Novel rotor spraying robot path optimization method based on machine vision
By introducing machine vision and neural network into the rotor spraying robot system, automatic optimization of spraying paths and postures is achieved, solving the problems of cross-spraying paths and uneven paint film thickness in the prior art, and improving the spraying efficiency and accuracy. Intelligent spraying of rotor bend areas is realized.
Patent Information
- Application Number
- CN202411207700.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-30
- Publication Date
- 2024-11-12
AI Technical Summary
The existing rotor spraying robot system has shortcomings in spraying path optimization and paint film thickness control, resulting in path crossing, uneven paint film thickness and insufficient amount of paint spraying at the rotor bends. Manual brushing is required, and intelligent production cannot be achieved. .
Using a path optimization method based on machine vision and neural network, the spray amount and paint film thickness are recognized in real time through the visual system, combined with a six-axis robot, the optimal spray path is independently planned, and the neural network system is used for parallel optimization to realize the spray path and posture. automatic adjustment.
The optimization of spray path and efficiency is improved, the accuracy of the paint film thickness and efficiency of the spray area is ensured, real-time identification and intelligent spraying of the rotor bend area are realized, and production time and material usage are reduced.
Smart Images

Figure CN118925983A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of rotor spraying robots, and in particular to a novel rotor spraying robot path optimization method based on machine vision. Background Art
[0002] At present, the existing generator rotor spraying consists of a single manipulator and a rotor spraying tooling. By rotating the manipulator base platform, the manipulator's sixth axis spraying system performs regular spraying, and the spraying path is completed by manual teaching. Without the introduction of any intelligent algorithm, the action cannot be optimized during the spraying process.
[0003] In addition, the existing rotor spraying robot system has intersecting paths during the spraying process, and simple position teaching cannot guarantee a more accurate paint film thickness at the intersection. In addition, the existing rotor spraying robot system cannot guarantee the amount of paint sprayed at the rotor bend, and manual secondary brushing is required, which is contrary to the original purpose of the intelligent production line. Summary of the invention
[0004] In order to solve the above problems, the present invention discloses a novel rotor spraying robot path optimization method based on machine vision.
[0005] The specific plan is as follows: A novel rotor spraying robot path optimization method based on machine vision is characterized in that the spraying path optimization is achieved by integrating the visual system and the neural network system into the spraying system; wherein the spraying system includes a spray gun, a paint can, a nozzle opening controller (paint spraying controller) and a six-axis manipulator; the visual system is used to perform real-time identification of the rotor magnetic paint spraying amount, combined with the action of the six-axis manipulator, reduce invalid paths, independently plan the optimal path, and reduce production time and material usage; the neural network system is connected to the visual system to continuously annotate and train the collected paint spraying pictures, and combines the visual system and the six-axis manipulator for parallel optimization.
[0006] As a further improvement of the present invention, it also includes a fan interface, an exhaust port and a rotating platform. The fan interface is used to connect the fan. The fan runs to exhaust air, and the exhausted gas is discharged after being filtered by activated carbon. The exhaust port is arranged under the workpiece to prevent paint mist from drifting. The rotating platform is equipped with a six-axis manipulator. When the size of the workpiece exceeds the operating range of the manipulator, mechanical zoning and blocking operations are realized.
[0007] As a further improvement of the present invention, the following steps are included: S1. Prepare a spraying system, wherein the visual system is composed of a visual camera, a lens and a light source, and is installed on a six-axis manipulator, and the neural network system is a neural network classification and recognition model; S2, calibrating the base coordinate system of the six-axis manipulator and the pixel coordinate system of the camera; S3, start the power supply, the camera will take 360° photos of the inner wall of the rotor that has been placed, and the number of photos is ten; S4, manually measuring the area of the positions of the ten photos taken; S5, taking photos with the camera and putting them into the neural network classification and recognition model, and manually marking the measured values; S6, manually teach the six-axis manipulator to perform rotor spraying action and switch to automatic spraying mode; S7, rotor enamel spraying is completed, the camera takes ten pictures of the sprayed product area; S8. Manually measure the paint film thickness at the photographed location; S9, putting the ten photos taken into the neural network classification and recognition model, and marking the measured values with artificial paint film thickness; S10, the number of photos taken is the number of times the product is sprayed n*2; S11, classification and identification area accuracy>99%, paint film thickness detection accuracy>99%, the number of product groups this time is recorded as t; S12: In the first t groups of products, the photo is taken after the action is completed. After the t+1th group of products, the photo is taken one beat after the action. S13, changing the running program; S14, start the equipment operation; S15, the six-axis manipulator takes a 360° picture of the inner wall of the rotor with a camera to identify the total area, and then proceeds to step S5; S16, spraying with the spray gun for a period of time, taking photos after spraying, confirming the uniformity of the thickness of the sprayed paint film, if the thickness of the AB paint film meets the process requirements, proceed to the next step; S17, the spray gun performs the next spraying, takes photos after spraying, identifies the thickness of the paint film, and performs step S9 at the same time; S18, guiding the six-axis manipulator with the spray gun to carry out the next spraying according to the spraying thickness; S19, spraying the bend, if the total area in step S15 is within a certain range specified by the process, the spray gun opening is automatically adjusted for spraying; S20, after spraying, the camera continues to take pictures around, detect the thickness of the paint film, and import the neural network classification model for annotation; S21, introduces a new model, in which the camera takes pictures to guide the six-axis robot and spray gun to automatically complete the spraying.
