Monitoring pollutants in the atmosphere, by flame spectrophotometry, comprises a learning and diagnostic process for identifying a current spectrum

摘要

A learning and diagnostic process for identifying a current spectrum (CS) that allows detection, in the atmosphere, of chemical and/or biological substances (A) by flame spectrophotometry, is new. Learning and diagnostic process for identifying a current spectrum (CS) that allows detection, in the atmosphere, of chemical and/or biological substances (A) by flame spectrophotometry comprises: (a) principal component analysis of reduced data from CS; (b) producing a matrix that represents the set of projections of the set of actives; (c) classifying the set into current groups (CG); (d) evaluating the potential membership of CS for all CG; (e) assigning CS to a CG if this potential is above a predetermined threshold; (f) triggering an alarm if one CG is present at a frequency above a predetermined threshold; and (g) rejecting CS and combining it with one of the groups being formed if CS is sufficiently distant from existing forms to be part of a new group.