In machine vision applications, the selection of light sources and lighting methods are crucial for image quality. For different detection objects, to highlight their measured features, specific requirements are placed on the structural shape, luminous angle, illumination size, etc. of the light source, resulting in vastly different imaging effects. A suitable light source can achieve optimal separation of target features and background information in the image, greatly reducing the difficulty of image processing and improving the accuracy and reliability of the system. This article combines the extensive project experience of Hanming to provide a brief understanding of how to choose light sources for different objects.
1、 Basic ideas
① Understand project requirements and clarify specific testing content, such as appearance inspection, OCR, size measurement, positioning identification, etc;
② View the material characteristics of the detection target, analyze the difference between the target and the background, and find the maximum difference in imaging between the two;
③ Ask the customer if there are any restrictions, such as production line light field, working distance, light source size, field of view, etc;
④ According to the specific situation, use actual light sources for repeated testing to obtain the best image effect, which is the best light source scheme.
2、 Lighting Techniques and Application Scenarios
2.1 Back light
Back lighting refers to placing a light source behind the tested object to highlight the product's contour, which can create a dark contour of the object against a bright background.
The most common applications are for detecting the presence of holes and gaps, edge detection, size measurement, and other requirements.
Backlight sources can be selected for illumination, which are mainly divided into the following three categories:
Blue backlight: With a short wavelength and weak penetration, it is generally used for size detection and defect detection of transparent products, such as glass and transparent films.
Infrared backlight: With a long wavelength and good penetration, it is generally used for the penetration of dark and translucent materials, such as dark oral liquid, dark liquid, and dark semi transparent leather detection.
White backlight: Good universality, high relative brightness, suitable for most application scenarios.
Application of Backlight Source in Detecting the Size of Drug Bottle Mouth
Application of Backlight Source in Screw Size Detection
2.2 High angle light
The direction of the line is relatively perpendicular to the detection surface, and the reflection on the flat surface is relatively easy to enter the lens, resulting in a brighter display in the image. Uneven areas such as dents, scratches, and other complex surface structures result in scattered reflections, with only a small amount of light being refracted into the lens, resulting in a darker effect in the image.
The most common applications include character recognition, pit and scratch detection, and size measurement.
High angle circular light source, coaxial light source, strip light source, and surface light can be selected for lighting.
Application of High Angle Ring Light Source in Nut Appearance Defect Detection
2.3 Low angle light
The direction of the line is relatively close to and parallel to the detection surface, and there is relatively no reflected light entering the lens on the flat surface. In the image, dark and uneven areas are displayed, such as pits, scratches, and other complex surface structures, with scattered reflections. Only a small portion of the light can be refracted into the lens, resulting in brighter performance in the image.
The most common applications include character recognition, pit and scratch detection, and size measurement.
Low angle circular light source, strip light source, and linear strip light can be selected for lighting.
Application of Low Angle Strip Light Source in Nut Appearance Defect Detection
2.4 Shadowless light
It combines the effects of high angle light and low angle light, allowing the tested object to be illuminated from multiple angles, weakening surface texture and wrinkles, and ensuring overall uniformity of the image.
The most common applications are positioning recognition, size measurement, and surface inspection of curved products.
Dome light source, circular shadowless light source, and light box can be selected.
Application of Shadowless Light Source in Flange Appearance Defect Detection
In fact, in addition to the above lighting methods, there are many other types of light sources to choose from, and valuable experience needs to be accumulated through continuous verification and testing. For example, in machine vision applications, attention should also be paid to the combination of object color and light source color. When shooting an object, if we want to color it white, we need to use a light source that is the same or similar to that color. If we want to color it black, we need to choose a light source that has a significant wavelength difference from the target color. The core idea of machine vision is to extract the required image features for the convenience of the vision system in the next step of work. Choosing the appropriate light source and lighting method can simplify the operation of the software system and improve the accuracy and stability of the entire system.