor belt part information acquisition module, and based on a part network model of the remote controller and the image information of the parts to be assembled on the conveyor belt, the type and size information of the parts on the conveyor belt and the information of the grabbing points are obtained;

the assembly robot interacts with the conveyor belt part identification module, moves based on the type and size information of the parts to be assembled and the information of the grabbing points, moves to the grabbing area corresponding to the grabbing points of the parts to be assembled, stops moving, and grabs the parts to be assembled;

and the mounting module is interacted with the assembly robot, and is used for mounting the grabbed parts to be assembled according to the assembly positions and the assembly sequence based on the preset assembly sequence and the assembly positions of each part until the remote controller is mounted.

6. The automatic remote controller part assembling system according to claim 5, wherein the assembling robot is provided with two end effectors for performing operations of sucking the parts and mounting the parts, respectively.

7. The remote control automatic parts assembly system of claim 5, further comprising an air pump for switching the end-effectors and controlling the end-effectors.

8. The automatic remote controller part assembling system according to claim 5, wherein the image collecting module and the conveyor belt part information collecting module are respectively performed by an industrial camera.

9. The automatic remote control part assembling system according to claim 5, wherein the mounting module is performed through an operation table;

and the operating platform is also provided with an industrial camera for acquiring image information of the installation process on the operating platform.

Technical Field

The invention belongs to the technical field of artificial intelligence and assembly robots, and relates to a method and a system for automatically assembling remote controller parts.

Background

For a long time, low labor cost is a great advantage of the manufacturing industry, and most remote controller assembly enterprises rely on manual assembly. With the continuous innovation of industrial products, the assembly process of the products is more and more complex, and the requirement on the assembly precision is higher and higher. In the conventional remote controller assembling process, as shown in fig. 1, a remote controller face cover 10, a PCB11 and a bottom cover 12 are assembled, and the process of manually assembling the remote controller is as follows: the PCB11 is mounted in the face cover 10 at a corresponding position, the bottom cover 12 is covered on the face cover 10, and the face cover 10 and the bottom cover 12 are fastened by pressing the bottom cover 12 with force. Due to the fact that the internal assembly space of the remote controller is narrow and small, high assembly precision is needed, workers work for a long time under the influence of visual fatigue factors, and the assembly efficiency is low and the assembly quality is poor.

Disclosure of Invention

The invention aims to overcome the defects of low manual assembly efficiency and poor assembly quality in the prior art, and provides a method and a system for automatically assembling remote controller parts.

In order to achieve the purpose, the invention adopts the following technical scheme to realize the purpose:

an automatic assembly method for remote controller parts comprises the following steps:

step 1) acquiring image information of all parts of a remote controller, taking the image information of all parts as training samples, and obtaining a part network model of the remote controller after training;

step 2) collecting image information above the conveyor belt, judging whether parts exist on the conveyor belt or not based on the image information above the conveyor belt,

if no part exists on the conveyor belt, continuing waiting;

if the parts exist on the conveyor belt, further acquiring image information of the parts on the conveyor belt, and combining a part network model of a remote controller to obtain the part types, size information, position information and rotation angles on the conveyor belt;

and 3) grabbing the parts on the transmission belt based on the part types, the size information, the position information and the grabbing point information on the transmission belt, and installing the grabbed parts to be assembled according to the assembly positions and the assembly sequence by combining the preset assembly sequence and the assembly position of each part until the installation of the remote controller is finished.

Preferably, the training in step 1) is performed using an instance segmentation network.

Preferably, in the step 2), the specific operation of judging whether the parts exist on the conveyer belt is as follows:

collecting a visual image above the conveyor belt, and converting the collected visual image into a gray scale image;

carrying out binarization processing on the gray level image according to the threshold value characteristics of the surface color of the conveyor belt, wherein the gray level image is a background when the threshold value characteristics of the surface color of the conveyor belt are reached, and otherwise the gray level image is a foreground;

calculating the maximum area of the foreground connected region;

if the area of the connected region is larger than or equal to a preset lowest threshold value, an object is considered to be on the conveyor belt, and at the moment, the frame of image is transmitted into a pre-trained network for recognition; if the connected area is smaller than the preset minimum threshold value, the conveyor belt is considered to be without objects.

