3-D imaging of blood flow based on 2-D ultrasound scanning and position sensor information can be used to portray clinical information not currently available in vascular and peripheral imaging. During surgery, information of blood flow is important to identify and avoid damage to important vessels. 3-D intraoperative ultrasound flow imaging has proven useful in this respect. However, due to Doppler angle-dependencies and the complexity of the vascular architecture, the generated 3-D image of blood vessels often have drop-outs in the vessel structure. In this work we aim to reduce the number of drop-outs in the 3-D blood flow images by combining 2-D flow images with different transmit angles. The compounded 2-D images are further reconstructed to a 3-D flow image based on information from an accurate position sensor system. The method was implemented on a research ultrasound scanner for real-time processing of the data, and in vitro trials showed promising results. The method also showed promising results in vivo using a clinical approved high-end scanner, improving the 3-D image quality of the feeding arteries of an arteriovenous malformation.