See-Through Soil: Underground Root Tuber Sensing With RF Sensor Networks

This article proposes a data-driven root tuber sens ing (RTS) framework that uses the received signal strength (RSS) data from a radio frequency (RF) sensor network to reconstruct cross-sectional images of root tubers in soils. We perform extensive experiments with our data acquisition system in various environments to build a wireless potato sensing (WPS) dataset. We propose to integrate multibranch convolutional neural networks with a di usion neural network to enable fine grained image reconstruction of root tubers. To deal with the multipath e ects on radio channels, we propose two domain adaptation methods: one-shot fine-tuning to update the neural network model online and disentangled representation learning (DRL) to transfer a pretrained model to unseen environments. Experimental results from over 1.7 million RF network measure ments show the e cacy of the proposed methods across di erent environments.