Piston blank multidirectional photographing recognition system

Abstract

The invention discloses a piston blank multidirectional photographing recognition system, relates to the technical field of visual recognition analysis, and solves the technical problem of low accuracy of acquired images due to the fact that the influence of the environment on a visual camera is not considered in the prior art. According to the method, the current recognition environment is detected, and the current recognition environment is compared with the preset labeling recognition environment to obtain the environment state; calculating an adjustment coefficient based on the abnormal environment state, comparing the adjustment coefficient with a preset adjustment threshold, and taking corresponding measures according to a comparison result; visual identification is conducted on the piston through the visual identification device, the manipulator is controlled to grab the piston to move according to the visual identification result, the outline image and the identification direction serve as input data to be input into the angle model to obtain the manipulator angle, and the manipulator is rotated according to the manipulator angle to adjust the direction of the piston; and the piston is subjected to secondary recognition through the secondary visual recognition device, and an alarm signal is generated.

Description

Technical Field

[0001] The invention belongs to the field of visual recognition and analysis, relates to Internet of Things technology, and specifically is a multi-directional photography and recognition system for piston blanks. Background Art

[0002] The piston is an important component of the internal combustion engine and is known as the "heart" of the internal combustion engine. Its performance directly affects the performance, emission standards and economy of the engine. The manufacturing methods of piston blanks include casting, forging and other forms; in the manufacturing process, the process parameters at each stage need to be strictly controlled to ensure the quality of the final product. Therefore, accurately identifying and completing the assembly task of the piston has become a major focus of internal combustion engine assembly.

[0003] The invention patent with application number CN2010102172254 discloses a method and device for assembling a piston, a piston pin and a connecting rod based on an industrial robot. The invention identifies the positions of the piston and the connecting rod through a visual recognition system, calculates the grasping position, controls the robot gripper to grasp and place the piston on the automatic assembly platform according to the calculation result, and completes the automatic assembly of the piston. When the invention identifies the piston position for grasping, it directly obtains and analyzes the image through a visual camera, and provides data support for subsequent automatic assembly according to the analysis result. However, when acquiring the image, factors such as light and vibration will affect the accuracy of the image collected by the camera. If factors such as light and vibration affect the accuracy of the collected image, it may cause errors in the subsequent robot gripper grasping or subsequent assembly due to insufficient image accuracy.

[0004] The present invention provides a piston blank multi-directional photographing and identifying system to solve the above technical problems. Summary of the invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes a piston blank multi-directional photographing and recognition system, which is used to solve the technical problem in the prior art that the image collected is not accurate because the impact of the environment on the visual camera is not taken into account.

[0006] To achieve the above-mentioned object, the first aspect of the present invention provides a piston blank multi-directional photographic recognition system, comprising: a database, a recognition environment control module, a visual recognition module and a manipulator control module;

[0007] Recognition environment control module: used to detect the current recognition environment, compare the current recognition environment with the preset standard recognition environment to obtain the environment state, and adjust the current recognition environment according to the environment state;

[0008] Visual recognition module: used to recognize the image of the piston transported by the conveyor belt and obtain primary recognition data and secondary recognition data;

[0009] Manipulator control module: used to control the primary manipulator to grab the piston according to the primary identification data, and control the secondary manipulator according to the secondary identification data and generate an alarm signal.

[0010] Preferably, the step of comparing the current recognition environment with a preset standard recognition environment to obtain the environment state includes:

[0011] Extracting the current recognition environment; setting a standard recognition environment; wherein the standard recognition environment includes standard lighting conditions, standard vibration conditions, and standard temperature conditions; the current recognition environment includes current lighting, current vibration, and current temperature of the current recognition environment;

[0012] Determine whether the current lighting meets the standard lighting conditions; if yes, mark the lighting state of the current recognition environment as normal lighting; if no, mark the lighting state of the current recognition environment as abnormal lighting;

[0013] Determine whether the current vibration meets the standard vibration condition; if yes, mark the vibration state of the current recognition environment as normal vibration; if no, mark the vibration state of the current recognition environment as abnormal vibration;

[0014] Determine whether the current temperature meets the standard temperature condition; if yes, mark the temperature state of the current identification environment as normal temperature; if no, mark the temperature state of the current identification environment as abnormal temperature;

[0015] Abnormal illumination, abnormal vibration and abnormal temperature are marked as abnormal environmental states, and the illumination state, vibration state and temperature state are combined to obtain the environmental state.

