Non-intrusive load monitoring (NILM) plays a crucial role in electricity management, and existing monitoring methods suffer from challenges related to the dependency on large historical datasets and powerful computational devices, making them unsuitable for deployment on edge devices. To address this issue, this paper proposes a load monitoring method based on a dual convolutional neural network (Dual-CNN) and employs K-means clustering for data labeling. The proposed method employs iterative K-means clustering to roughly categorize and label short-term electricity consumption data, thereby avoiding the uncertainty of the K value and the problem of imbalanced data classification granularity in traditional K-means algorithms. Subsequently, a dual convolutional neural network model is constructed for load identification and disaggregation, and an integrated edge device is designed to collect data and perform load monitoring. Experimental results demonstrate that the proposed method successfully achieves load monitoring with the advantage of unsupervised learning. The method exhibits stable performance on the designed edge device, and its overall performance in non-complex electricity systems is above 0.9.