Seismic noise attenuation is a critical step to suppress the unwanted noise that masks the true geological structures. To handle the complexity and high amplitude of the coherent noise, a typical denoise workflow often requires multiple passes of denoise algorithms to target each noise type separately. We propose a novel one-step method to handle diverse high-amplitude noise. The method consists of automatic detection of coherent noise based on a global signal trace within a gather plus a low-rank eigenimage filter for suppressing the noise on identified noisy traces only. This one-step approach avoids the conventional workflow that uses multiple passes of various denoise algorithms. In addition, this method is effective to handle a variety of coherent noise and also reduces the data processing turn-around time. We apply this method to a 3D field data set from the Llanos Basin, Colombia, which has a variety of strong coherent noise. The proposed method reduces the strong coherent noise considerably without any observable loss of primary energy.