[0008] As a further improvement of the present invention, in steps S3 and S7, the photos cover the entire area of the painted surface.
[0009] As a further improvement of the present invention, in step S8, a single-area six-point method is used to measure and obtain an average value.
[0010] The beneficial effects of the present invention are: 1. The introduction of vision and neural networks enables more efficient improvement and optimization of spraying path, posture and efficiency.
[0011] 2. By adopting the visual system, when the spraying areas intersect, the unsprayed areas can be accurately identified, and the robot's spraying path and posture can be adjusted in real time to ensure the paint film thickness and spraying efficiency in the spraying area.
[0012] 3. Real-time identification can be performed in the rotor bending area, and the posture and paint amount can be adjusted in real time to achieve intelligent spraying after full adjustment. BRIEF DESCRIPTION OF THE DRAWINGS
[0013] Figure 1 It is a framework diagram of the present invention. DETAILED DESCRIPTION
[0014] The present invention will be further explained below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are only used to illustrate the present invention and are not used to limit the scope of the present invention.
[0015] As shown in the figure, the present invention provides a novel rotor spraying robot path optimization method based on machine vision, which realizes spraying path optimization by integrating a visual system and a neural network system into the spraying system; wherein the spraying system includes a spray gun, a paint can, a nozzle opening controller (paint spraying controller) and a six-axis manipulator; the visual system is used to perform real-time identification of the rotor magnetic paint spraying amount, combined with the action of the six-axis manipulator, reduce invalid paths, autonomously plan the optimal path, and reduce production time and material usage; the neural network system is connected to the visual system to continuously annotate and train the collected paint spraying pictures, and combines the visual system and the six-axis manipulator for parallel optimization.
[0016] In this embodiment, it also includes a fan interface, an exhaust port and a rotating platform. The fan interface is used to connect the fan. The fan runs to exhaust air, and the exhausted gas is discharged after being filtered by activated carbon. The exhaust port is arranged under the workpiece to prevent paint mist from drifting. A six-axis manipulator is installed on the rotating platform. When the size of the workpiece exceeds the operating range of the manipulator, mechanical zoning and blocking operations are realized.
[0017] In this embodiment, the following steps are included: S1. Prepare a spraying system, wherein the visual system is composed of a visual camera, a lens and a light source, and is installed on a six-axis manipulator, and the neural network system is a neural network classification and recognition model; S2, calibrating the base coordinate system of the six-axis manipulator and the pixel coordinate system of the camera; S3, start the power supply, the camera will take 360° photos of the inner wall of the rotor that has been placed, and the number of photos is ten; S4, manually measuring the area of the positions of the ten photos taken; S5, taking photos with the camera and putting them into the neural network classification and recognition model, and manually marking the measured values; S6, manually teach the six-axis manipulator to perform rotor spraying action and switch to automatic spraying mode; S7, rotor enamel spraying is completed, the camera takes ten pictures of the sprayed product area; S8. Manually measure the paint film thickness at the photographed location; S9, putting the ten photos taken into the neural network classification and recognition model, and marking the measured values with artificial paint film thickness; S10, the number of photos taken is the number of times the product is sprayed n*2; S11, classification and identification area accuracy>99%, paint film thickness detection accuracy>99%, the number of product groups this time is recorded as t; S12: In the first t groups of products, the photo is taken after the action is completed. After the t+1th group of products, the photo is taken one beat after the action. S13, changing the running program; S14, start the equipment operation; S15, the six-axis manipulator takes a 360° picture of the inner wall of the rotor with a camera to identify the total area, and then proceeds to step S5; S16, spraying with the spray gun for a period of time, taking photos after spraying, confirming the uniformity of the thickness of the sprayed paint film, if the thickness of the AB paint film meets the process requirements, proceed to the next step; S17, the spray gun performs the next spraying, takes photos after spraying, identifies the thickness of the paint film, and performs step S9 at the same time; S18, guiding the six-axis manipulator with the spray gun to carry out the next spraying according to the spraying thickness; S19, spraying the bend, if the total area in step S15 is within a certain range specified by the process, the spray gun opening is automatically adjusted for spraying; S20, after spraying, the camera continues to take pictures around, detect the thickness of the paint film, and import the neural network classification model for annotation; S21, introduces a new model, in which the camera takes pictures to guide the six-axis robot and spray gun to automatically complete the spraying.
