The illicit importation, development and deployment of nuclear or radiological dirty bombs poses a serious threat to the United States. Portal monitoring and inspection systems provide a major domestic layer of defense. Once radiological materials are inside the country, the detection, localization and identification of these threats becomes increasingly more difficult due to complex radiological backgrounds, which can easily confound the detection of shielded or low activity threat signatures. Technology development efforts seek to significantly enhance the capability of detection systems in low signal-to-clutter ratio (SCR) environments. Physical Sciences Inc. (PSI) has developed a novel approach for radiological background estimation that improves the detection and discrimination capability of on-the-move medium resolution detectors, as well as handheld isotope identifiers used in small area searches. The algorithm processes energy spectra to perform clutter suppression and applies statistical hypothesis testing resulting in the detection and identification of low-activity threats in noisy environments. The performance is achievable at low integration times necessary in a high throughput and dynamic operational environment. We present a detailed quantitative analysis of algorithm performance against field data collected in a variety of environments with embedded check sources. We show that the algorithm achieves a high probability of detection and identification with low false warning rates under low SCR conditions.