Optimization of Plant as Required by the Shop Floor! 3D Picking with Image Processing and Robot ControlI-CUBE TECHNOLOGY Co.,Ltd
Construction of a high-quality system with high added value at low cost by combining image processing with robot control.
Points to be noted
- Rational promotion of development with industry-academia collaboration
- Construct a reasonable system with the current system and low-cost robots
3D picking is required for automation systems of the production process in factories. ICube developed an automation system by combining robot control with image processing as basic technology. 3D images are prepared from 2 cameras, and objects are recognized using the images. Camera cost is reduced by using existing systems. Research is in progress on 3D object recognition in collaboration with professor Hashimoto of the Department of Mechanical and Systems Engineering, School of Engineering, Chukyo University.
The main product of the present factory automation (FA) industry is flat stack picking. Parts are placed without overlap on flat areas such as belt conveyors, and they are picked up. Since there are many new types and small quantities of parts that must be picked, dedicated part feeders are required, labor and time are needed to change the feeders and requires storage space. One of the solutions for this complicated task is 3D picking.
The 3D picking system of ICube Technology, for example, can pick various parts kept in a box by searching spatial coordinates with a stereoscopic vision like human eyes. Dedicated part feeders are required each different part, and one of the drawbacks is that part feeder size increases with an increase in part size. 3D picking system can be used by merely replacing part information. Replacing the part feeder or changing the work setup is not required.
Mr. Hidetomo Narita, President, and CEO says “Humans can easily distinguish shape, color, and orientation. Though it is difficult for robots to do such work, the final goal of factory automation is to reduce workers. The goal of our development is to meet this need”.
Professor Hashimoto is responsible for developing the algorithm. 3D object recognition technology “Vector pair matching” developed by us independently uses only a few special features that are effective for recognition. There are three major advantages.
First, high speed. Calculations extract special features with independent judgment reducing the data volume. This increases the picking speed.
Next, highly accurate and reliable recognition. Since only valid features are extracted in advance, the risk of incorrect recognition is low, and the reliability of recognition is improved.
Finally, compact and low cost. Since calculations are reduced, a compact and low-cost system is realized.
Due to aging and declining birthrate in Japan, there is a continuous strict requirement for improving production efficiency and quality control, even though there is a scramble for securing workers in the manufacturing field. Though there is a growing need for automation, low-cost systems that can be purchased by small and medium sized companies are not available.
ICube Technology collaborates with Professor Hashimoto, who is an authority on robotics research for factory automation, to offer proper solutions for user requirements. Our objective is to develop a system that can manage and operate at a low cost, even if it is not specialized. We will continue to contribute to automation by combining image processing and robots and are also looking to enter the medical field.
The situation of the algorithm which will become the mainstay in the field of 3D picking is still fluid, and companies are engaged in fierce competition. We would like to focus on how the system of ICube Technology can change the shop floors of Japan.
Interview Date: July 5, 2018
Our Company provides an integrator for automation systems, which is help in solving current issues in manufacturing, such as vanishing workforce and improvement in production efficiency. The automation system proposed by our company has the advantage of being able to be customized according to customer specifications through the application of image processing technique + robot + application software.
DESIGNOFFICE SEE Co.,Ltd
A non-invasive “Body glucose balance meter” to easily measure blood sugar levels and “Optical agrisensor” that can quantify the state of plant growth
Kiryu Electronics Development. Corp
"Caster & Central lock solution system" to enable free design with wire.
TOSHIN MFG Co.,Ltd