[0018] In this embodiment, in steps S3 and S7, the photos cover the entire area of the painted surface.
[0019] In this embodiment, in step S8, a single area six-point method is used to measure and obtain an average value.
[0020] The technical means disclosed in the scheme of the present invention are not limited to the technical means disclosed in the above-mentioned implementation mode, but also include technical schemes composed of any combination of the above-mentioned technical features. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications are also regarded as the protection scope of the present invention.
Claims
1. A novel rotor spraying robot path optimization method based on machine vision, characterized in that: By integrating the visual system and the neural network system into the spraying system, the spraying path optimization is achieved; the spraying system includes a spray gun, a paint can, a nozzle opening controller and a six-axis robot; the visual system is used to identify the amount of rotor magnetic paint sprayed in real time, combined with the action of the six-axis robot, to reduce invalid paths, independently plan the optimal path, and reduce production time and material usage; the neural network system is connected to the visual system to continuously annotate and train the collected paint pictures, and combined with the visual system and the six-axis robot, parallel optimization is performed.
2. According to a novel machine vision-based rotor spraying robot path optimization method according to claim 1, it is characterized in that: It also includes a fan interface, an air exhaust port and a rotating platform.
3. According to a novel machine vision-based rotor spraying robot path optimization method according to claim 2, it is characterized in that: The fan interface is used to connect a fan, and the fan operates to extract air, and the extracted gas is filtered through activated carbon and discharged.
4. According to a novel machine vision-based rotor spraying robot path optimization method according to claim 2, it is characterized in that: The air exhaust port is arranged below the workpiece to prevent paint mist from drifting.
5. According to a novel machine vision-based rotor spraying robot path optimization method according to claim 2, it is characterized in that: The rotating platform is equipped with a six-axis manipulator, and when the size of the workpiece exceeds the operating range of the manipulator, mechanical zoning and blocking operations are implemented.
6. A novel rotor spraying robot path optimization method based on machine vision according to claim 1, characterized in that: The following steps are involved: S1. Prepare a spraying system, wherein the visual system is composed of a visual camera, a lens and a light source, and is installed on a six-axis manipulator, and the neural network system is a neural network classification and recognition model; S2, calibrating the base coordinate system of the six-axis manipulator and the pixel coordinate system of the camera; S3, start the power supply, the camera will take 360° photos of the inner wall of the rotor that has been placed, and the number of photos is ten; S4, manually measuring the area of the positions of the ten photos taken; S5, taking photos with the camera and putting them into the neural network classification and recognition model, and manually marking the measured values; S6, manually teach the six-axis manipulator to perform rotor spraying action and switch to automatic spraying mode; S7, rotor enamel spraying is completed, the camera takes ten pictures of the sprayed product area; S8. Manually measure the paint film thickness at the photographed location; S9, putting the ten photos taken into the neural network classification and recognition model, and marking the measured values with artificial paint film thickness; S10, the number of photos taken is the number of times the product is sprayed n*2; S11, classification and identification area accuracy>99%, paint film thickness detection accuracy>99%, the number of product groups this time is recorded as t; S12: In the first t groups of products, the photo is taken after the action is completed. After the t+1th group of products, the photo is taken one beat after the action. S13, changing the running program; S14, start the equipment operation; S15, the six-axis manipulator takes a 360° picture of the inner wall of the rotor with a camera to identify the total area, and then proceeds to step S5; S16, spraying with the spray gun for a period of time, taking photos after spraying, confirming the uniformity of the thickness of the sprayed paint film, if the thickness of the AB paint film meets the process requirements, proceed to the next step; S17, the spray gun performs the next spraying, takes photos after spraying, identifies the thickness of the paint film, and performs step S9 at the same time; S18, guiding the six-axis manipulator with the spray gun to carry out the next spraying according to the spraying thickness; S19, spraying the bend, the total area in step S15 is within the specified range of the process, and the spray gun opening is automatically adjusted for spraying; S20, after spraying, the camera continues to take pictures around, detect the thickness of the paint film, and import the neural network classification model for annotation; S21, introduces a new model, in which the camera takes pictures to guide the six-axis robot and spray gun to automatically complete the spraying.
7. The novel machine vision-based rotor spraying robot path optimization method according to claim 6 is characterized in that: In steps S3 and S7, the photos cover the entire area of the spray-painted surface.
8. The novel machine vision-based rotor spraying robot path optimization method according to claim 6 is characterized in that: In the step S8, a single area six-point method is used to measure and obtain an average value.
