在水质总磷的在线检测过程中，由于微型光谱仪的测量精度限制以及一些外界干扰导致光谱数据的不稳定，影响了检测的稳定性和测量精度。在紫外分光光度法的基础上通过卡尔曼滤波和非线性回归等数据处理方法对光谱数据进行处理，建立吸光度与浓度的关系方程，计算得到总磷的含量。通过这种数据处理方法提高在线检测的稳定性和测量精度，很好地解决了一些随机干扰和系统误差的问题，适合总磷在线监测装置的现场实时运行，具有很好的推广作用。

In the online detection of total phosphorus for water quality, due to the limitation of measuring accuracy of micro-spectrometer and some external interference, the stability of spectral data is damaged so that the detection stability and precision is affected. Kalman filter and non-linear regression method are used to deal with spectral data on the basis of ultraviolet spectrophotometry, and then the relationship equation between absorbance and concentration is established. Thus the content of total phosphorus can be calculated. The method can improve the stability and precision of online detection. The problem of stochastic disturbance and system error can be solved very well. It can be used for real-time monitoring of total phosphorus.