With the advancement of Industry 4.0, the integration of manufacturing and intelligent technologies has become crucial for industrial upgrading. Tool wear in milling processes significantly impacts workpiece quality and accuracy. Therefore, online tool wear identification and real-time monitoring have emerged as vital research directions in intelligent manufacturing to enhance machining efficiency and reduce defects. This study proposes a tool wear condition monitoring model using Piecewise Cubic Hermite Interpolation Polynomial (PCHIP) for data augmentation, addressing the issue of insufficient tool wear data compared to sensor data. PCHIP establishes a one-to-one correspondence between sensor and tool wear data, preventing signal loss due to missing wear values and maximizing sensor data information utilization. The proposed CNN-Transformer model combines the rapid feature extraction and dimensionality reduction of Convolutional Neural Networks (CNN) with the long-term dependency learning of Transformer, enabling direct learning from sensor signal feature sequences without manual feature extraction. Experimental results show that the model achieves over 95% accuracy, outperforming the baseline model in all evaluation metrics, thus providing superior monitoring performance.