In addition to Weibo, there is also WeChat
Please pay attention
WeChat public account
Shulou
2025-01-19 Update From: SLTechnology News&Howtos shulou NAV: SLTechnology News&Howtos > IT Information >
Share
Shulou(Shulou.com)11/24 Report--
Beijing, December 11 (Xinhua) Robots may be the future, but robotic arms are obviously not good at using an ancient and stable form of technology: bar codes. E-commerce giant Amazon said on Friday that bar codes are hard to find and may be posted on grotesque products, a problem that robots can't solve well. So Amazon is going to eliminate the technology.
Amazon wants to kill the bar code with robots. Amazon's solution is to train computer models with pictures of items in Amazon's warehouse. The company has developed a camera system that monitors items passing through the conveyor belt one by one to make sure they match the pictures. Eventually, Amazon's artificial intelligence and robotics experts hope to combine the technology with robots, allowing robots to identify objects while picking and flipping them.
Amazon's artificial intelligence experts must first build a product image library, and there is no reason for the company to do so before the project. The image itself and the data about the size of the product provide material for the early version of the computer vision algorithm, and the camera continues to capture new images of the product to train the model.
"it is essential to solve this problem so that robots can pick up objects and process them without having to find and scan bar codes," said Nontas Antonakos, applied science manager of Amazon's computer vision division in Berlin. "this will help us deliver packages to customers faster and more accurately."
The system, known as multi-pattern recognition, will not soon completely replace bar codes. Amazon said the system has been put into use in warehouses in Barcelona, Spain and Hamburg, Germany, speeding up the processing of parcels there. The technology will be shared across Amazon's business units, so it could be used in whole Foods or other Amazon physical chains in the future.
Amazon said that the error rate of the system is not high, and the problem of sending incorrect goods to customers does not occur frequently. But considering that a warehouse has to deal with many items a day, even unusual errors can lead to serious delays.
Amazon's algorithm has an accuracy of between 75% and 80% when it is used for the first time, which Amazon sees as a good start. Amazon says the current accuracy has reached 99%. The system encountered a failure when it was initially used and was unable to capture the color difference. During the Prime Day promotion, the system cannot distinguish between two different colors of Echo Dots smart speakers. The only difference between these packages is a small blue or gray dot. With some improvements, the recognition system can now assign a confidence score to its level (whether it can correctly understand the intention), which marks only items that it is very certain to be incorrect.
The Amazon artificial intelligence team says fine-tuning multiple pattern recognition systems to evaluate products processed by people will be a challenge, which is why the ultimate goal is for robots to handle these products.
Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.
Views: 249
*The comments in the above article only represent the author's personal views and do not represent the views and positions of this website. If you have more insights, please feel free to contribute and share.
Continue with the installation of the previous hadoop.First, install zookooper1. Decompress zookoope
"Every 5-10 years, there's a rare product, a really special, very unusual product that's the most un
© 2024 shulou.com SLNews company. All rights reserved.