Data processing is an essential component of heavy-metal ion detection. Most of the research now uses a conventional data-processing approach, which is inefficient and time-consuming. The development of an efficient and accurate automatic measurement method for heavy-metal ions has practical implications. This paper proposes a CNN-based heavy-metal ion detection system, which can automatically, accurately, and efficiently detect the type and concentration of heavy-metal ions. First, we used square-wave voltammetry to collect data from heavy-metal ion solutions. For this purpose, a portable electrochemical constant potential instrument was designed for data acquisition. Next, a dataset of 1200 samples was created after data preprocessing and data expansion. Finally, we designed a CNN-based detection network, called HMID-NET. HMID-NET consists of a backbone and two branch networks that simultaneously detect the type and concentration of the ions in the solution. The results of the assay on 12 sets of solutions with different ionic species and concentrations showed that the proposed HMID-NET algorithm ultimately obtained a classification accuracy of 99.99% and a mean relative error of 8.85% in terms of the concentration.

Traditional unsupervised domain adaptation (UDA) usually assumes that the source domain has labels and the target domain has no labels. In a real environment, labelled source domain data usually comes from multiple different distributions. To handle this problem, multi‐source unsupervised domain adaptation (MUDA) is proposed. Multi‐source unsupervised domain adaptation aims to adapt the model trained on multi‐labelled source domains to the unlabelled target domain. In this paper, a novel MUDA method by domain‐specific feature recalibration and alignment (FRA) is proposed. Specifically, to achieve feature recalibration, the authors leverage channel attention to pick out significant channels and spatial attention to focus on important features in different channels. Such integration of channel and spatial attention can lead to effective domain‐specific feature recalibration that may be of great importance to MUDA. In addition, to achieve better MUDA, the authors propose domain‐specific feature alignment which consists of Maximum Mean Discrepancy and JS‐divergence loss. Maximum Mean Discrepancy can reduce the difference between the source domain and target domain. Meanwhile, JS‐divergence loss may ensure the prediction consistency of different classifiers in the source domains. Four experiments have proved that FRA can achieve significantly better results in popular benchmarks for MUDA.