Longª²Duration Energy Accumulated Detection
Based on Timeª²Frequency Morphological Filtering
Shang Haiyan 1£¬2£¬Shui Penglang1£¬Zhang Shouhong1£¬Su
Hongtao1£¬Zhang Yabin1
£¨1£®National Laboratory of Radar Signal Processing£¬Xidian¡¡University£¬Xi'an 710071£¬China£»2£®Electronic
Engineering Department£¬Xi'an Shiyou University£¬Xi'an 710065£¬China£©
Abstract£ºA new detection method based on timeª²frequency analysis and morphological
filtering is proposed in order to detect the maneuvering target with unknown waveform£¬
long accumulated time and low signalª²toª²noiseª²ratios £¨SNR£© under a severe noise
background£® Utilizing the characteristics that the timeª²frequency distribution
congregates the signal energy nearby the instant frequency curve and the white noise
energy is dispersed over the timeª²frequency plane£¬ the timeª²frequency distribution£¬
threshold processing and morphological filtering are used to estimate the support region
of high energy£¬ and then these regional energy is accumulated to construct the detecting
statistics to perform the statistic decision£® The proposed method does not need the
prior waveform information£¬ the accumulated time of detection is not restricted by
handling methods and the weak maneuvering target can be detected in longer observing time
under noisy background£® The simulation results show that when the false alarm rate is 10-5and
SNR is -9 dB the detection probability can achieve 99£¥£®
Keywords£ºtimeª²frequency distribution£» morphological filtering£» long accumulated
time£» maneuª²vering target detection