Fujitsu Laboratories’ new technology consists of three building blocks: The teacher, grader, and the teaching material, which are described by Fujitsu as follows:
- The teacher building block: A program that finds the desired image can be created by combining image processing functions in a tree structure, but with conventional methods, the combinations would approach infinite sizes. The new technique imposes limits on the order of processes when forming the tree structure, based on knowledge of feature emphasis, which reduces the number of combinations so that the target program can be quickly generated.
- The grader building block: Rather than evaluating image quality, the new software evaluates programs created in the automatic generation process on the basis of the shape of the component and its similarity to a target, in order to detect position.
- The teaching material building block: Tp reduce learning time by minimizing the number of pairs of training images and targets, the new software categorizes multiple learning-candidate images based on feature values for factors such as brightness, contrast, and detail. Representative images were selected from each category to ensure that programs would be generated that could handle the full variety of images based on a small amount of training data.
In trials conducted by Fujitsu Laboratories to assess the positional detection of components during assembly, a recognition rate of 97% was achieved. In addition, the amount of time it took to revise image recognition programs was reduced to one-tenth of the previous amount of time. Fujitsu Laboratories is currently refining the software with the goal of implementing it on Fujitsu's own production lines during this fiscal year. In addition, the company plans to investigate broader applications of this technology as a solution in vehicle onboard cameras, monitoring cameras, and medical imaging cameras.
Read additional details about the new software.
Share your vision-related news by contacting James Carroll, Senior Web Editor, Vision Systems Design
To receive news like this in your inbox, click here.
Page 1 | Page 2