KR19990060082A - Automatic Image Recognition Method of Component Mounter - Google Patents

Abstract

The present invention relates to an image recognition automatic recognition method of a surface mounter, which inputs basic data about a component to an image recognition unit, classifies the type of the component conveyed by the supply device into data stored in the image recognition unit, and at the same time. It decides which nozzle, camera, lighting and suction head to use, and after absorbing the parts, it is transferred to the commanded camera side and imaged in 3D again. Create a template for the analyzed part in the state, check it with the database, compare the generated template with other newly adsorbed parts to recognize the part, and repeatedly determine whether the part has the compensation value for the template within a certain value. After storing the data, the image recognition unit that receives the image captured by the camera Enter a minimum number of items of basic data of the product, build a database that inputs other data, and input the data into a separate database to shorten the recognition time for the parts and the number of types mounted on the printed circuit board. The data for many parts can be easily recognized, and in particular, it can cope with model changes or small quantity production of a variety of products, thereby improving productivity.

Description

Automatic Image Recognition Method of Component Mounter

The present invention relates to a method for automatically recognizing an image of a component mounter, and more particularly, to input a minimum data that is a reference for image recognition of a component in advance, and other data are automatically measured and inputted by the mounter. By shortening the programming time for the parts, as well as improving the production efficiency, and more particularly, it relates to the automatic image recognition method of the component mounter that can flexibly respond to the model exchange.

In general, the mounter for mounting the surface-mount electronic component on the printed circuit board recognizes the distortion caused when the component is sucked by the vacuum pressure, and then corrects the position by the amount of the mounting to mount the component at the correct center position of the printed circuit board. As shown in FIG. 1, the mounter is capable of absorbing (holding) while moving the device supplied from the supply device 10 and the device supplied from the supply device 10 in an up and down direction, as shown in FIG. 1. Recognizing the components sucked by the nozzle 12 and the drive actuator 16 for transferring the suction head 14 in the X and Y axis directions to the center of the component with the nozzle 12 It is composed of a part recognition device 18 for.

The component recognition apparatus includes a camera for capturing an image of a component, an image recognition unit connected to the camera, and a monitor displayed so that an operator can visually check an image of the component recognized by the image recognition unit.

In the image recognition unit, recognition (specification) for various kinds of parts is input. That is, the information of the basic data {name, type, lead type, focus and size (width x length x thickness)) for each part, and feeder data (type of feeder, supply direction, quantity and Information of the tray supply data) and operation data (movement speed, adsorption time for adsorption, mounting and rotation, height of adsorption, position correction, adsorption detection vacuum degree, camera designation, illumination) and image recognition data (lead shape) Data related to component image recognition, such as information on leads on each side (width, pitch, measurement center distance, lead length, number of leads, number of leads removed from each side), and the like.

As such, in order to input data for each component, a conventional method is input by the image recognition recognition method as shown in FIG. 2.

That is, after the input step (S10) and the input step (S10) for inputting the parts, feeder, motion and recognition-related data, the search step (S20) and the search step (S20) for searching the data, and then the template generation step Search step (S40) to search the recognition template and the actual part generated in (S30) and the recognition step (S50) to recognize the image by retrieving the image by repeating the search 2-3 times by absorbing the actual part and the part after recognition In step S60 of determining the change of the determined value, and if the recognized value of the determined part is within a predetermined value, the operation is recognized as a storage step S70 of storing data related to image recognition.

At this time, if there is deleted data from the above data in the search step (S20) for the parts and the mounting device, if the recognition template is insufficient in the process of checking the recognition template, the routine is repeatedly executed if the data is insufficient. Modify or reenter the data to save the data for one part.

Thereafter, when the parts supplied by the feeder are picked up by the camera, the picked-up parts image is transferred to the image recognition unit, and the image recognition unit outputs the pre-stored data to identify the same as the adsorbed parts in the search step S40. After the judgment, display on a monitor to determine the center position of the part.

Then, the adsorption head 14 adsorbs the parts in the state of being transferred in the X-axis and / or Y-axis direction along the drive actuator 16 to the center position of the parts to mount the parts on the printed circuit board.

However, in the conventional image recognition recognition method as described above, since the operation of the feeder and the suction head and the information of the part image recognition data are inputted according to the individual parts, a small number of kinds of parts are easily mounted on the printed circuit board. However, when the number of types to be installed is large, it is necessary to input many data items to recognize each part, and it takes a lot of input time, which causes worker fatigue. There was a problem of deterioration.

The present invention is to solve the above problems, an object of the present invention is to reduce the entry of the basic data (size, supply lane and type of parts) of the parts to the image recognition unit that receives the image captured by the camera to the minimum In addition, by constructing a database in which other data is input, the data is input to a separate database to shorten the recognition time of the parts, and to easily recognize the data of a large number of types mounted on the printed circuit board. In particular, the present invention provides an automatic image recognition method of a component mounter that can cope with a model change or a small quantity production of a variety of products, thereby improving productivity.

