Right inspection system for finding defective pieces of a system or machine is very important as right inspection decides the faith of working of a machine. If parts of machine are defective, the investment into the making of machine may go waste. Hence, inspection starting from machine testing to human testing is very crucial.
Basically, a specimen test is carried out through vision sensors, photoelectric sensors, human eye inspection, etc. Today, however, the most advanced vision sensors are combining the best features of those older methods with unprecedented advantages in speed, accuracy, size and cost.
A vision sensor works by conducting scanning and analyzing. It first collects visual data through the camera, analyzes it in the microprocessor and makes a decision based on that analysis. The analysis is done with the help of software that has some images stored in its memory. The software compares the extracted information with data provided by the user. This includes counting, locating, verifying, measuring, and identifying objects in the image. This helps the sensor to differentiate between good and defective piece.
Finally, during the determination phase, the software determines whether an inspection has passed or failed, according to the predetermined criteria.
The sensor then outputs its results to the machine.
A good inspection of a product is only possible if it has been analyzed taking all parameters into consideration like location, or orientation; shape; light intensity and coun.
These parameters are crucial and have to be taken into account because of following factors:
- Orientation of specimen to be tested gives an idea where and in what direction specimen lies.
- The shape of specimen gives an idea about dimension of specimen.
- The light intensity over specimen allows accurate inspection.
- Count of specimen also helps to determine features of specimen.
Illumination is the most important aspect of specimen inspection. The type of illumination and its location are crucial for creating a high-contrast image that has the greatest possible contrast between the target and the background
Hence to insure a good quality image, it is necessary to choose right type of light source like LED, fluorescent, halogen, etc and to locate the light source relative to the part so as to maximize contrast and resolution.
Image capturing is done by sensor`s lens. This occurs when the light reflected off the target is focused through the sensor`s lens on to the imager chip. The imager comprises of pixels (an array of tiny light sensitive cells), which convert the light into an image.
The resolution of the imager is measured in terms of the number of pixels. Generally, size of image is 640 x480 pixels.
The brightness of the signal is converted to analog signal. The image is then digitized by the Analog to digital converter and stored as an 8-bit gray scale, video signal. An analog signal can have any value but a digital signal has a fixed value: 1, 3, 6, or 8 bit. Each output digital value is the closest to the allowable analog value.
Vision sensor analyzes the captured images using algorithm.
There are two types of algorithm for analyzing the specimen undergoing the scanning test. These two algorithms are linear algorithm and area algorithm.
Both look for a small variation (defect) in the image, but in different ways.
· A linear algorithm is better when the area of interest can be determined, because they are quick and more reliable than area tools.
· An area algorithm is the required when the location of the target could change. For instance, this algorithm could scan and analyze the specimen, and would search for deviation from normal state.
The vision sensor ability to operate at much higher speeds; conduct repetitive, multiple, and consistent inspections all at significantly lower cost is making Vision sensor to evolve as a great player in industrial application. In fact, vision sensor has the potential to change the working of industrial world, making working process quicker and simpler. But, the challenge doesn`t end here. In the upcoming days, the biggest challenge for the engineers will be to manufacture a foolproof vision sensor.