Shadowgraphy is one of the most popular imaging techniques to characterize moving particles by their size, geometry as well as velocity, due to its simplicity. However, it requires advanced image processing to handle various image defects such as non-uniform illumination, overlapped particles, etc., which are normally only solved for individual applications. This study proposes a robust image processing method for particle shadowgraphy, aiming to process imperfect particle shadow images. The proposed method first detects qualified particles from particle shadow images, and then processes detected particles individually. Therefore different defects from different particles can be handled separately and locally. An overlapped particles detection and separation algorithm is also implemented to improve the accuracy of size and geometry characterization. The proposed method is first proved by synthetic generated particle shadow images, followed by a proof test with shadow images from a transparent dot pattern target. Finally this method is successfully applied to a shadow image acquired from a water spray and proved to be able to handle various issues of shadowgraphy.
KeywordsShadowgraphy, image processing, particle characterization
IntroductionOver past few years, there has been a growing interest and also increasing availability of commercial image analysis software for atomization characterization by means of shadowgraphy for spherical as well as irregular shape particles. Compared with point measurement techniques such as phase Doppler anemometry, it is easy to setup and the results are more straightforward to understand. Additionally, it can be applied for particles with any shape [1, 2] and it is not affected by multi-scattering effects [3,4]. However, most of imaging process methods for shadowgraphy are made for individual applications and often demand high image quality [5], which cannot be achieved in many cases due to the complexity of the spray or the limitation from optical access. In general, these image processing methods are facing various challenges like too many droplets with various shapes in the field of view -which results in overlapped particles [6] or out-of-focus particles, non-uniform illumination, etc. In this study, an advanced shadowgraphy image processing method is proposed to handle shadow images from complex sprays. The principle of this method will be firstly introduced, followed by verification from synthetic images, dot target images with known dot sizes, and then a validation with shadow images from a real spray.