What is computer vision used for?#
It automates the analysis of visual information, allowing a business to instantly identify manufacturing defects, monitor a physical space for safety compliance, or categorise images without direct human intervention. This dramatically increases the speed, accuracy, and scale of any process that relies on visual inspection.
How does computer vision work?#
A computer vision system is "trained" by being shown thousands or millions of labelled images. For example, to learn to identify a "cracked part," it is fed countless pictures of both perfect and cracked parts, with each image clearly marked. The system learns to recognise the specific patterns, shapes, and textures associated with each category.
Once trained, the system can analyse a new, unseen image and predict what it contains based on the patterns it has learned. It essentially breaks the new image down into its fundamental components and compares them against its training to make a determination—such as "cracked" or "not cracked"—with a calculated degree of confidence.
How is computer vision different from image processing?#
Computer vision is focused on extracting meaning and understanding from an image or video. This is different from basic image processing, which focuses only on enhancing or manipulating an image (like adjusting brightness or applying a filter) without necessarily understanding what the image contains.