AirVisual sensors provide highly reliable PM2.5 measurements at an accessible price, allowing everyone the opportunity to better understand the air they breath. 

Through a combination of technologies, AirVisual sensors reliably deliver a high correlation against governmental Beta Attenuation reference monitors. These applied technologies include:
​

AirVisual sensors use an internally-developed advanced light-scattering laser sensor to measure PM2.5. Laser sensors represent the most accurate means of measuring PM2.5 in a consumer device. 

The sensor works by shining a laser within a measuring chamber and counting the irradiated light reflected from the microscopic particulate matter (from 0.3 to 10 μm in size). 

Disturbed air flow which is not precisely accounted for, is a common source of measurement error in low cost monitors. The AirVisual sensors use a small fan within the device’s sensor chamber to ensure a constant, calculated flow of air through the measuring chamber.

Picture 1. Simplistic diagram of AirVisual's laser-based PM2.5 sensors

Each monitor is carefully tested and calibrated in a factory setting. Using a controlled pollution environment, devices are exposed to varying pollution levels and calibrated via a computer controlled system. This method ensures high precision to reference monitors and a low intra-model variability.

Picture 2. AirVisual Pro monitor shows 0.99 correlation rate with professional monitoring equipment
​

A light scattering laser sensor’s reading is subject to the impact of environmental factors such as humidity, temperature and pollution composition. 

High humidity, for example, causes more light irradiation and makes particles appear denser - causing inflated PM2.5 concentration figures.

To reduce the influence of these environmental factors, measurements are further adjusted using cloud-based calibration algorithms, which improve accuracy by taking local environmental conditions into account. 

This cloud calibration method was developed in-house, by our data scientists, who used machine learning to analyze meteorology data, satellite imagery pollution composition and aggregated PM2.5 measurements from both governmental and low-cost sources. By recognizing past data trends and correlations, the system learned relationships between various environmental parameters and the laser sensor PM2.5 measurements, as compared to reference monitors. In doing so, it created a correction algorithm to improve the AirVisual sensors accuracy in global environments.

Outdoor deployed AirVisual sensors, which have been made public, are further cross-calibrated and validated against nearby reference monitors. A data point is only published to the AirVisual Platform when it fits the general trend of nearby monitors. In processing and learning from real-time and historical air quality data points, the system can identify anomalous values and remove these from publication before causing undue misrepresentation.