Buried cross-country carbon-steel pipelines are used for transportation of petroleum products over large distances. They need to be regularly inspected to ensure its structural integrity. This is a mandatory requirement as per Oil Industry Safety Directorate (OISD) guidelines. Instrumented Pipeline Inspection Gauge (IPIG), an inline inspection tool is used for non-destructive evaluation (NDE) of these pipelines. It travels inside the pipeline with the flow of petroleum products, carrying an array of NDE sensors for automated inspection of these pipelines. Ultrasonic sensing, among others, is one such technique, used for assessing the health of these pipelines. Due to the rough surface of these pipelines, along with motion-induced vibrations of the tool, ultrasonic echo received by the sensor is often buried in heavy noise. Uncertainty in the region of occurrence of corrosion defects (internal or external to pipe surface) implies that echoes received from both near and far surfaces of pipe wall need to be independently identified and analyzed. In this paper, an algorithm to independently identify two different echoes from a single ultrasonic A-scan and estimating remaining pipe-wall thickness for carbon-steel pipelines is proposed. An array of such UT transducers kept in a customized configuration leads to estimation of length, width and depth of corrosion defects. The algorithm uses concepts of Stationary Wavelet Transform (SWT), envelope extraction by Hilbert transform and remaining pipe-wall thickness measurement in the spectral domain. The technique is validated experimentally on carbon steel plates with template defects fabricated and has achieved an accuracy within 3.50% of pipe-wall thickness.