Sprecher
Beschreibung
High-speed side-view videos of sliding drops enable researchers to investigate drop dynamics and surface properties. However, understanding the physics of sliding requires knowledge of the drop width, which necessitates a front-view perspective. The drop’s width is a crucial parameter due to its association with the friction force. Incorporating extra cameras or mirrors to monitor changes in the width of drops from a front-view perspective is cumbersome and limits the viewing area. This limitation impedes a comprehensive analysis of sliding drops, especially when they interact with surface defects.
To address these challenges, our study explored the use of regression and multivariate sequence analysis (MSA) models to estimate the drop width at a solid surface solely from side-view videos. This approach eliminates the need for additional equipment, ensuring an unlimited viewing area for sliding drops. Among the tested methods, the Long Short Term Memory (LSTM) model with a 20-frame sliding window achieved the best performance, with a root mean square error (RMSE) of approximately 68 µm. Considering the range of drop widths in our dataset (1.6 to 4.4 mm), this corresponds to an error margin of about 2.5%. Notably, the LSTM model provided drop width predictions across the entire sliding length of 5 cm, a capability previously unattainable.
Building on this foundation, our subsequent research advanced the methodology by departing from the reliance on time-series data of specific parameters like contact angles. Instead, our approach processes raw video footage to dynamically identify features most indicative of drop width. This method significantly enhances measurement precision across various environmental conditions and demonstrates robustness against noise, blur, and brightness variations. Our refined neural network architecture achieved an RMSE of approximately 55 µm, improving upon prior results and highlighting the potential of direct video analysis to extract more effective features. This advancement simplifies experimental workflows and improves measurement accuracy in drop dynamics research, enabling more detailed insights into interactions between sliding drops and surface defects.