What is Data Fusion in a Machine Vision context?

I am writing a paper on Data Fusion (Main Topic) in the context of Machine Vision. If someone can define or lead me to a tutorial explaining it head to IEEE transactions to get started reseraching on the topic.

• First of all, what is Data Fusion in Machine Vision Vision Context? Can you give examples

Note: I will place the sub-topics here after someone answers.

[Walter Roberson]

Example: suppose your mobile robot has a lidar and a camera, then you could combine the distance information from the lidar with the color information of the camera to help make better sense of your environment.

[Leonardo Wayne]

Hi, Walter Roberson

Thanks for your reply. A follow up to your reply is here in these questions:

1. So in a Machine Vision system, all data coming from the sensors are all visual which are basically collecting, and processing information using techniques manipulating the electromagnetic spectrum. Is this correct? Does this include light, radio, and microwave, do they all fall under Machine Vision?
2. Also, is it correct to categorise Data Fusion in Machine Vision applications/context as "Sensor Fusion"? And if so, can this be data coming from a combination involving non-visual sensors, say hypothetically; an Ultrasonic sensor data (uses infomration about sound) coupled with LIDAR, or two sensors data that don't even use the EM spectrum?

Please feel free to highlight any misconceptions stated above.

[Walter Roberson]

So in a Machine Vision system, all data coming from the sensors are all visual

No, a machine vision system can use whatever information can be deduced from its sensors. That could include audio in the form of sonar, and could include human audio range to process dopler effect, and could include human audio range to process air movement to detect which areas are likely to be open or to target certain sounds (e.g., you might be designing something that heads towards the source of pedestrian signals for the blind.)

The general topic of Machine Vision would certainly include simulating a mosquito which does not have good vision but homes in on CO2 emissions from mammals.

There is perhaps some technicality that you cannot really call it "Machine Vision" unless at least one EM sensor is involved.

But none of this has anything to do with MATLAB and you should be doing your research elsewhere.

[Leonardo Wayne]

Thanks, Walter Roberson