Paper
Recent years have assisted a widespreading of RF-based tracking and mapping algorithms for a wide range of applications, ranging from environment surveillance to human-computer interface. This work presents a material identification system based on a portable 3D imaging radar-based system, the Walabot sensor by Vayyar Technologies; the acquired three-dimensional radiance map of the analyzed object is processed by a Convolutional Neural Network in order to identify which material the object is made of. Experimental results show that processing the three-dimensional radiance volume proves to be more efficient thas processing the raw signals from antennas. Moreover, the proposed solution presents a higher accuracy with respect to some previous state-of-the-art solutions.
|
Example of the data recorderd by the Wallabot device.
|
Dataset
We collected four different datasets (D1, D2, D3, D4) with different materials, which are reported in Table 1. Dataset D1 consists in 4 material classes, namely wall, glass, wood1, wood2,1 which are made of about 100 acquisitions of the 40 raw signals of the antennas and of the scene reflectance tensor. The acquisitions D1 were used to evaluate the performances using the first or the second type of data. These classes were extended in dataset D2, including two different desks, different types of bricks and mortars, plastic material, and composed materials (floor). The final datasets consists of 12 different material types, which were acquired on the same day. Dataset D3 and D4 consists in 5 of the materials contained in D2 (wall, floor, wood1, wood2 and glass) but captured on different days and instances. This can be used to test the reproducibility of the classification methods and permitted evaluating the robustness of the approach.
Downloads
- Wallabot dataset for material classification
- Paper
At the address http://lttm.dei.unipd.it/nuovo/datasets.html you can find other ToF and stereo datasets from our research group.
Contacts
For any information on the data you can write to gianluca.agresti@dei.unipd.it or simone.milani@dei.unipd.it. Have a look at our website http://lttm.dei.unipd.it for other works and datasets on this topic.