Did you know that the manufacturing industry is one of the main sources of employment in our “hypertechnology” society? Yes, we refer to the one that is dedicated exclusively to the transformation of the raw material into final consumer goods. This is another area of the human economy that is undergoing a real revolution, thanks to the automation of tasks based on the latest advances in the [campo de la inteligencia artificial](https://data-speaks.luca-d3.com/2018/02/diferencia-IA-ICognitiva.html) and Deep Learning.
One of the key components of this automation is the ”computer vision”, which consists of identifying objects from images. It is literally about getting a computer to” see ” how a human does it, through the acquisition, processing and analysis of digital images and videos.
It is not difficult to find opensource tools that allow the detection of generic objects such as trees. These are very basic systems, which are already pre-trained. However, the object identification using artificial vision requires the training of algorithms capable of identifying much more specific details of an image, such as distinguishing different facial expressions.
We are going to see examples of how technologies based on artificial vision are beginning to be applied to different phases of the production process, from quality management, to the classification and packaging of products.
Quality control
The quality control process is highly dependent on a person’s visual and adaptive ability, so mistakes can be more common than they should be. AI can identify defective products automatically and at high speed. And what’s better, allows for corrective action with the same precision and agility. This is especially useful in very dynamic environments, where things are constantly changing.
Inventory management
Real-time inventory management can prove to be an incredibly complex task for an organization. The use of AI allows, again, automate this task, eliminating the risk of human error. This makes inventory maintenance more accurate and efficient.
Rating
Manual sorting is a long and costly process, in which human errors are also frequent. AI-based technologies make it possible to product tracking and classification, selecting certain very specific parameters, and generating corresponding statistics of the number of objects displayed. As a result, the assembly lines are much more flexible, and the number of sorting anomalies is greatly reduced.
Assembly line
In the manufacturing industry, almost all assembly lines are fully automated. While the use of robotics in this field is extremely useful, the use of technologies based on artificial intelligence, to locate, correctly identify products, and locate them where they correspond at all times, will open doors to production improvements and the efficiency of the workers. And it is precisely machine vision, that is, the target detection based on AI techniques, which allows this possibility to become a reality.
Detection of”custom objects”
Technologies based on the detection of “custom objects ” allow to meet the specific needs of manufacturing highly specialized or niche products. Objects can have a variety of shapes, and algorithms usually need to be trained with thousands of examples to learn how to differentiate them. With this technology, programmers can greatly reduce those needed for the algorithm to work accurately and efficiently.
In short, advances in machine vision are a clear benefit for the industry, since automation immediately translates into improved productivity and error elimination associated with the human factor, allowing factory workers to dedicate their time to tasks with greater added value.
Original post written by Lucy Beardsley for LUCA’s blog. Translated by Paloma Recuero.
