Machine Vision Camera Selection Guide (1): Choosing the Resolution

Machine Vision Camera Selection Guide: Choosing the Resolution

Machine vision cameras are used for product inspection and dimensional measurement, and are an important tool for accurately capturing object sizes and defects. Here we explain, in plain terms and with concrete examples, how to choose the resolution.

1. What Is Resolution?

Resolution is expressed as the number of pixels on the camera's sensor. It determines how accurately the fine details of an object can be captured. In general, the higher the resolution, the more clearly smaller features can be captured.

2. How to Choose the Resolution: Consider the Minimum Feature Size

One of the key factors in deciding the resolution is the minimum feature size. This is the size of the smallest object you want the camera to detect. To capture this feature accurately, it is recommended that it be covered by at least 3 to 5 pixels. This reduces false detections and yields reliable data.

3. Worked Examples and How to Calculate

Next, let's look in detail at how to calculate the resolution, using two concrete examples.

A. Measuring a person 2 m tall in 1 mm increments

• Object size: 2,000mm

• Minimum feature size: 1mm

• Required number of pixels:
  If the minimum feature is captured by 1 pixel, the required resolution is as follows.
  Required resolution = 2000 mm / (1 mm) = 2000 pixels
  Therefore, a resolution of at least 2000 x 2000 pixels (4 megapixels) is required. For example, a camera with a Sony IMX255 sensor (2448 x 2048 pixels) would be suitable.

B. Detecting a 5 mm scratch on a printed circuit board 300 mm wide and 100 mm high

• Object size: 300mm x 100mm

• Minimum feature size: 5mm

• Pixel coverage requirement: Ideally 5 pixels per feature.

  • Required resolution in the width direction:
    300 mm / (5 mm/5 pixels) = 300 pixels

  • Required resolution in the height direction:
    100 mm / (5 mm/5 pixels) = 100 pixels
    The required resolution is 300 x 100 pixels, but in practice a VGA camera (640 x 480 pixels) is suitable, to allow some margin.

Table 1. Sensor resolutions

Standard	Horizontal	Vertical	Total pixels
HD	1,280	720	Approx. 0.92 megapixels
HD+	1,600	900	Approx. 1.44 megapixels
Full HD (2K)	1,920	1,080	Approx. 2.07 megapixels
Full HD+	2,400	1,350	Approx. 3.24 megapixels
4K	3,840	2,160	Approx. 8.29 megapixels
8K	7,680	4,320	Approx. 33.17 megapixels

Summary

Resolution is a key factor in selecting a machine vision camera. Choosing a camera with the right resolution for the size of the object and the smallest feature you want to detect makes accurate inspection and measurement possible. Optimizing the resolution should give you more reliable inspection results.