The number of end devices that use the last mile wireless connectivity is dramatically increasing with the rise of smart infrastructures and require reliable functioning to support smooth and efficient business processes. To efficiently manage such massive wireless networks, more advanced and accurate network monitoring and malfunction detection solutions are required. In this paper, we perform a first time analysis of image-based representation techniques for wireless anomaly detection using recurrence plots and Gramian angular fields and propose a new deep learning architecture enabling accurate anomaly detection. We examine the relative performance of the proposed model and show that the image transformation of time series improves the performance of anomaly detection by up to 29% for binary classification and by up to 27% for multiclass classification. At the same time, the best performing model based on recurrence plot transformation leads to up to 55% increase compared to the state of the art where classical machine learning techniques are used. We also provide insights for the decisions of the classifier using an instance based approach enabled by insights into guided back-propagation. Our results demonstrate the potential of transformation of time series signals to images to improve classification performance compared to classification on raw time series data.