色阶
光学
高斯分布
共焦
色差
傅里叶变换
职位(财务)
流离失所(心理学)
波长
相(物质)
信号处理
观测误差
连贯性(哲学赌博策略)
信号(编程语言)
物理
亚像素渲染
准确度和精密度
频域
光谱密度
数学
材料科学
计量系统
相位相关
测量不确定度
高斯函数
计算机科学
算法
高斯过程
财产(哲学)
信噪比(成像)
解析信号
作者
Peidong Xu,Xiaokun Yan,Junteng Ma,Xiaomian Cai,Xiu Zheng,Youhua Chen,Cuifang Kuang
标识
DOI:10.1109/tim.2026.3671906
摘要
Peak extraction methods based on chromatic confocal displacement sensors have shown promising results; however, several challenges remain unresolved. Therefore, this paper proposes a method for determining the peak wavelength in chromatic confocal systems using time-frequency analysis. The peak coordinates are derived from the spectral characteristics through time-frequency analysis. The position is measured by applying the time-shifting property of the Fourier transform to identify the peak coordinates via the signal phase spectrum. The coherence property of dual-pulse functions in the complex domain is used for thickness measurement to determine the peak coordinates through the phase spectrum slope and magnitude spectrum zero-pole characteristics. The proposed approach demonstrates a lower systematic error and higher anti-aliasing capability than traditional localization methods. The proposed method maintains an accuracy comparable to Gaussian fitting approaches while reducing the indistinguishable distance between overlapping peaks by more than half.The position of each peak is obtained even when processing severely mixed double-peak signals. Moreover, The proposed method significantly improves the accuracy and measurement range of chromatic confocal systems for thickness measurement of transparent objects. The experimental results showed a significantly smaller asymmetry-induced error, achieving an 84% improvement in the localization accuracy over Gaussian fitting. The computational efficiency of the proposed method is 7.4 times higher than that of Gaussian fitting, thereby offering a new solution for peak wavelength determination in chromatic confocal systems.
科研通智能强力驱动
Strongly Powered by AbleSci AI