Modern civilization largely depends on electricity to run everything from businesses and houses to technology and medical facilities. It is vital for supporting infrastructure, promoting economic growth, and improving quality of life. The detection of anomalous power consumption is similarly crucial since it aids in identifying unusual patterns in energy usage, reduces waste, improves efficiency, and protects against possible electrical problems or security breaches, all of which eventually contribute to sustainability and cost savings. This work proposes a convolutional neural networks (CNN)-based anomalous power consumption detection using quick response (QR) code images. SimDataset, one of the widely used labeled datasets in anomalous power consumption detection, has been used in this work. Each instance of the dataset is converted to the QR code image, and the generated QR code image dataset is used to train and test the proposed approach. The outcomes of the experiments showed that the proposed approach is superior to the state-of-the-art in standard performance evaluation metrics, resulting in an overall accuracy and F1-score of 96.7% and 0.94, respectively.