3D machine vision is rapidly replacing 2D machine vision for quality inspection and defect detection. Advantages of 3D machine vision include: 1) effective for low-contrast and highly-reflective products; 2) does not require complex lighting; 3) surpasses the capability of legacy 2D vision systems; 4) provides additional dimensional data to improve vision inspection tasks; 5) high resolution; 6) high speed; 7) low cost; and, 8) high accuracy.

In the past, laser triangulation sensors have been used for 3D inspection tasks. However, laser triangulation laser/camera systems are often too expensive for many applications. With recent breakthroughs in 3D time of flight (TOF), stereoscopic, and structured-light sensors, 3D industrial machine vision can now be performed with low-cost commodity sensors that were developed for a number of other applications, including gaming, entertainment, and robotics. For example, Shape Ape Standard was released with the Microsoft Kinect v2, a high resolution 3D TOF sensor that is effective for a wide range of 3D inspection tasks.

Defect detection is one area of 3D machine vision inspection that is growing rapidly. Defect detection can be performed at regions of interest on the part. In such cases, the user defines thresholds for each region of interest, and measurements above that threshold are tagged to indicate a defect on the part. Another way of performing detection detection with 3D machine vision is to develop a catalog of 3D profiles of “perfect” parts and measure the full 3D profile of each manufactured part. Thresholds can be defined by the user to determine whether or not a part is defective, based on the difference between the measured shape and the shape of the perfect part. 3D inspection for defect detection is extremely valuable for cases where 2D inspection struggles, for example, when there is very little contrast in the 2D image. 3D machine vision, on the other hand, compares deviations of the part’s profile, which can be much more effective for defect detection.