Detection of fetal umbilical cord is usually done by a gynecologist by naked eye using ultrasound equipment. However, the results of the ultrasound image having a low image quality leading to problems of its own to identify the umbilical cord. Digital image processing on a computer may increase or improve the quality of the image and finding a particular area, the advantages are used to detect the umbilical cord to the fetus. Pipeline image processing of this application consists five phases: the Image Acquisition stage, Image Enhancement stage, Image Segmentation stage, Representation Description stage to get the area of the cord and morphological Image stage to separate areas with the baby's umbilical cord. This applications built with C ++ programming language, and tested using a functional testing for each stage. Functional testing is intended to test the algorithms used. The test results of using contour tracing method required USG image with contrast color between the baby and umbilical cord is needed, so it goes well in image segmentation stage. The experimental results of 10 samples, 5 samples contained detectable, or 50% success rate. Success rate can be raised by using an image that have a contrast color between the baby and the umbilical cord.