Global Sources
EE Times-Asia
Stay in touch with EE Times Asia
EE Times-Asia > Manufacturing/Packaging

Software automates image-recognition in assembly lines

Posted: 10 Sep 2014 ?? ?Print Version ?Bookmark and Share

Keywords:image recognition? assembly? Fujitsu?

Towards smarter machine-learning process

To make the machine-learning process more efficient, Fujitsu Laboratories devised three building blocks: the teacher, the grader, and the teaching material.

The teacher building block is a program that finds the desired image can be constructed by combining image-processing functions in a tree structure, but with conventional methods, the combinations would approach infinite size, making this technique impractical. Fujitsu imposed limits on the order of processes when forming the tree structure, based on the type of process and the flow, among others. This dramatically reduces the number of combinations so that the target program can be generated quickly.

Teacher building block

Figure 3: How image-recognition programs are created automatically using genetic algorithms. Source: Fujitsu

The grader building block evaluates the programs created in the automatic-generation process on the basis of the shape of the component and its similarity to the target, in order to detect position. For example, if a program detects a straight line, and if the line's angle () and position () match up well with the target, the program is considered to be good. This makes it possible to automatically generate programs that can be used for positional detection even if image quality is poor.

Grader building block

Figure 4: How generated images are evaluated. Source: Fujitsu

The teaching-material building block categorises multiple learning-candidate images based on feature values for factors such as brightness, contrast, and detail. Representative images are selected from each category to ensure that programs can handle the full variety of images based on a relatively small amount of training data.

Teaching-material building block

Figure 5: How training data is selected. Source: Fujitsu

Accuracy in automation

Trials to assess positional detection of components during assembly, showed dramatic improvement from below 50 per cent all the way to 97 per cent. The time required to revise image-recognition programs was also reduced to one-tenth the previous time. The high recognition rate halves positional deviations during component assembly, cutting overall assembly time to two-thirds.

The trial demonstrates that it will be possible to run a production line with stable recognition rates, without stoppages, resulting in high-quality, high-efficiency manufacturing.

The technology has potential applications beyond component assembly to machining, inspecting, and other industrial processes where image processing can be used on a production line. Fujitsu said it plans to use this technology to help make production lines better able to respond to changes in their operating environment without long downtime.

Fujitsu expects to implement the technology on its own production lines within the year. The company will also explore broader applications of this technology as a solution in vehicle onboard cameras, monitoring cameras, and cameras for medical purposes.

?First Page?Previous Page 1???2

Article Comments - Software automates image-recognition...
*? You can enter [0] more charecters.
*Verify code:


Visit Asia Webinars to learn about the latest in technology and get practical design tips.

Back to Top