Preferably, in step 3), the specific operation of acquiring the rotation angle is:

the method comprises the steps of collecting visual images above parts on a conveyor belt, extracting rectangular outlines based on the part images so as to obtain a minimum external rectangle of the parts, and calculating the rotation angle of the parts of the remote controller according to four vertexes of the minimum external rectangle.

An automatic remote controller part assembling system, comprising:

the image acquisition module is used for acquiring image information of all parts of the remote controller;

the image information processing module is interacted with the image acquisition module and is used for processing the image information of all parts of the remote controller, and the image information of all parts of the remote controller is used as a training sample for training to obtain a part network model of the remote controller;

the conveyor belt part information acquisition module is used for acquiring image information above the conveyor belt, judging whether parts exist on the conveyor belt or not based on the image information above the conveyor belt, and continuing to wait if no parts exist on the conveyor belt; if the conveyor belt is provided with the part, the part is the part to be assembled, and image information of the part to be assembled on the conveyor belt is further acquired;

the conveyor belt part identification module is respectively interacted with the image information processing module and the conveyor belt part information acquisition module, and based on a part network model of the remote controller and the image information of the parts to be assembled on the conveyor belt, the type and size information of the parts on the conveyor belt and the information of the grabbing points are obtained;

the assembly robot interacts with the conveyor belt part identification module, moves based on the type and size information of the parts to be assembled and the information of the grabbing points, moves to the grabbing area corresponding to the grabbing points of the parts to be assembled, stops moving, and grabs the parts to be assembled;

and the mounting module is interacted with the assembly robot, and is used for mounting the grabbed parts to be assembled according to the assembly positions and the assembly sequence based on the preset assembly sequence and the assembly positions of each part until the remote controller is mounted.

Preferably, the assembly robot is provided with two end effectors for performing operations of sucking the parts and mounting the parts, respectively.

Preferably, the system further comprises an air pump for switching the end-effector and controlling the end-effector.

Preferably, the image acquisition module and the conveyor belt part information acquisition module are respectively performed by an industrial camera.

Preferably, the installation module is carried out by means of an operating console;

and the operating platform is also provided with an industrial camera for acquiring image information of the installation process on the operating platform.

Compared with the prior art, the invention has the following beneficial effects:

the invention discloses an automatic assembly method for remote controller parts, which is characterized in that images of different parts of a remote controller are collected as training samples, and a network model is obtained after training. Starting the conveyor belt and running a program, firstly judging whether parts exist on the conveyor belt, inputting the acquired images into a trained network if the parts exist, outputting information such as the types of the parts and the like, and calculating the information of the grabbing points and the angles of the parts. Along with the motion of conveyer belt can lead to the remote controller part to appear rotating, need calculate rotation angle before snatching the remote controller part to guarantee that every remote controller part can arrive the mounted position accurately. And (4) combining the preset assembly sequence and the assembly position of each part, and installing the grabbed parts to be assembled according to the assembly positions and the assembly sequence until the installation of the remote controller is completed. The remote controller part recognition accuracy is high, the recognition speed is high, the robustness is high, and the assembly can be completed quickly.

The invention also discloses an automatic assembly system for the remote controller parts, which comprises an image acquisition module, an image information processing module, a conveyor belt part information acquisition module, a conveyor belt part identification module, an assembly robot and an installation module. The collected images are accurately divided by using an image processing algorithm to obtain the information of the grabbing points and the grabbing angles, the assembly robot grabs the remote controller parts on the operating platform and assembles the remote controller parts according to the installation sequence of the remote controller. The system fully embodies the intelligent automation concept of the assembly robot, and the assembly robot can complete the assembly of the remote controller. The tail end of the assembling robot is provided with an industrial camera to replace human eyes to detect and classify remote controller parts, and the assembling robot replaces human hands to assemble the remote controller in a real environment. The method lays a foundation for further researching the operation of an intelligent automatic system of the assembly robot, has high intelligent level, and is a typical application of combining an artificial intelligence technology and an assembly robot technology.

