<p>This paper presents a constant false alarm rate (CFAR) detector for the use in dense multi-target environments. The proposed CFAR detector is based on the Ordered Data Variability (ODV) statistic, allowing for adaptive truncation of interference units. The method generates a Range-Doppler map using a two-dimensional fast Fourier transform and reduces both the interference from nearby targets and the high noise by configuring protection and training units. ODV statistics are then used to estimate the energy of the training units, adaptively removing the excessive energy. Finally, the noise in the remaining training units is estimated. A target detection model is developed through simulation to evaluate the algorithm's performance in dense multi-target environments. Additional simulation experiments are conducted to assess its detection performance in dense multi-target environments under the presence of multi-source noise of radar system. Experimental results demonstrate the algorithm's detection capability in dense multi-target environments, and it also exhibits robust detection performance even in the presence of multi-source noise in radar system.</p>