Timely failure detection for bearings is of great importance to prevent economic loses in the industry. In this article we propose a method based on Convolutional Neural Networks (CNN) to estimate the level of wear in bearings. First of all, an automatic labeling of the raw vibration data is performed to obtain different levels of bearing wear, by means of the Root Mean Square features along with the Shannon's entropy to extract features from the raw data, which is then grouped in seven different classes using the K-means algorithm to obtain the labels. Then, the raw vibration data is converted into small square images, each sample of the data representing one pixel of the image. Following this, we propose a CNN model based on the AlexNet architecture to classify the wear level and diagnose the rotatory system. To train the network and validate our proposal, we use a dataset from the center of Intelligent Maintenance Systems (IMS), and extensively compare it with other methods reported in the literature. The effectiveness of the proposed strategy proved to be excellent, outperforming other approaches in the state-of-the-art.