- Recognizing pallets on images
- Counting the number of pallets based on the color and edges
- Storing information on the server system and integrated with the Android application
- Deep learning
- Python (Yolov 5 model)
- Team size: 2 people
- Development time: 1 year
The client is a logistics solutions company from Japan. To reduce the time checking and counting the number of pallets loaded on trucks, they wanted to create an Android application that uses AI to detect and count the number of pallets automatically.
The solution uses AI to recognize real-time images of pallets loaded onto trucks. The Android application counts pallets based on the color characteristics and the pallets’ edges.
The collected information will be stored on the server system and integrated with Android applications to manage and track goods on trucks.
The client provided us with samples of pallet images to analyze. To get high detection accuracy, we had to determine the optimal setting that balance all indicators (pallets' colors, length, edges, etc.). At the same time, our deep learning models need to analyze the images on a mobile application in real time.
The client initially just wanted a solution to streamline the process. But they have decided to go with a visual detection solution after a demo with our team. They realized this would be more time-efficient for their drivers and have higher accuracy than manual counting.
We analyzed the client's images and created more lookalike cases to train the model. And thanks to the abundant data, the deep learning model gradually gets higher accuracy.
The AI system's accuracy is currently at about 98.5%, with a scan time of one second per image.
Now, the application is in its trial phase with six phones to fulfill the client’s daily tasks. In the following production phase, we expect to have the application installed on 20 phones before launching it to thousands of truck drivers in the company.
Looking for an expert in AI?
1. GEM Corporation is an IT Outsourcing company experienced in developing AI-based applications. We have worked on developing AI projects for various Japanese corporations across industries.
3. We have more than 8 years of experience. Our offices are based in Hanoi, Vietnam, and Tokyo, Japan.
4. We have successfully built more than 100 successful projects for our clients in the US, UK, Europe, Japan, Korea, Singapore, and many more.
5. Let us know how we can help you build your banking solution. Contact us now and get a quote for your project.