高锌背景下光谱法同时检测痕量多金属离子浓度时, 由于微型光谱仪光源能量辐射不均匀性, 并且混合溶液中不同离子对不同波段紫外可见光会选择性吸收, 因而如果选取微型光谱仪的积分时间过大, 可能导致光谱能量值达到饱和, 选取积分时间过小可能导致光谱信号的信噪比很低。 积分时间的选择往往取决于研究者的经验和待测离子对紫外可见光的吸收特征。 为了实现能够自动选取微型光谱仪积分时间参数, 提出了一种基于二分搜索的高质量紫外可见光谱信号重构算法, 用于重构由不同积分时间组成的图谱特征更加明显的紫外可见光谱信号。 该方法首先采集不同积分时间下参比溶液的紫外可见光谱能量信号, 然后给定参比溶液的不同目标重构光谱能量信号值, 在每一波长点使用二分搜索算法寻找合适的积分时间采样参数; 然后根据紫外可见光谱的特点, 定义了表示重构后的光谱能量值与目标设定值接近程度的重构精度指标和表示重构信号后与重构信号前的图谱特征区分程度的重构特征显著度指标, 最后, 选取搜索区间范围内重构精度最高的光谱信号作为重构信息量, 利用光谱信号重构信息量重构待测溶液紫外可见光谱能量值, 最终得到待测溶液的重构光谱吸光度信号。 实验结果表明, 该算法能够快速自动地选定目标积分时间采样参数值对紫外可见光谱进行信号重构, 来得到高质量紫外可见光谱信号。 该算法可以使信号重构精度达94.84%, 并且重构特征显著度有所提升。 同时, 相对于重构前的光谱信号, 重构后的光谱吸光度信号得到一定程度增强, 信号信噪比也大大提升, 而且避免了积分时间采样参数需要依靠研究者主观判断选择的问题, 为检测多种痕量金属离子的浓度信息提供了高质量的模型数据。

In the background of high zinc, when the concentration of trace polymetallic ions is simultaneously measured by spectroscopy, the energy radiation of the light source of the micro-spectrometer is not uneven, and different ions in the mixed solution can selectively absorb the ultraviolet-visible light at different bands. Therefore, too long integration time may lead to the saturation of spectral energy value, and if the integration time is insufficient, the signal-to-noise ratio of the spectral signal may be very low. The selection of integral time depends on the existing experience of researchers and the characteristics of the UV-Vis spectral signals of ions to be measured. In order to automatically select the integral time of micro-spectrometer, a high-quality UV-Vis spectral signal reconstruction algorithm based on the binary search is proposed to reconstruct the UV-Vis spectral signal with more distinct characteristics composed of different integral time. Firstly, the UV-Vis spectral energy reference solution is collected at the different integral time. Then the spectral energy signal target values of the reference solution are given, and the appropriate integral time sampling parameters are found at each wavelength point by using the binary search algorithm. Next, according to the characteristics of UV-Vis spectrum, the reconstruction accuracy index which is determined to represent the degree of proximity between the reconstructed spectral energy value and the target setting value and the feature saliency index which is used to indicate the degree of feature saliency before and after signal reconstruction are defined. Finally, the spectral signal with the highest reconstructed accuracy in the search interval is selected as the reconstructed information. The reconstructed information is used to reconstruct the UV-Vis spectral energy value of the solution to be measured. The reconstructed spectral absorbance signal of the solution to be measured is obtained at the end. The experimental results show that the proposed algorithm can quickly and automatically select the lintegral target time to reconstruct the UV-Vis spectrum signal, and to obtain high quality UV-Vis spectrum signal. The accuracy of signal reconstruction can reach 94.84%, and the reconstructed feature is improved significantly. At the same time, the absorbance signal of the reconstructed UV-Vis spectral signal is enhanced. Compared with the UV-Vis spectral signal before reconstruction the signal-to-noise ratio of the spectral signal after reconstruction is greatly improved, and the problem of choosing the parameters of integral time based on the subjective judgement of researchers is avoided. It provides high quality model data for detecting the concentration information of various trace metal ions.