[0016] Preferably, the adjusting the current recognition environment according to the environmental state includes:

[0017] Extract the environment state;

[0018] Determine whether the environmental state of the current identification environment contains an abnormal environmental state; if so, generate adjustment data based on the abnormal environmental state and the corresponding standard identification environment, adjust the current identification environment based on the adjustment data, and re-detect after the adjustment is completed; if not, generate an identification signal and start the visual device to perform the first identification of the piston.

[0019] Preferably, the step of generating adjustment data according to the abnormal environmental state and the corresponding standard recognition environment includes:

[0020] Extract abnormal environmental status and corresponding standard recognition environment;

[0021] The differences between the abnormal environment state and the corresponding standard recognition environment are marked as temperature abnormality difference, vibration abnormality difference and light abnormality difference respectively;

[0022] The temperature abnormal difference is marked as WDX, the vibration abnormal difference is marked as ZDX, and the illumination abnormal difference is marked as GZX; the adjustment coefficient TJX is calculated by the formula TJX=α1×WDX+α2×ZDX+α3×GZX; wherein α1, α2 and α3 are weight coefficients, which are set according to the influence of illumination, vibration and temperature on the recognition of the piston by the recognition device; α1+α2+α3=1;

[0023] Setting an adjustment threshold; the adjustment threshold is set based on comprehensive cost and adjustment difficulty;

[0024] Determine whether the adjustment coefficient is greater than or equal to the adjustment threshold; if yes, the current environment is not suitable for adjustment, and the recognition environment needs to be transferred or the equipment needs to be adjusted; if no, adjust the abnormal recognition environment to meet the standard recognition environment.

[0025] Preferably, controlling the manipulator to grab the piston once according to the identification data once includes:

[0026] Extracting primary recognition data, contour samples in the database, recognition direction and corresponding historical manipulator angles; wherein the primary recognition data includes the piston center point position and piston contour recognized by the visual device;

[0027] Adjust the position of the manipulator. When the center point of the manipulator coincides with the center point of the piston, start the manipulator to grab the piston to the side recognition area to identify the piston contour.

[0028] The contour samples, corresponding recognition directions and corresponding historical robot angles are used as training data and test data. The training data is used to train the neural network model. After the training, the test data is used to test the trained neural network model. The neural network model is adjusted according to the test results to obtain an angle model whose input data is the piston contour and recognition direction and whose output data is the robot angle. The piston contour and recognition direction are input into the angle model to obtain the robot angle corresponding to the current piston. The angle of the robot is adjusted according to the robot angle. The robot places the piston on the conveyor belt according to the adjusted angle, and the piston follows the conveyor belt to the secondary recognition area.

[0029] Preferably, the controlling the secondary manipulator and generating an alarm signal according to the secondary identification data comprises:

[0030] Extracting secondary identification data; wherein the secondary identification data includes a secondary piston profile obtained by a secondary identification device performing secondary identification on the piston;

[0031] The secondary piston contour and the identification direction of the secondary identification device are input into the angle model to obtain the secondary angle; whether the secondary angle is the same as the adjusted angle is determined; if yes, the secondary manipulator is started to grab the piston to the configuration area; if not, an alarm signal is generated to stop the operation of the equipment.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] 1. The present invention detects the current recognition environment and compares the current recognition environment with the preset marked recognition environment to obtain the environmental state. By ensuring that the environmental conditions meet the standards, the accuracy and reliability of the visual recognition system can be improved, and the current recognition environment can be adjusted according to the environmental state to make it more suitable for recognition; the adjustment coefficient is calculated by abnormal environmental state, and the adjustment coefficient is compared with the preset adjustment threshold to distinguish whether the current environment is suitable for adjustment. In consideration of the comprehensive cost and adjustment capacity, other schemes or measures can be taken to reduce the production cost; the piston is visually recognized by the visual recognition device, and the manipulator is controlled to grab the piston and move it according to the result of the visual recognition, the neural network model is trained by historical data, the contour image and the recognition direction are input as input data into the neural network model to obtain the manipulator angle, the manipulator is rotated according to the manipulator angle to adjust the direction of the piston, and after being placed on the conveyor belt, the piston is secondary recognized by the secondary visual recognition device, and by comparing the directions of the piston before and after transportation, it can be obtained whether there is a problem with the equipment and whether the production task can be completed. An alarm signal is generated according to the comparison result, and the production capacity of the piston assembly equipment can be tested to prevent the production quality and production efficiency from being affected by equipment problems.

