“We need a camera for inspection.” It is a familiar starting point, but not enough to select the right product. One application may be covered by a configured vision sensor. Another may need an industrial camera, a separate computer and software owned by the customer or an integrator.

The deciding question is not which camera has more megapixels. First establish: do you need images or the result of a working inspection, and who will implement the processing?

1. Do you need an image or an inspection result?

An industrial camera is one building block in PC-based machine vision: it supplies images that software on the computer evaluates. The final output may be a dimension, position, count or pass/fail signal. Purchasing the camera alone does not provide those results.

A configured vision sensor, such as Baumer VeriSens, processes images using its supported inspection tools and can communicate results to the control system. Configuration and machine integration are still necessary. The question is whether the selected variant’s toolset covers the task under actual production conditions.

These are not inherently “good” and “bad” architectures. They involve different engineering scope and freedom. A well-defined inspection does not automatically justify a custom PC-based system; equally, custom algorithms or complex multi-camera processing should not be forced into an unsuitable toolset.

2. Which architecture fits the task?

There are further options, including programmable smart cameras such as the Baumer AX family. These are neither simply configured sensors nor cameras that only forward images. Custom software and its long-term owner remain important in this architecture too.

Integrated optics and lighting do not, by themselves, mean a completed inspection. Baumer IX all-in-one GigE Vision cameras, for example, are compact imaging components: image processing still has to be provided. Examine the product’s function, not just the shape of its housing.

3. What makes the system complete?

Usable images and reliable decisions require the whole chain to fit. In simple terms: part → illumination and optics → image acquisition → processing → machine response.

• Optics: select around field of view, working distance and the feature to inspect. Sensor size, lens and mechanical space must be considered together.
• Lighting: make the relevant difference visible on the actual surface. More megapixels do not automatically fix unsuitable illumination on reflective material.
• Data transfer and computer: interface, cable length, camera count and processing load together determine the required architecture.
• Environment: dust, water, cleaning, temperature and vibration can affect the camera, lens, connectors and installation. Check the protection of the complete assembly.
• Machine communication and acceptance: when is an image acquired, how does the result reach the controller, and what happens if an image is invalid or missing?

Not every application needs every accessory. The objective is to cover the required functions, not to produce the longest possible component list.

4. Two questions that quickly clarify the direction

“Do you have your own vision team or an appointed integrator?” If so, establish the images, interface and software environment they need. If not, discuss how the application will be implemented before choosing a camera reference.

“What counts as a successful inspection?” Detecting a missing component and checking a dimensional tolerance can require different optics, software and testing. Define success using good and defective samples, production variation and agreed acceptance criteria.

More megapixels are not the same as better measurement accuracy. Field of view, contrast, lens, calibration, object motion and the complete measurement method all affect the result.

5. What should you provide before requesting a quotation?

1. What must be detected or measured, and what result does the controller need?
2. What are the field of view and smallest relevant feature or deviation?
3. What surface, color and position variation apply? Can you provide images or samples?
4. What is the production speed, and how many views/images are needed per part?
5. What space and working distance are available, and what is the environment?
6. Which interface, controller and software must the system work with?
7. Who will integrate it, and what quantities, schedule and budget are planned?

These answers establish both camera requirements and the boundary of the complete system. When replacing a camera in an existing machine, involving the machine builder and software owner is especially important; a similar designation does not prove compatibility.

Let us discuss your application

Virtech Sensor helps clarify the task and select a suitable Baumer camera or vision sensor. If a complete system is required, integration scope and responsibilities need to be agreed separately.

Send the task description, field of view, production speed, working distance, environmental conditions and existing software environment.

Discuss your application with the Virtech Sensor team

Related sources

• Baumer industrial cameras – official portfolio (EN)
• Baumer VeriSens vision sensors – brochure (EN, architectural background)

Examples explain system architecture rather than promise a particular product’s performance. Check selected-model functions and compatibility against current documentation, and validate the application with representative samples.