KR100276430B1 - Ccd camera based automated visul inspection system and met hod to the rear side specification in an autobody assembly process - Google Patents

Abstract

PURPOSE: An automatic appearance testing system and its method are provided to automate the entire processes. CONSTITUTION: An automatic appearance testing system comprises a computer(30) to analyze an image and to check a tested object; a sensor(40) to inform a loading of a car body; an image processing board(50) to memorize shape information by inputting an image of a camera; CCD(Charge Coupled Device) cameras(60,70) to take images of a rear portion in the car body; a tag reader(80) to offer a model and specification of the car body by reading information of a tag; and a controller(90) to stop and operate the car body and to inform through a warning lamp. An emergency stop signal is detected. An emergency stop of the car body is checked. A sensor signal is detected. Information of a stopped vehicle is analyzed. Images of a roof and a rear portion are tested. A defect is tested and the defective car body is processed. A conveyer is driven.

Description

Automatic Appearance Inspection System and Method for Rear Specification of Automobile Body Assembly Using CD Camera

The present invention relates to an automatic appearance inspection system and method for the rear specification of a vehicle body assembly using a CCD camera, and more particularly, to the specification of the roof of the vehicle body and the left side lamp device and the right lamp at the rear when the automobile body assembly is produced. A method and apparatus for determining whether a device is assembled into parts of a specification suitable for a car model.

Conventionally, it was manually determined whether the mounting parts of the rear lamps mounted by the operator were assembled into suitable parts. However, as many different cars enter the assembly line on the same production line, parts are sometimes replaced by parts of different models. Since many cars are constantly entering the body assembly production line, it is often the case that an erroneous assembly is not found due to fatigue or other thinking of the operator. If there are any misassemblies, they will be found in the final inspection process after the vehicle has been assembled. The loss is enormous. The economic loss of using the entire car and all the work in the meantime are futile. The loss caused by this mistake is a loss of manpower input and time that has been started since the misassembly, which can lead to the inability to use the entire body or enormous losses in manpower, time and cost of repairing parts. Secondly, the failure rate is higher, which lowers the quality of the car body. The only way to prevent this loss was to be careful to prevent assembly by humans.

Accordingly, the present invention provides a method for automatically finding a defect assembly to be discovered by a person using a computer and a CCD camera to solve the above problems. Using the computer and machine vision technology, it is possible to determine in real time whether the mounted rear lamp unit is incompatible with the car model, to emit a warning sound when misassembly, and to stop the line for maintenance. In this case, it aims to automate all the processes to make the line progress when assembled correctly.

In order to achieve the above object, the shape of the part to be inspected is stored in a computer and compared with the shape of the part of the body to be inspected, if it is the same as the shape of the body model to be inspected; If there are several types of bodies, the shape of each part is stored and stored for several models.When the body to be inspected comes in line, the information about the body model is read first, and then the shape of the object to be inspected and the shape of the model stored By comparison, it is determined whether the shapes are similar.

In the present invention, the configuration and the system diagram of the entire inspection system are described, and the technology for storing and comparing the shape of the rear lamp device is described in the present invention.

1 is a schematic view of the automatic appearance inspection system of the vehicle body specification according to the present invention

2 is a vehicle inspection process flow chart according to the present invention.

3 is a process flowchart of the conveyor belt stationary vehicle information analysis method in FIG.

4 is a process flowchart of a roof burn inspection method in FIG.

FIG. 5 is a process flowchart of the rear image inspection method in FIG. 2; FIG.

FIG. 6 is a flowchart illustrating a method of processing a bad vehicle body in FIG. 2.

* Explanation of symbols for main parts of the drawings

10: conveyor

20: inspection vehicle

30: computer

40: sensor

50: image processing board

60, 70: CCD camera

80: Tag reader

90: controller

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

1 is a schematic diagram of an automatic appearance inspection system of an automobile body specification according to the present invention, FIG. 2 is a flowchart of a vehicle body inspection process according to the present invention, FIG. 3 is a flowchart of a conveyor belt stationary vehicle information analysis method of FIG. FIG. 4 is a process flowchart of the roof image inspection method in FIG. 2, FIG. 5 is a process flowchart of the rear image inspection method in FIG. 2, and FIG. 6 is a process flowchart of the bad vehicle body processing method in FIG. 2.