[0034] 2. When detecting the recognition environment for identifying the piston, the present invention not only compares the light and vibration that will affect the photographic recognition, but also detects the temperature factor. The temperature will not only affect the efficiency of the devices in the equipment, but also affect the accuracy of the sensor. If the light sensor and the vibration sensor are affected by the temperature and the detection results are not accurate enough, it will also affect the recognition accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0036] Figure 1 The following is a flowchart of an embodiment of the present invention.

[0037] Figure 2It is a system composition diagram of the present invention.

[0038] Figure 3 The figure is a complete flow chart for obtaining the environment status in one embodiment of the present invention. DETAILED DESCRIPTION

[0039] The technical solution of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0040] See also Figure 1-Figure 3 , the first aspect of the present invention provides a piston blank multi-directional photographic recognition system, including: a database, a recognition environment control module, a visual recognition module and a manipulator control module;

[0041] Recognition environment control module: used to detect the current recognition environment, compare the current recognition environment with the preset standard recognition environment to obtain the environmental state, and adjust the current recognition environment according to the environmental state.

[0042] Exemplarily, the light intensity of the current identification environment collected by the light sensor, vibration sensor and temperature sensor set in the piston assembly equipment is 600 lux, the vibration amplitude is 50 microns, and the temperature is 25 degrees; the preset standard lighting conditions in this embodiment are 700 lux to 1000 lux, the standard vibration conditions are 10 microns to 40 microns, and the standard temperature conditions are 20 degrees to 28 degrees.

[0043] Determine whether the current illumination meets the standard illumination conditions; if it is determined that the current illumination does not meet the standard illumination conditions, mark the illumination state of the current recognition environment as abnormal illumination; determine whether the current vibration meets the standard vibration conditions; if it is determined that the current vibration does not meet the standard vibration conditions, mark the vibration state of the current recognition environment as abnormal vibration; determine whether the current temperature meets the standard temperature conditions; if it is determined that the current temperature meets the standard temperature conditions, mark the temperature state of the current recognition environment as normal temperature; obtain the environmental state of the current recognition environment.

[0044] It should be noted that light, vibration and temperature will affect the accuracy of visual recognition equipment to varying degrees. Excessive light, weak light and excessive vibration will affect the image quality, and temperature will affect the accuracy of the sensor and the imaging quality of the equipment, resulting in low visual recognition accuracy of the piston.

[0045] In some other embodiments, if the temperature state of the current identification environment is abnormal, it is necessary to adjust the temperature and re-acquire the data of the light sensor and the vibration sensor to prevent the sensor detection data from changing due to abnormal temperature.

[0046] Determine whether the environmental state of the current recognition environment contains an abnormal environmental state. If it is determined that the environmental state of this embodiment contains abnormal illumination and abnormal vibration, then generate adjustment data according to the illumination, vibration and the corresponding standard recognition environment, and mark the difference between the abnormal environmental state and the closest value of the standard recognition environment as abnormal illumination difference and abnormal vibration difference, respectively. If the temperature meets the standard temperature condition, there is no need to adjust the temperature. The abnormal illumination difference is 100 lux, and the abnormal vibration difference is 10 microns. In this embodiment, weight coefficients α1=0, α2=0.9, and α3=0.1 are set. The adjustment coefficient TJX=19 is calculated by the formula TJX=α1×WDX+α2×ZDX+α3×GZX. In this embodiment, the adjustment threshold is set to 20, and it is determined whether the adjustment coefficient is greater than or equal to the adjustment threshold. If it is determined that the adjustment coefficient is less than the adjustment threshold, the illumination is increased and the vibration amplitude is reduced.

[0047] It should be noted that by calculating the adjustment coefficient and comparing it with the adjustment threshold, it is possible to determine whether the current environment is suitable for adjustment. The setting of the adjustment threshold can further determine the necessity of adjustment after an abnormal environmental state is determined. By determining in advance whether the environment needs to be adjusted, unnecessary adjustment costs can be avoided. Especially in cases where the adjustment cost is high, you can choose to change the operating time, change the operating environment, or adjust the equipment to reduce overall production costs.

