A pipeline leak detection and localization technique is crucial in a structural health monitoring system to prevent water wastage at an early stage. The main aim of this approach is to propose a standalone architecture for leak detection and localization using a single sensor. The sensor used in this approach is an Acousto-optic vibration sensor, which is highly sensitive to capture the vibrations caused by the pipeline leak. The proposed standalone architecture contains two steps: 1) Feature extraction and 2) leak detection and localization. This approach uses a one-dimensional convolutional neural network (1DCNN) for feature extraction. This paper tunes the AdaBoost to have support vector machines (SVM), Decision Trees (DT), and multi-layer perceptron (MLP) instead of the inbuilt weak estimators to give improved performance. The modified AdaBoost detects and localizes the leak by classifying the leak locations. The proposed 1DCNN-modified AdaBoost's performance is cross-verified with nine models and cross-correlation. All the models are tested with 200000 and 300000 Pascal pressure to check the stability. The proposed 1DCNN-modified AdaBoost outperforms all the other methods implemented in this research. In the future, this research can be extended with different leak sizes and pipeline materials and real-time pipeline environments with longer distances. © 2013 IEEE.