摘要
在基于FPGA的时频分析研究设计中,针对直接显示的时频谱图无法体现频谱细节变化的问题,提出了一种适合硬件实现的谱图增强算法.该算法以正弦信号的频谱结构为原型,选用反比例函数逼近正弦信号的频谱包络,并结合实际应用情况推导出具体的增强算法,利用该算法不仅对时频谱图的灰度级进行了压缩,而且进行了谱图增强处理.实验结果表明,此算法可有效显示出时频谱图的细节变化,与对数变换相比,具有运算量小,实时性好,适合硬件实现等特点.
During the research and design of time‐frequency analysis based on FPGA , the direct indication of the time‐frequency spectrogram is not able to reflect the variation of spectrum in detail . To solve this problem , an image enhancement algorithm of time‐frequency spectrogram based on FPGA is presented in this paper . The algorithm takes sinusoidal signal spectrum structure as design prototype , chooses the inverse proportion function to approximate the spectral envelope of sinusoidal signal , combines with the practical application to derive the specific enhancement algorithms . T he algorithm can not only compress the gray level of time‐frequency spectrogram , but also enhance the image . Experimental results show that more detail variation of time‐frequency spectrogram can be obtained by the algorithm effectively . Algorithm in this paper has the characteristics of much less computation , strong real‐time property and suitable for hardware implementation compared with logarithmic transform .
出处
《西北师范大学学报（自然科学版）》
CAS
北大核心
2016年第3期57-60,共4页
Journal of Northwest Normal University(Natural Science)
基金
西北师范大学青年教师科研能力提升计划资助项目(NWNU-LKQN-14-21)
关键词
现场可编程门阵列
时频谱图
短时傅里叶变换
图像增强
反比例函数
field programmable gate array
time-frequency spectrogram
short-time fourier transform
image enhancement
inverse proportional function
