The objective with this thesis has been to find out if it is possible to utilize low cost image sensors combined with fast digital logic as building blocks in laser based dust monitors. The field of application of these kinds of monitors is within environmental monitoring in incinerators and industrial processes. A new dust monitor based on the principles outlined in this thesis, could potentially reduce the cost of the monitor significantly, since expensive optical and mechanical parts as well as a custom detector can be replaced by a high volume image sensor and digital logic circuits. Additionally it has been an objective to compare the new monitor with a traditional monitor based on mainly optical solutions.
As part of the implementation of a prototype during the work with this thesis, a simple optical and mechanical system has been designed. A suitable image sensor has been selected and a printed circuit board with code for a field programmable array and an embedded processor has been developed.
Three measurement concepts have been evaluated and tested. Two concepts based on scattered light have been tested, of which one was based on blocking the direct laser beam in the center, while the second allowed also the central main part of the beam to reach the sensor. The third measurement concept was based on the measurement of the direct beam or the optical light transmission through the dust.
The camera based dust monitor has been compared with the commercially available LaserDust instrument which is based on measurement of a combination of scattered and direct light. Dust monitors will normally be calibrated in-situ in the process, since different types of dust can lead to different response and therefore an absolute calibration will normally not be possible for dust monitors. The camera based dust monitor and the LaserDust instrument have been compared mounted in two different set-ups. The first set-up was below the ceiling in an office environment corridor and the second in a specially designed dust test chamber.
Measurement results from the comparisons show satisfactory correlation for most tests, but the camera based dust monitor has a somewhat higher noise level than the LaserDust monitor. This appears to be caused by two main issues, namely that the prototype does not fully utilize the available time for measurement and interference and speckle phenomena related to laser light, optics and the image sensor surface. The complexity of these issues and the extensive work needed to resolve them, made optimization a task beyond the scope of this thesis.
Design of an optical system with reduced interference and speckle problems related to the laser light is one of the fields where further work and optimization is needed. Work to increase the effective measurement time should also be done to improve the signal to noise ratio. Room for improvement is also present in algorithms for dust concentration calculation based on image sensor data. It has also been observed that defective pixels might influence the measurements and a system for handling dead or deviating pixels should be implemented in future versions of the dust monitor.
The work during this thesis has shown that it is possible to measure dust based on low cost digital image sensors and digital logic. Even though comparisons between the prototype and a commercial monitor displayed an advantage for the commercial monitor, the camera based monitor has a significant room for improvement by increasing the efficient measurement time and optimizing algorithms for the dust concentration calculation. An improved dust monitor based on image sensors will therefore have a potential to achieve the same performance as a traditional dust monitor, but with reduced component cost.