Stereo vision based automated solder ball height detection

Jinjin Li, Bonnie L. Bennett, Lina Karam, Jeff S. Pettinato

Research output: Chapter in Book/Report/Conference proceedingConference contribution


Solder ball height inspection is essential to the detection of potential connectivity issues in semi-conductor units. Current ball height inspection tools such as laser profiling, fringe projection and confocal microscopy are expensive, require complicated setup and are slow, which makes them difficult to use in a real-time manufacturing setting. Therefore, a reliable, in-line ball height measurement method is needed for inspecting units undergoing assembly. Existing stereo vision measurement techniques determine the height of objects by detecting corresponding feature points in two views of the same scene taken from different viewpoints. After detecting the matching feature points, triangulation methods are used to determine the height or depth of an object. The issue with existing techniques is that they rely on the presence of edges, corners and surface texture for the detection of feature points. Therefore, these techniques cannot be directly applied to the measurement of solder ball height due to the textureless, edgeless, smooth surfaces of solder balls. In this paper, an automatic, stereo vision based, in-line ball height inspection method is presented. The proposed method includes an imaging setup together with a computer vision algorithm for reliable, in-line ball height measurement. The imaging set up consists of two different cameras mounted at two opposing angles with ring lighting around each camera lens which allows the capture of two images of a semi-conductor package in parallel. The lighting provides a means to generate features on the balls which are then used to determine height. Determining and grouping points with the same intensity on the ball surface allows the formation of curves, also known as iso-contours, which are then matched between the two views. Finally, an optimized triangulation is performed to determine ball height. The method has been tested on 3 products and exhibits accuracy within 4um mean squared error compared to confocal ground truth height, and the coplanarity of BGA package as derived from calculated substrate depth results.

Original languageEnglish (US)
Title of host publicationIPC APEX EXPO 2014
PublisherIPC - Association Connecting Electronics Industries
StatePublished - 2014
EventIPC APEX EXPO 2014: New Ideas... For New Horizons - Las Vegas, United States
Duration: Mar 25 2014Mar 27 2014


OtherIPC APEX EXPO 2014: New Ideas... For New Horizons
Country/TerritoryUnited States
CityLas Vegas

ASJC Scopus subject areas

  • Hardware and Architecture
  • Electrical and Electronic Engineering


Dive into the research topics of 'Stereo vision based automated solder ball height detection'. Together they form a unique fingerprint.

Cite this