A high-resolution, digital x-ray imaging system has been demonstrated during Phase I to provide extremely high quality images of surface-mount printed circuit boards. Defects in solder joints could in many cases be easily identified through visual image inspection,-and could also be identified and characterized through quantitative image analysis. We propose in phase ii to optimize this powerful x-ray imaging system design and develop image analysis algorithms for automatic identification and characterization of solder joint defects. We will evaluate the possibilities for using a similar system as an important part of a near-real-time feedback loop in an integrated quality assurance system in production-line situations. This approach offers important advantages over other inspection modes. Surface-mount methods make many joints inaccessible for visible-light inspection, and other x-ray imaging methods are either much slower (film), or have much poorer spatial and contrast resolution (fluoroscopy). Our patented detection system provides near-real-time data with high dynamic range (12 bits) and sensitivity to defects as small as 0.001 inches in size.