Computer vision has been with us for a few years now, but there are still doubts and mistakes when applying it. In today’s post we give you 5 tips you should know about computer vision. So that applying it in your project is easy, comfortable and effective. Let’s get started!
Having more camera resolution is not always better
The resolution tells us the number of theoretical pixels generated by the image of the camera sensor we have selected. In those pixels are stored details such as color, shadows and contrasts, among others. For example, if we are analyzing a 30 mm part and we process it at 720 pixels, we will have a (theoretical) resolution of 0.04 mm/pix. That is why we must take into account that this value is not perfect, since there are parameters such as aberrations or focus. But it does not end here, there are techniques to improve resolution by image processing that can be used when necessary. The important and primordial thing is to select the appropriate optics for the size of the object and the defect we are going to analyze.
Not just any light can be used
Lighting is key in a computer vision application. Selecting the right lighting makes it possible to detect defects more reliably. In addition, by selecting the right source it is possible to visualize defects that would otherwise not be detectable in the application. It’s fantastic!
If I see it ‘by eye’, the camera will be able to see it. No!
This phrase that has been heard so much is not true. Our human eye has, according to a study, 576 MPx, that is, much more than a smartphone. If we bring the part closer to the eye (keeping the focus) we can achieve higher resolutions than those of many cameras, as well as being more adaptive or placing the part in the most suitable angle of light. Therefore, to check whether an application is suitable for machine vision processing and to verify its suitability, experience and testing are required.
More sensitive computer vision system will give you more false negatives
When configuring the upper and lower limits for the machine computerreport a part as defective, it must be taken into account that if the sensitivity of the system is increased, some parts will be rejected when, in fact, they are not. Therefore, at this point it will depend on the necessary confidence in the application to configure the system in one way or another.
Computer vision can only be applied in laboratories. No!
This sentence is a false belief. ATRIA has been installing and developing robust and reliable computer vision systems for factories for years. Factories in the automotive, agri-food, electronics and pharmaceutical sectors already have these systems in their production or assembly lines. Being able to count on these systems in the factories, allows a 100% inspection of the parts, so that the quality process becomes more reliable.
Do you want us to implement a machine vision system in your process? Contact us!
- Position detection and presence of stickers on plastic parts
- Automatic cuvette cleaning
- Detection of defects in pressed parts