Statistical Video Based Control of Mobile Robots

摘要: A typical approach to the control of mobile robots is based on analysis signals from various kinds sensors (e.g. infra-red or ultrasound). Another type input data used for motion can be video acquired by cameras. In such case many algorithms utilised e.g. SLAM (Simultaneous Localization and Mapping) its modifications, being still an active field research (Kalyan et al., 2010), which require usually full images (or frames) (Se 2001; 2002), in some applications using also additional operations as Principal Component Analysis (Tamimi & Zell, 2004). Similar problems may occur robot localisation map building processes (Hahnel 2003). An interesting alternative application fast image techniques statistical experiments. Presented related first step systems, considered pre-processing technique, reducing amount further with similar accuracy. Discussed methods proportional steering line following (Cowan 2003) a representative example discussions. Nevertheless, other presented are possible object classification shape (Okarma Lech, 2009) detection, well reduced-reference quality estimation 2008b).