Application of machine vision products in quality inspection of LED car lights

Machine vision detection technology is a non-contact detection method based on machine vision methods, which comprehensively uses technologies such as image processing, precision measurement, pattern recognition, and artificial intelligence. The basic principle is to analyze the measured target image obtained by the machine vision system to obtain the required measurement information, and to determine whether the measured target meets the specifications (that is, qualified or unqualified) according to the existing prior knowledge.

With the popularization and application of computer integrated manufacturing systems, traditional manufacturing enterprises are seeking to develop towards flexibility and automation. In the process, the demand for computer-assisted quality systems has been raised. With the popularization of flexible manufacturing systems, people have put forward high requirements on the intelligence, flexibility, and speed of detection systems to meet the needs of multi-variety and small-batch production. This is also the origin of the current popular Industry 4.0. Because machine vision systems can obtain a large amount of information quickly, and are easy to integrate with design information and process control information, machine vision systems for working condition monitoring and field monitoring have been widely used in industrial, commercial, and other fields. Among them, there is a very important application field, that is, the automotive industry. This article is based on the application experience of iJUNCO, taking the quality inspection of LED lights as an example to introduce our products in the automotive industry.

LED car light detection

With the rapid development of the automotive industry, the requirements for car lights are getting higher and higher, and the design of new car lights has attracted more and more attention. New, intelligent LED lights are one of the dazzling products in automotive electronics. In automotive lighting systems, the rear light assembly is a key component of driving safety. There are corresponding standards and specifications for the manufacture and installation of lighting components, and they are strictly adjusted before leaving the factory to ensure that the lighting meets the standards.

  The headlight assembly is a combined lighting component. Under the current existing technical conditions, the quality of the LED rear light assembly is only tested for a single item of lighting performance, and manual visual inspection methods are used, which directly affects product quality and production efficiency . In fact, the quality inspection of the lamp assembly should also include the detection of the power of the light source and the quality of parts assembly.

  Introduce machine vision technology to the quality inspection of LED headlight assemblies, use image processing methods to detect the lighting performance of LED headlights, comprehensively inspect the light source power and component assembly quality of headlight assemblies, etc. The quality of different types of lights is comprehensively tested, which is highly flexible and intelligent.

  The system uses machine vision technology and virtual instrument technology to design an effective image acquisition system, which includes MV-EM / E series CCD and CMOS Gigabit Ethernet industrial cameras, megapixel industrial lenses, trigger devices, etc. Based on the lighting image, research and apply digital image processing technology to obtain the lighting quality of the car lights in real time, and judge the failure and area of LED lighting, and realize the online dynamic detection of LED light power and light intensity parameters. The system can complete the automatic detection of various types of car lights through screen settings or software adjustments. The image processing process of the lighting performance of the LED rear lamp assembly is shown in the figure:

Image processing process of LED car light assembly lighting performance

After processing through a series of image processing methods, the size of the particles in the positioning interval can be analyzed to determine whether the LED light in the area is invalid. Large particles are obviously missing in the area shown in the figure, indicating that the LED has a failure area.

The system introduces machine vision technology to the LED vehicle light detection system. It realizes the flexibility and flexibility of LED vehicle lighting performance and light source power detection through coarse image positioning and LED gray particle statistics methods. It is a set of parts detection equipment. By adjusting the camera system and setting the software parameters, the system can also detect different models of LED car light products, which has better flexibility. At present, this type of inspection equipment using Vessel image digital cameras has been successfully applied Among different brands such as Mitsubishi and Audi, they have been well received by users. With the increase of the types of LED lights used in the future and the increase in the number of LEDs for each set of lights, the size of the automotive LED market will also continue to increase, with broad market prospects.