Drawings

FIG. 1 is a schematic diagram of an intelligent remote controller assembly machine;

FIG. 2 is a main flow chart of the remote controller assembly;

FIG. 3 is a flow chart of remote control cover installation;

FIG. 4 is a flow chart of remote control PCB installation;

FIG. 5 is a flow chart of the bottom cover installation of the remote controller;

wherein: 1-a computer; 2, an air pump; 3-quickly changing a bracket by the end effector; 4-a first end effector; 5-a second end effector; 6-assembling the robot; 7-an industrial camera; 8-an operation table; 9-remote controller fixing device; 10-face cover; 11-PCB; 12-a bottom cover; 13-a remote controller; 14-a conveyor belt.

Detailed Description

The invention is described in further detail below with reference to the accompanying drawings:

example 1

An automatic assembly method for remote controller parts is shown in fig. 2, and comprises the following steps:

step 1) acquiring image information of all parts of a remote controller, taking the image information of all parts as training samples, and obtaining a part network model of the remote controller after training;

step 2) collecting image information above the conveyor belt, judging whether parts exist on the conveyor belt or not based on the image information above the conveyor belt,

if no part exists on the conveyor belt, continuing waiting;

if the parts exist on the conveyor belt, further acquiring image information of the parts on the conveyor belt, and combining a part network model of a remote controller to obtain the part category, size information, position information and grabbing point information on the conveyor belt;

and 3) grabbing the parts on the transmission belt based on the part types, the size information, the position information and the grabbing point information on the transmission belt, and installing the grabbed parts to be assembled according to the assembly positions and the assembly sequence by combining the preset assembly sequence and the assembly position of each part until the installation of the remote controller is finished.

Example 2

An automatic assembly method for remote controller parts comprises the following steps:

step 1) acquiring image information of all parts of a remote controller, taking the image information of all parts as training samples, and obtaining a part network model of the remote controller after training;

step 2) collecting image information above the conveyor belt, judging whether parts exist on the conveyor belt or not based on the image information above the conveyor belt,

collecting a visual image above the conveyor belt, and converting the collected visual image into a gray scale image;

carrying out binarization processing on the gray level image according to the threshold value characteristics of the surface color of the conveyor belt, wherein the gray level image is a background when the threshold value characteristics of the surface color of the conveyor belt are reached, and otherwise the gray level image is a foreground;

calculating the maximum area of the foreground connected region, if the area of the connected region is larger than the minimum threshold value, determining that an object exists on the conveyor belt, and transmitting the frame of image into a pre-trained network to be identified as a certain part; if the connected region area is less than the minimum threshold, the conveyor belt is deemed to be free of objects.

If no part exists on the conveyor belt, continuing waiting;

if the parts exist on the conveyor belt, further acquiring image information of the parts on the conveyor belt, and combining a part network model of a remote controller to obtain the part category, size information, position information and grabbing point information on the conveyor belt;

and 3) grabbing the parts on the transmission belt based on the part types, the size information, the position information and the grabbing point information on the transmission belt, and installing the grabbed parts to be assembled according to the assembly positions and the assembly sequence by combining the preset assembly sequence and the assembly position of each part until the installation of the remote controller is finished.

Example 3

A remote controller part automatic assembly method comprises the following algorithms of remote controller part identification, remote controller part grabbing point positioning and remote controller part grabbing angle:

and collecting images of different parts of a plurality of remote controllers as training samples, and training by using the example segmentation network to finally obtain a network model. Starting the conveyor belt and running a program, firstly judging whether an object is detected on the conveyor belt, inputting the acquired image into a trained network if the object is detected, outputting the type of the object and the coordinates of the upper left corner and the lower right corner of the object, and calculating the coordinates to obtain the information of the grabbing point and the angle.

