Bitte wählen Sie ihr Lieferland und ihre Kundengruppe
Infrared imagery is an appropriate technology for a large number of outdoor robotic applications. However, infrared images have considerable sensitivity to illumination conditions, limiting in many cases the use of autonomous robotic systems. A general infrared vision system is presented for outdoor applications. Two main approaches have been considered to solve this mentioned sensitivity to illumination conditions. The first one consists in adapting automatically the camera settings in order to control the visibility of the images. The identification of infrared cameras obtains a linear model. This model is employed to develop control schemes using visual feedback techniques. The proposed vision system also includes a fuzzy multi-resolution threshold selection method that considers knowledge on the application and information on illumination conditions to select automatically an adequate threshold value. The adaptation of the method to specific applications can be done easily. The proposed method shows considerable robustness to illumination changes. This vision system has been implemented in several infrared applications for outdoors, one of which is described including some examples. Future work will include the integration in field robotics systems for applications such as surveillance and forest-fore fighting.