Advanced computing and image processing utilised in dashcam imagery study

This paper discussed the use of advanced computing and image processing technique to study dashcam imagery based on collinearity and epi-polar conditions, and multiple-polynomial regression. Existing software was improved to examine the problems relating to multi-component lens distortions conditions. The objective of this paper was to report detail of the study findings based on real-world test-fields. A high quality (professional) digital camera was utilised in the study for imagery quality comparison purposes. The results showed that dashcam exhibits multi-component lens and sensor distortions which required precision real-world test-field calibration. Overall, the study found the distortion properties of dashcam were of roughly 15 times higher than professional digital cameras. Based on these findings, dashcam imageries have limited application in high-precision scientific, engineering or forensic applications. However, coarse estimation of projectile velocity and direction of projectile motion is promising.