The detection steps of whether the object exists on the conveyor belt are as follows: the method comprises the steps of collecting a visual image above a conveyor belt, converting the image into a gray image, and carrying out binarization processing on the image according to the threshold value characteristics of the surface color of the conveyor belt, wherein the threshold value meeting the color of the conveyor belt is a background, and otherwise, the threshold value is a foreground. And judging whether an object exists on the conveying belt or not according to the size of the maximum area of the foreground communication area.

The remote controller part rotation angle detection algorithm comprises the following steps: the method comprises the steps of collecting visual images above remote controller parts, extracting rectangular outlines of the remote controller part images to obtain a minimum external rectangle of the remote controller parts, and calculating the rotation angle of the parts according to four vertexes of the minimum external rectangle.

Example 4

The same as example 3 was repeated except for the following.

Step 1) acquiring image information of all parts of a remote controller, taking the image information of all parts as training samples, and obtaining a part network model of the remote controller after training;

and 2) as shown in fig. 2, when the remote controller is installed, firstly installing a remote controller surface cover, then installing a remote controller PCB, then installing a remote controller bottom cover, and finally placing the installed remote controller on a conveyor belt for subpackaging.

As shown in fig. 3, when the remote controller cover is installed, image information above the conveyor belt is first collected, whether a part is on the conveyor belt is judged based on the image information above the conveyor belt,

if no part exists on the conveyor belt, continuing waiting;

if the parts exist on the conveyor belt, further acquiring image information of the parts on the conveyor belt, combining a part network model of the remote controller to obtain whether the parts on the conveyor belt belong to the remote controller parts, if not, continuing to wait, and if the parts belong to the remote controller parts, continuing to calculate to obtain size information, position information and rotation angle of the remote controller parts, and further acquiring a grabbing point and a grabbing angle. The remote controller part grabbing can be divided into two stages of coarse positioning and fine positioning, wherein the coarse positioning refers to grabbing the remote controller part from the dynamic conveyor belt and placing the remote controller part on the operating platform, and the grabbing precision is difficult to guarantee to reach the assembling precision due to the fact that the remote controller part is grabbed from the dynamic conveyor belt in the process, so that fine positioning operation is needed; the fine positioning refers to the process of grabbing remote controller parts from the operation table and placing the remote controller parts into the remote controller assembly position, and the remote controller parts are statically grabbed from the operation table in the process, so that the assembly precision can be met. And the assembly robot grabs and assembles the parts of the remote controller according to the preset assembly position and the assembly sequence.

Similarly, according to fig. 4 and 5, the PCB and the rear cover of the remote controller are mounted to complete the mounting of the remote controller.

Example 5

An automatic remote controller part assembling system, comprising:

the image acquisition module is used for acquiring image information of all parts of the remote controller;

the image information processing module is interacted with the image acquisition module and is used for processing the image information of all parts of the remote controller, and the image information of all parts of the remote controller is used as a training sample for training to obtain a part network model of the remote controller;

the conveyor belt part information acquisition module is used for acquiring image information above the conveyor belt, judging whether parts exist on the conveyor belt or not based on the image information above the conveyor belt, and continuing to wait if no parts exist on the conveyor belt; if the conveyor belt is provided with the part, the part is the part to be assembled, and image information of the part to be assembled on the conveyor belt is further acquired;

the conveyor belt part identification module is respectively interacted with the image information processing module and the conveyor belt part information acquisition module, and based on a part network model of the remote controller and the image information of the parts to be assembled on the conveyor belt, the type and size information of the parts on the conveyor belt and the information of the grabbing points are obtained;

the assembly robot interacts with the conveyor belt part identification module, moves based on the type and size information of the parts to be assembled and the information of the grabbing points, moves to the grabbing area corresponding to the grabbing points of the parts to be assembled, stops moving, and grabs the parts to be assembled;

and the mounting module is interacted with the assembly robot, and is used for mounting the grabbed parts to be assembled according to the assembly positions and the assembly sequence based on the preset assembly sequence and the assembly positions of each part until the remote controller is mounted.

