This paper tackles the detection of radar targets in presence of K-distributed clutter. A detection scheme composed by a preprocessing stage and a likelihood ratio test is proposed for adapting the detection threshold in order to maintain a constant false alarm probability. The conventional Cell-Averaging Constant False Alarm Rate (CA-CFAR) detector is considered as reference one. CA-CFAR technique uses output of the square law detector which cannot be the optimum one. The proposed detection scheme allows the use of a fixed threshold based on the finite number of clutter samples of the reference window. Considered detectors are evaluated with real data acquired by IPIX radar. Results prove that the designed adaptive threshold technique based on the likelihood ratio presents better detection capabilities than CA-CFAR, maintaining for both schemes a similar constant false alarm probability.