Monitoring pollutant emissions in an urban area is of primary importance in order to guarantee high levels of human health (by reducing the impact of respiratory diseases on urban inhabitants) and good preservation of objects (such as historical sites or buildings). The European Environment Agency (EEA) has established that the maximum concentration of PM2.5 concentration for long-term exposure is 10 μg/m3. By measuring the light that has been backscattered by atmospheric particulates, the lidar technique seems to be a valid method to provide a warning signal when a particulate concentration threshold is exceeded. In addition, the lidar system can also provide information about the distance at which the particulate threshold is exceeded. In this research measurements of traffic-originated particulate in the area near the University of Rome “Tor Vergata” is analysed at different times on a weekday to obtain continuous monitoring of its backscattering coefficient, values of particulate concentration and details about time and distance from the lidar system at which a warning signal is provided. The University of Rome “Tor Vergata” is in an area where traffic is very heavy, so high pollutant concentration values are to be expected.

Optical spectroscopic techniques, such as Laser-Induced Breakdown Spectroscopy (LIBS) or Laser-Induced Fluorescence (LIF), have already been used to study and detect Biological Agents (BAs). Unfortunately, BAs usually share similar-shaped emitted spectra and low-signal intensities, making their detection and classification difficult to assess. Least-Square Minimisation (LSM) based algorithms are usually deployed to measure the concentration of agents from spectra. Recently, it has been shown how the use of ad hoc weights can help in improving the performance of the concentration evaluation. More specifically, it has been observed that the “weight matrix” should be modelled as a function of the boundary conditions of the problem. This work proposes a new weight matrix that is based on the Signal-to-Noise Ratio (SNR) of the measurements. The idea is based on the fact that more noisy data are less reliable and therefore weight should be lowered. The paper, after a brief introduction and review of the LSM applied to spectra, will show the new methodology. A systematic analysis of the new algorithm is done and the comparison with the other LSM algorithms is presented. The results clearly show that there is a range of parameters for which the new algorithm performs better.