1 is a view schematically showing a general component mounter,

FIG. 2 is a conventional flowchart for recognizing image recognition of a component adsorbed by the component mounter of FIG. 1;

3 is a flowchart illustrating an image recognition automatic recognition method of a component mounter according to the present invention.

Explanation of symbols on the main parts of the drawings

S100: input step S110: determination step

S120: part adsorption step S130: analysis step

S140: confirmation step S150: adsorption step

S160: recognition step S170: determination step

S180: Save step

Features of the present invention for achieving the above object, the input step of inputting the basic data for the part to the image recognition unit; A classification step of classifying the types of parts conveyed by the supply apparatus by the input step into data stored in the image recognition unit, and determining which of nozzles, cameras, lights and suction heads for the parts are to be used; An analysis step of adsorbing the component in the determining step, transferring the component to the commanded camera, and then imaging in 3D to analyze the data of the component as data previously stored in the database; Generating a template for the component analyzed in the analyzing step and confirming the result in a database; A recognition step of recognizing the part by comparing the template generated in the step with another newly adsorbed part; A determination step of repeatedly determining whether the component corrects a template within a predetermined value in the recognition step; The step of storing the data identified in the determination step is accomplished by automatically recognizing the image recognition of the part.

Hereinafter, a preferred embodiment of the automatic image recognition method of the component mounting machine according to the present invention will be described in detail with reference to the accompanying drawings.

Since the configuration of the component mounter according to the present invention is the same as the configuration shown in FIG.

3 is a flowchart illustrating an automatic image recognition method according to the present invention, which includes a camera for capturing a component, and an image recognition unit is connected to the camera to display an image of the component on a monitor. The database is connected.

In the image recognition unit, basic data of parts, that is, information of minimum items (size, supply lane, type of parts) for recognizing the image of the parts are input.In the database, lighting, lead information, and exposure information according to the type and type of parts are input. Data such as visual field information, adsorption and mounting speed for each component type, and adsorption head for adsorption of components and nozzles mounted thereon are input.

Thereafter, when data on the size, supply lane, and type of the parts transferred by the supply device 10 are input to the image recognition unit in the input step S100, the image recognition unit classifies the parts and then absorbs the parts. After deciding which nozzle, camera, and lighting and adsorption head to use, from the determination step S110, the component is adsorbed by the component adsorption step S120, transferred to the commanded camera side, and imaged again to obtain the component size and lead. Analyze information and location information.

The picked-up component is analyzed by the data already input to the database in the analysis step S130, and then generates a template, and is confirmed by the database from the verification step S140. The template is then displayed on the monitor for visual and visual confirmation.

In this way, after generating a template that is a reference to the other components by the adsorption head nozzle by the adsorption step (S150) and then compared to the template to recognize the components in the recognition step (S160). In other words, after positioning on the template, it is determined repeatedly by the determination step S170 whether the correction value for the template is within a predetermined value.

If it is within the correction value, the data is stored from the data storage step (S180). If the correction value is greater than or equal to, the part is recognized again to generate a new template, and then the data is stored in comparison with the correction value for the template.

Although the component is measured in three dimensions and inputted as basic data in the present invention, the component captured by the camera can be automatically input as three-dimensional data.

As described above, according to the image recognition automatic recognition method of the component mounter according to the present invention, the image recognition unit receiving the image picked up by the camera is input to the image recognition unit with the minimum items of the minimum data, and other data is input. By constructing the input database, the data is input to a separate database, which shortens the recognition time for the parts, and can easily recognize the data for the large number of parts mounted on the printed circuit board. It can cope with change or small quantity production and improve productivity.

The present invention is not limited to the above-described embodiments, and various modifications can be made by those skilled in the art without departing from the gist of the present invention as claimed in the claims. Such changes are intended to fall within the scope of the claims.

Claims (4)

An input step of inputting basic data on the part to the image recognition unit; A classification step of classifying the types of parts conveyed by the supply apparatus by the input step into data stored in the image recognition unit, and determining which of nozzles, cameras, lights and suction heads for the parts are to be used; An analysis step of adsorbing the component in the determining step, transferring the component to the commanded camera, and then imaging in 3D to analyze the data of the component as data previously stored in the database; Generating a template for the component analyzed in the analyzing step and confirming the result in a database; A recognition step of recognizing the part by comparing the template generated in the step with another newly adsorbed part; A determination step of repeatedly determining whether the component corrects a template within a predetermined value in the recognition step; And automatically recognizing image recognition of the component by storing the data identified in the determination step. According to claim 1, wherein the component adsorbed in the determination step is imaged in two dimensions by the camera, the image recognition of the component mounter further comprises analyzing the imaged data in advance stored in the database Cognitive Method. The image recognition of the component mounter according to claim 1, wherein an image of the component is displayed on the monitor through an image recognition unit in order to visually identify the component located on the template in the recognition step of the component image. Automatic recognition method. The image of the component mounter according to claim 1, further comprising a template generation step of re-recognizing the component and generating a new template when the correction value of the component for the template is greater than or equal to a predetermined value in the determination step. Recognition automatic recognition method.