[0048] Visual recognition module: used to identify the image of the piston transported by the conveyor belt and obtain primary recognition data and secondary recognition data.

[0049] Exemplarily, the system includes three visual recognition devices, namely a primary visual recognition device, a side visual recognition device and a secondary visual recognition device. The primary visual recognition device is arranged above the visual recognition area, the side visual recognition device is arranged on the side of the side recognition area, and the secondary visual recognition device is arranged above the secondary recognition area, which are respectively used for visual recognition of the top and sides of the piston. When the piston reaches the corresponding area, the piston is visually recognized.

[0050] It should be noted that the visual recognition device includes image acquisition equipment, image processing software and data transmission interface, which can acquire and analyze images, convert digital signals into electrical signals and transmit them to the overall control system.

[0051] A conveyor belt transports the piston blank to a visual recognition area, a visual recognition device is started to capture an image of the piston blank, the image is binarized to obtain a contour map of the piston blank in the image, and the center point of the contour map is marked as the piston center point position, a moving path for the center point of a manipulator to reach the piston center point position is planned, the manipulator is controlled to move to the top of the piston blank according to the moving path and then grab the piston blank, grab the piston blank to the side recognition area, capture an image of the piston blank through a side visual recognition device, capture a side image of the piston blank, and binarize the side image to obtain a side contour of the piston blank.

[0052] Extract the side profile samples, identification direction and corresponding historical manipulator angles of the piston blank from the database, use the profile samples, corresponding identification direction and corresponding historical manipulator angles as training data and test data, the training data is used to train the neural network model, and the test data is used to test the trained neural network model after the training, and the neural network model is adjusted according to the test results to obtain an angle model with the piston profile and identification direction as input data and the manipulator angle as output data; in this embodiment, the side recognition is on the right side of the piston blank, and the piston profile and identification direction are input into the angle model to obtain the manipulator angle corresponding to the current piston; rotate the primary manipulator to the same angle as the manipulator angle, keep the manipulator angle unchanged, place the piston blank on the conveyor belt, and the piston blank moves with the conveyor belt to the transfer area, grab the piston blank to the secondary visual recognition area by the secondary manipulator to collect the secondary piston image of the piston blank, and extract the profile image of the secondary piston image, and input the profile image and the identification direction into the angle model to obtain the secondary angle; judge whether the secondary angle is the same as the adjusted angle. In this embodiment, if the secondary angle is different from the adjusted angle, an alarm signal is generated, the equipment is stopped, and maintenance personnel are dispatched to check the equipment.

[0053] It should be noted that the secondary angle is different from the adjusted angle, indicating that there is a problem with the conveyor belt in front of the transfer area, resulting in different directions before and after transportation, which will affect production, so it needs to be stopped and repaired.

[0054] Part of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is a formula closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0055] Working principle of the present invention:

[0056] The present invention collects environmental data of the identification environment through sensors, and compares it with a preset standard identification environment to obtain an environmental state. If an abnormal environmental state occurs in the environmental state, an adjustment coefficient is calculated based on the abnormal environmental state and the corresponding standard identification environment, and the adjustment coefficient is compared with a preset adjustment threshold value, and a decision is made based on the comparison result whether to perform environmental adjustment. After the environmental adjustment is completed, the piston blank is transported by a conveyor belt, and after arriving at the identification area, the contour and center point of the piston are obtained by a visual recognition device, and a manipulator is controlled to grab the piston to a side identification area based on the position of the center point, and the direction of the piston is identified by a side identification device, and the manipulator is controlled to adjust the direction and place it on a conveyor belt for transportation to a transfer area, and the visual recognition device is used again to detect whether the direction before and after transportation has changed, and an alarm signal is generated based on the judgment result.

[0057] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. A multi-directional photography and recognition system for piston blanks, characterized in that: include: Database, recognition environment control module, visual recognition module and manipulator control module; Recognition environment control module: used to detect the current recognition environment, compare the current recognition environment with the preset standard recognition environment to obtain the environment state, and adjust the current recognition environment according to the environment state; Visual recognition module: used to recognize the image of the piston transported by the conveyor belt and obtain primary recognition data and secondary recognition data; Manipulator control module: used to control the primary manipulator to grab the piston according to the primary identification data, and control the secondary manipulator according to the secondary identification data and generate an alarm signal.

