ABSTRACT
At present, the height of a baby is detected manually by a nurse. In order to realise rapid non-destructive testing for the detection of a baby’s height and weight, a module was established to detect the height of baby based on machine vision in a smart baby weighing system. This paper proposed a fast detection method for finding a baby’s height using the GrabCut segmentation algorithm, based on the image captured by a monocular camera in a smart baby weighing system. First, a baby was marked by preprocessing and a body detector. Second, by accurately extracting the feature points of the top of the head and the feet, the real height of the baby is calculated accurately using the theory of the GrabCut segmentation algorithm. The test results proved that the method has obtained the precision of the manual method at present and been adaptive to the angle of camera and baby moving posture. The simple and efficient structure of the detecting system ensures its application.