As shown in FIG. 1, the automatic appearance inspection system of the vehicle body specification of the present invention includes a computer system 30, a sensor 40, a camera 60, 70, a tag reader 80, and software (second). A device that automatically detects misassembly by using two systems. The system is largely composed of hardware and software.

1. Configuration and function of hardware part are as follows.

Reference numeral 10 causes the inspection vehicle to enter the conveyor. 20 moves on the conveyor to the inspection vehicle. 30 is a computer which interprets the image and determines whether the test object is a pass or fail. 40 informs the arrival of the car body with a sensor. 50 inputs an image of a camera to an image processing board to store shape information. 60 and 70 take images of the rear portion of the vehicle body with a CCD camera. 80 is a tag reader that reads the tag information attached to the vehicle body and provides model and specification information of the vehicle body. The 90 serves as a controller to stop the vehicle, resume operation and sound a warning light.

2. The composition and function of the software are as follows.

As shown in FIG. 2, the vehicle 20 enters the inspection site by using the conveyor belt 10 in the vehicle body production line, and the system performs inspection in real time on a specific inspection site input by two cameras, and displays the result. Inform

In the automatic appearance inspection method for the rear specification of the vehicle body assembly using a CCD camera, the vehicle body exterior inspection method detects an emergency stop signal for a vehicle entering the inspection site by a conveyor belt in the vehicle body production line (step 100). If the emergency stop is not performed (step 200), if the emergency stop does not occur, the process returns to the previous step 100, and in case of an emergency stop, the sensor signal is detected (step 300), and the conveyor belt stop vehicle information analysis (step 400) is performed. If it is not the inspection vehicle, the vehicle is driven (step 1000) to return to step 100, and if it is the inspection vehicle, roof image inspection (step 600) and rear image inspection 700 are executed. If it is not good (step 800), and if it is bad, perform a bad body treatment (step 900). System during the inspection sikieo driving the conveyor 1000 and performs the step 100 to repeat steps 1000.

This inspection algorithm will be described in more detail with reference to FIGS. 2 to 6.

1. First, when the system is operated and the inspection is performed, the system checks the sensor signal every moment to determine whether the vehicle 20 has entered (steps 100 to 300 of FIG. 2).

2. When the inspection vehicle 20 is detected by the sensor 40, the detection signal is transmitted to the PLC and the system at the same time at this time, the PLC stops the conveyor 10 in response and the system starts the actual inspection. (Step 400 of FIG. 2).

3. Model identification, one of the important information for inspection, consists of reading and interpreting the tag attached to the bottom of the vehicle. To this end, the Tag reader (RFID) delivers the information contained in the tag to the system, and the system interprets the received information and checks the necessary information (type of vehicle, P / O number, presence of roof panel holes). Used to perform However, since a bad tag is sometimes found, if the system does not immediately receive the information sent by the tag reader, the system re-requests the information transmission from the tag reader. If the operation is repeated three more times, and there is no proper response, the system causes the operator to review the situation by buzzing the alarm with a tag bad message (steps 400 to 403 of FIG. 3).

4. If the model is found in the analysis of the tag in FIG. 2, first, the vehicle that is not the inspection target is transferred to the next line without inspection, and the inspection starts with the inspection of the roof panel (step 500 of FIG. 2).

5. When the image of the roof panel is input from the camera, the system searches for a predefined inspection area (ROI) in the image, finds a part similar to the registered reference template, and checks the various inspection results accordingly. Display through a monitor (steps 600 to 603 of FIG. 4).

