Image analytics, the means by which computers can visually identify and classify real-world objects, is seeing more widespread use given recent advances in computer vision and machine learning technology. Accurate analytics can now be achieved in real-time, and powerful (often cloud-based) computing has made the development and operationalization of these technologies much faster.
In 2012 a group of researchers from the University of Toronto won the ImageNet Computer Vision competition using a deep convolutional neural network (CNN). The objective was to classify millions of images from thousands of categories. They not only won the competition by a great margin, but also cast a spotlight on the use of this technology for image processing.
CNNs have really taken the world by storm and their use for advanced analytics on images and video has increased dramatically with applications varying from facial recognition, self-driving cars, and analyses of medical images to help doctors make informed diagnoses.
A collaborative report by DHL and IBM entitled “Artificial Intelligence in Logistics” was published in early 2018, and highlights how Artificial Intelligence (AI) is playing an increasingly central role in the digital transformation of the logistics industry. The report goes further to say that there has probably never been a more exciting time for collaboration between logistics and technology professionals as they enable AI in this industry.
There are many use cases for AI in supply chain automation. Often, AI technologies that have been successful in other industries find direct application in supply chains. These include intelligent robotic process automation or machine learning for improved forecasting. Image analytics, however, has opened a door to an exciting field of application for improvement of how supply chains operate.
One very important task, that often goes astray, is management of everything in your warehouse. It is necessary to continuously keep track of what has arrived at the warehouse, the waybill needs to correctly describe what has actually arrived, and the condition of everything that arrived has to be verified. Image analytics eases the burden of checking in wares and verifying the condition, which can then be validated against the waybill information.
Image analytics can further be used to keep track of goods as they move around the warehouse. Cameras mounted on forklifts, for example, can keep track of the bins into which certain goods are loaded, how full the bins are, and where empty bins are located for future loading. Simple data analysis of picking actions can help plan which bins to use for which goods. Frequently accessed items can be placed in easily accessible bins, and items that commonly ship together can be placed close to each other. Warehouse efficiency can also be improved through loitering detection from video feeds.
Identifying bottlenecks in the supply chain is a critical aspect of improving overall efficiency. Machine learning, AI, and image analytics can be leveraged right now for major improvements to warehouse operations. These tools will play an important role in the digitalisation of supply chains, a trend which is accelerating globally and which companies should follow to avoid being left behind.
Onpro Consulting provides data science solutions that include machine learning, AI, and image analytics. Combining effective business consulting with state-of-the-art data science solutions is one of Onpro Consulting’s strengths.