Example 6

As shown in fig. 1, an automatic assembling system for remote controller parts comprises a computer 1, a remote controller assembling robot 6, a first end effector 4, a second end effector 5, an air pump 2 for switching the end effectors and controlling the end effectors, an operation table 8 and a conveyor belt 14, wherein a remote controller fixing device 9 and a mounted remote controller 13 are arranged on the operation table 8, and the remote controller 13 can be divided into a surface cover 10, a PCB11 and a bottom cover 12. The computer 1 executes a remote controller component detection algorithm on the acquired image, converts the coordinates of the grabbing points of the remote controller components in the image according to a certain proportional relation to obtain the motion coordinates of the assembly robot under a coordinate system, performs corresponding grabbing operation by using the first end effector 4 by combining the installation sequence of the remote controller, the installation height of the remote controller components and the identified remote controller component type information, and places the assembly robot into the fixed remote controller device 9 by using the second end effector 5 for installation.

The assembling robot 6 performs object recognition by the industrial camera 7 and then guides the assembling robot 6 to perform a grasping operation. The calibration of the assembly robot 6 and the industrial camera 7 is required, and the specific operation steps are as follows:

step 1: the computer 1 controls the assembling robot 6 to move so that the X and Y axis coordinates on the operation table are perpendicular in the image taken by the industrial camera 7 and the vertical distance of the industrial camera from the operation table 8 is 40 cm, and records the coordinates (X) of the assembling robot 6 at this time_p，y_p，z_p，A_p，B_p，C_p)。

Step 2: the assembly robot 6 is controlled to move down along the Z-axis until the first end effector 4 contacts the operation table 8, then the first end effector 4 is marked as the origin at the position where it contacts the operation table 8, and the assembly robot 6 returns to the shooting position (x)_p，y_p，z_p，A_p，B_p，C_p) On the table 8, marks are made by 10 cm vertical to the X direction of the origin and 10 cm vertical to the Y direction of the origin, respectively.

And step 3: the pixel coordinates (x) of the origin on the operation table 8 can be obtained by turning on the industrial camera 7₀，y₀) And the pixel coordinate along the X direction is (X)₁，y₁) And the pixel coordinate along the Y direction is (x)₂，y₂) Here x₀And x₂And y₀And y₁Are approximately equal. The offset proportionality coefficient delta along the X-axis and the Y-axis can be calculated through pixel coordinates_xAnd Δ_y。

And 4, step 4: if a point D is selected in the imaging area on the operation table 8, the pixel coordinate of the point D is (x)_D，y_D) Then the assembly robot 6 should arrive at point D and the assembly robot 6 coordinate should be (X)_D，Y_D，Z_D，A_p，B_p，C_p) Wherein Z is_DTo assemble the robot 6 to the height of the operating table 8, X_DAnd, Y_DCan be calculated by the following formula.

X_D＝Δ_x(x_D-x₀)+x_p

Y_D＝Δ_y(y_D-y₀)+y_p

And 5: optionally, a point in the image area is recorded on the control panel 8, the assembly robot 6 is controlled to be guided visually to the point on the control panel 8, and if a deviation occurs, Δ is adjusted_xAnd Δ_ySo that the grabbing operation of the assembly robot 6 is more accurate.

The tail end quick-change device of the assembling robot 6 is connected with the air pump 2, the computer 1 is directly connected with the controller of the assembling robot 6 through a network cable, and the linear motion of the assembling robot 6 and the state of an input port are controlled in an Ethernet communication mode, so that the grabbing operation of the input port state switching end effector and the first end effector 4 and the installing operation of the second end effector 5 are controlled.

