Developing Warehouse Robots to Improve Inventory Cycle Count

#Machine Learning
August 18, 2017 3 min read

Imagine a warehouse, just like the one pictured below, packed with thousands of boxes. Now imagine a vast network of warehouses with a myriad of new packages arriving daily.  

Warehouse workers walk from one aisle to another, manually count the number of boxes on each shelf, record the data, and verify if this number corresponds to the quantity indicated in the database. 

Inventory cycle count optimization picture

At Innovecs, we have been working on a state-of-the-art AI-powered robot which can cope with this task. It is capable of saving a lot of time and budget. Even though it may sound like something out of a sci-fi movie, Innovecs is making it a reality. We already have a prototype developed and the robot itself will be fully operational in 6 months.

This is a very time consuming and costly process. Workers spend hours walking around the premises to keep track of all the inventory. Furthermore, human involvement is error-prone. It is easier to make a mistake while counting and record wrong information. This can wreak havoc in the warehouse management and result in running out of stock when it is required and delays in shipping. Automated warehouse seems to be the right way out. Using automated warehouse robots can significantly optimize the entire inventory management process. 

The Warehouse Robot Basic Capabilities 

The project goal is to develop the warehouse automation software enabling the robot to calculate boxes within an image using the OpenCV and Tesseract technologies. These boxes will be stored in a big warehouse with labelled rows such as in the picture below.

inventory cycle

The robot must be able to: 

  • Read the barcode and letters on the warehouse shelf, which contains information about the quantity of boxes located on it 
  • Identify an object in the image (pallet, box, label, etc.) 
  • Count the number of boxes in the image 
  • Verify if the number of boxes in the image matches the number stored in the database 

Challenges Innovecs Needs to Overcome 

Developing the warehouse automation robot which can accomplish all the above stated tasks poses difficulties, yet interesting challenges to overcome. They run as follows:

  • Detecting, identifying and quantifying objects in an image
  • Synchronizing information received with the warehouse database
  • Reading and recognizing the barcode on a warehouse shelf
  • The boxes are stored in the warehouse in subzero temperatures (-30°C), therefore the camera must be made operable in these conditions
  • The warehouse is equipped with low lighting, which is why the camera must also be able to detect the objects in the image and count them in the dimly lit premises

    The Warehouse Object Detection and Analysis 

    To make the robot cope with the inventory count, it should be able to do two things: detect an object and analyze it.

    For the object detection, the warehouse robot defines whether the object is in the frame or not and figures out its approximate location. We use OpenCV for that, since its features are better suited. For the object analysis, we use Tesseract, since OpenCV does not include OCR (Optical Character Recognition) libraries. OCR is a popular technology used for text recognition inside the image. There are two general settings where OCR is applied:

    Controlled: Images taken from a scanner where the target is a document with clear fonts, perspective, orientation, and background consistency.

    Uncontrolled: Live photos taken from a camera where OCR helps identify traffic lights, a stop sign, number plate, etc.

    Tesseract is better when used in the controlled setting, such as a warehouse, that is why we decided to make use of it. Generally, retraining OCR will not directly improve detection, but may improve recognition, especially for the scene text OCR.

    The Final Product: AI-Driven Automated Warehouse Robot 

    The final product is going to be a warehouse robot placed on the forklift (like the one shown below) and powered by artificial intelligence. Thanks to an advanced computer vision, it will be able to accurately identify, locate and record in-store package data even if boxes are placed just above the ceiling.

    AI-Driven Warehouse Robot

    Equipped with such inventory count robots, the warehouse managers can set inventory sweeps at schedules they need and increase their frequency.  It will help receive updated stock data every one or two hours and significantly improve storage and placement as well as minimize ineffective movements of the warehouse workers and machines used. Furthermore, such robots are very compact allowing more space for the inventory storage.

    It is only one out of many examples of the innovative software solutions Innovecs can develop from scratch. We are always open to new, exciting and challenging projects which require the application of the emerging technologies including machine learning, artificial intelligence, the Internet of Things, and more. Get in touch with our team and learn what type of software we can create for your business to prosper.

    4.91 5.0
    16 Reviews
    Dmytro Ryabokon
    Dmytro has PhD in Engineering in the field of Artificial Intelligence. His professional experience includes over 10 years in teaching and scientific activities (Kyiv Polytechnic Institute, Institute for Applied Systems Analysis; Taras Shevchenko National University of Kyiv, Faculty of Cybernetics). Additionally, he has more than a decade of project management experience in product R&D companies (Finance, Logistics) as well as big outsourcing programs (Deutsche Bank, LG Electronics).