Adaptive detection of range-spread target is addressed in compound-Gaussian clutter.Based on the secondary data

the adaptive modified scatterer density dependent (AMSDD) detector and the adaptive-dynamic-threshold (ADT) detector are designed

with the sample covariance matrix (SCM) or the recursive estimation (RE) matrix.The theoretical analyses of constant false alarm rate (CFAR) show that

the SCM-based detectors only hold the adaptiveness to the clutter covariance matrix structure (CCMS)

while the RE-based detectors can ensure CFAR to both of CCMS and texture components.The performance assessment shows that

as for AMSDD and ADT

the RE-based detectors outperform the SCM-based ones

in terms of CFAR properties and detection performance.Moreover

the ADT based on RE performs robustly in the mismatch cases of the expected number of scatterers.