Example 7

The contents are the same as those of example 5 except for the following.

Due to the narrow space for assembling the remote controller 13, the remote controller cannot be installed easily if the remote controller is not accurately grabbed, so that the remote controller needs to be grabbed coarsely and finely. In the rough grabbing stage, the precision requirement for grabbing objects is not high, the objects only need to be grabbed, the assembling robot 6 collects images above the conveying belt 14 through the industrial camera 7, whether the objects exist in the images is judged, if the objects exist, the collected images are stored and sent to the Yoract network to obtain information of the type, grabbing point and pose of the remote controller parts, and the assembling robot is guided by vision to grab and place the remote controller parts in the remote controller part placing area on the operating table 8. The grabbing precision requirement of the object in the fine grabbing stage is high, the assembling robot 6 collects images above a placing area of remote controller parts of the operating platform 8 through the industrial camera 7, the computer 1 obtains accurate coordinate positions and included angle parameters of the object through image processing, and then the assembling robot 1 is controlled to grab the remote controller parts and place the remote controller parts to the fixed remote controller device 9. When the remote control is mounted, it is withdrawn from the stationary remote control unit 9 and replaced on the conveyor belt.

Example 8

A remote controller part automatic assembly method based on the system comprises the following steps:

step 1: and starting the intelligent remote controller assembly machine to complete various initialization work including initialization of the vision assembly robot, initialization of the assembly robot, starting of the conveyor belt and the like.

Step 2: the visual assembly robot 6 of the intelligent assembly machine for the air conditioner remote controller starts a coarse grabbing stage for remote controller assembly, images right above a conveyor belt are collected, if objects exist in the images, a remote controller part detection and classification algorithm is called to obtain grabbing point and pose information of remote controller parts, and the computer 6 grabs the remote controller parts by controlling the assembly robot and places the remote controller parts in corresponding areas on an operation table 8. When the face cover 10, PCB11, and bottom cover 12 of the remote control 13 are placed within the area of the remote control components, the rough gripping stage of the remote control assembly is complete.

And step 3: and in the remote controller equipment fine grabbing stage, images of remote controller part areas on the operating platform 8 are collected, and accurate grabbing points and included angle parameters of the remote controller parts can be obtained by using the instance segmentation network. Firstly, the remote controller face cover 10 is grabbed, the calculator 1 controls the assembling robot 6 to grab the remote controller face cover 10, and then the remote controller face cover 10 is inserted into the fixed remote controller device 9; secondly, the PCB11 of the remote controller is grabbed and put into the remote controller face cover 10 in the previous step; finally, the bottom cover 12 of the remote controller is grasped and grasped to ensure that the four corners of the bottom cover 12 and the face cover 10 are aligned, and then the four corners of the bottom cover of the remote controller are properly pressed downward by using the second end effector 5, whereby the installation of the remote controller 13 is completed.

And 4, step 4: the first end effector 4 is used to push the remote controller 13 out by two thirds from the fixed remote controller device 9, then the first end effector 4 is controlled to grab the remote controller, the remote controller is moved to the position right above the conveyor belt, whether an object exists in the image shot by the industrial camera 7 or not is judged, and if no object exists, the remote controller is placed on the conveyor belt.

And 5: when the mounted remote control 13 is placed back on the conveyor belt, the remote control is assembled.

In conclusion, the method and the system for automatically assembling the remote controller parts provided by the invention can accurately complete the remote controller assembly, have good detection classification robustness and high recognition speed of the remote controller parts under different illumination conditions, can be used as a model system of an intelligent automatic system for deep research, fully embody the concept of machine intelligence instead of manual operation, are typical applications of combination of artificial intelligence technology and assembly robot technology, and provide a foundation for further research on the intelligent automatic system of the assembly robot.

The above-mentioned contents are only for illustrating the technical idea of the present invention, and the protection scope of the present invention is not limited thereby, and any modification made on the basis of the technical idea of the present invention falls within the protection scope of the claims of the present invention.