2. The multi-directional photography and identification system for piston blanks according to claim 1 is characterized in that: The step of comparing the current recognition environment with a preset standard recognition environment to obtain the environment state includes: Extracting the current recognition environment; setting a standard recognition environment; wherein the standard recognition environment includes standard lighting conditions, standard vibration conditions, and standard temperature conditions; the current recognition environment includes current lighting, current vibration, and current temperature of the current recognition environment; Determine whether the current lighting meets the standard lighting conditions; if yes, mark the lighting state of the current recognition environment as normal lighting; if no, mark the lighting state of the current recognition environment as abnormal lighting; Determine whether the current vibration meets the standard vibration condition; if yes, mark the vibration state of the current recognition environment as normal vibration; if no, mark the vibration state of the current recognition environment as abnormal vibration; Determine whether the current temperature meets the standard temperature condition; if yes, mark the temperature state of the current identification environment as normal temperature; if no, mark the temperature state of the current identification environment as abnormal temperature; Abnormal illumination, abnormal vibration and abnormal temperature are marked as abnormal environmental states, and the illumination state, vibration state and temperature state are combined to obtain the environmental state.

3. The multi-directional photography and identification system for piston blanks according to claim 1 is characterized in that: The step of adjusting the current recognition environment according to the environmental state includes: Extract the environment state; Determine whether the environmental state of the current identification environment contains an abnormal environmental state; if so, generate adjustment data based on the abnormal environmental state and the corresponding standard identification environment, adjust the current identification environment based on the adjustment data, and re-detect after the adjustment is completed; if not, generate an identification signal and start the visual device to perform the first identification of the piston.

4. The multi-directional photography and identification system for piston blanks according to claim 3 is characterized in that: The step of generating adjustment data according to the abnormal environment state and the corresponding standard recognition environment includes: Extract abnormal environmental status and corresponding standard recognition environment; The differences between the abnormal environment state and the corresponding standard recognition environment are marked as temperature abnormality difference, vibration abnormality difference and light abnormality difference respectively; The temperature abnormal difference is marked as WDX, the vibration abnormal difference is marked as ZDX, and the illumination abnormal difference is marked as GZX; the adjustment coefficient TJX is calculated by the formula TJX=α1×WDX+α2×ZDX+α3×GZX; wherein α1, α2 and α3 are weight coefficients, which are set according to the influence of illumination, vibration and temperature on the recognition of the piston by the recognition device; α1+α2+α3=1; Set the adjustment threshold; determine whether the adjustment coefficient is greater than or equal to the adjustment threshold; if yes, the current environment is not suitable for adjustment, and the recognition environment needs to be transferred or the equipment needs to be adjusted; if no, adjust the abnormal recognition environment to meet the standard recognition environment.

5. The multi-directional photography and identification system for piston blanks according to claim 1 is characterized in that: The method of controlling the manipulator to grasp the piston according to the identification data comprises: Extracting primary recognition data, contour samples in the database, recognition direction and corresponding historical manipulator angles; wherein the primary recognition data includes the piston center point position and piston contour recognized by the visual device; Adjust the position of the manipulator. When the center point of the manipulator coincides with the center point of the piston, start the manipulator to grab the piston to the side recognition area to identify the piston contour. The contour samples, corresponding recognition directions and corresponding historical robot angles are used as training data and test data. The training data is used to train the neural network model. After the training, the test data is used to test the trained neural network model. The neural network model is adjusted according to the test results to obtain an angle model whose input data is the piston contour and recognition direction and whose output data is the robot angle. The piston contour and recognition direction are input into the angle model to obtain the robot angle corresponding to the current piston. The angle of the robot is adjusted according to the robot angle. The robot places the piston on the conveyor belt according to the adjusted angle, and the piston follows the conveyor belt to the secondary recognition area.

6. The multi-directional photography and identification system for piston blanks according to claim 1 is characterized in that: The method of controlling the secondary manipulator and generating an alarm signal according to the secondary identification data comprises: Extracting secondary identification data; wherein the secondary identification data includes a secondary piston profile obtained by a secondary identification device performing secondary identification on the piston; The secondary piston contour and the identification direction of the secondary identification device are input into the angle model to obtain the secondary angle; whether the secondary angle is the same as the adjusted angle is determined; if yes, the secondary manipulator is started to grab the piston to the configuration area; if not, an alarm signal is generated to stop the operation of the equipment.

Images