6. Then, when the camera receives the image of the rear lamp area, the system checks the left and right rear lamp area and the key hole part of the trunk lid in the same way as in step 5, and then the test result. The distance between the left lamp and the right lamp, the left lamp and the keyhole is calculated for the three parts having the high similarity, and it is checked whether it is within the registered threshold value (steps 700 to 703 in FIG. 5).

7. In Fig. 2, if the above 5 and 6 results meet various criteria related to each model, the system finally displays the result of the pass, otherwise fail, on the monitor and informs the operator. If a vehicle is found, a warning light and a buzzer are generated respectively so that the workers can take appropriate action accordingly, and then press the reset switch to restart the inspection (steps 900 to 903 of FIG. 6).

As described above, when the automatic appearance inspection system of the present invention is put into the intermediate process of the assembly production line of the vehicle body, all misassembly can be prevented in advance. The effect of preventing such misassembly is largely as follows.

First, it improves the quality of automobiles in terms of quality, thereby enhancing its competitiveness.

Since automobiles are made by assembling 20,000 parts, the goal of automobile production technology is to eliminate defect rate by 0.0005% in each process to minimize defects. In order to achieve this goal, the inspection system is put in the middle of the assembly process as in the present invention, so that a system capable of detecting 100% of defective products must be operated to make a product that can compete with foreign countries.

Second, to prevent manpower and economic losses due to assembly.

If the misassembly of a vehicle is found in the final inspection process, the entire vehicle may become a waste. Therefore, all the manpower and time spent in assembling the car will be lost, and material loss will be caused to the car itself, causing damage to the production cost of the car. Since these cases often occur, the use of this system is an effective invention that can prevent losses.

Claims (6)

In the appearance inspection system of the vehicle body 20 which enters the inspection place by the conveyor belt 10 in the vehicle body production line, the computer 30 which analyzes an image and determines whether the inspection object passes or fails, and the arrival of the vehicle body Sensor 40 for notifying the user, an image processing board 50 for inputting an image of the camera to store shape information, CCD cameras 60 and 70 for taking an image of the rear portion of the vehicle body, and a tag attached to the vehicle body. Automated appearance inspection system comprising a tag reader (80) for reading the information of the vehicle body and provides the model and specification information, and a controller (90) to stop and restart the vehicle body and to sound a warning light. In the automatic appearance inspection method for the rear specification of the vehicle body assembly using a CCD camera, the vehicle body appearance inspection method comprises the steps of detecting an emergency stop signal for the vehicle entering the inspection site by the conveyor belt in the body production line; A step 200 for checking whether the emergency is stopped, a step 300 for detecting a sensor signal, a step 400 for interpreting the conveyor belt stop vehicle information, a step 500 for checking whether the vehicle is an inspection target, a step 600 for inspecting a roof image, Steps 100 to 100 are performed during a rear image inspection step, a step 800 for executing the step 700, a step 800 for checking for defects, a step 900 for processing a bad car body, and a step 1000 for driving a conveyor and performing an appearance inspection. Automatic appearance inspection method, characterized in that to repeat 1000 The method of claim 1, wherein the conveyor belt stop vehicle information analysis method of step 400 is performed by stopping the conveyor belt, reading the tag information 402, and analyzing the vehicle information and the specification information by the tag 403. Auto Appearance Inspection Method Featuring 2. The automatic appearance inspection method according to claim 1, wherein the roof image inspection of step 600 is performed by inputting a car roof image, step 601, analyzing an image, and inspecting and determining 603. The rear image inspection method according to claim 1, wherein the rear image inspection method of step 700 is performed by inputting a rear body image of the vehicle body, acquiring vehicle body type identification, image analysis, and searching for a reference template 702 and inspection determination step 703. Auto appearance inspection method The method of claim 1, wherein the method for treating the bad body of the step 900 is performed by the step 901 of sounding a warning light and a buzzer, the step 902 of stopping the buzzer sound and the operator removing the bad body, and turning on a switch to continue the inspection. Auto Appearance Inspection Method Featuring

Images