Photogrammetry has changed\u00a0substantially in the past three decades. When Jeff Yates began in the business 32 years ago, he used a projection system known as a Kelsh two-bucket plotter. \u201cThat was four generations ago of photogrammetric systems hardware,\u201d he recalls. \u201cFrom there, we went to the analog stereo plotter. That was the second big change in photogrammetric equipment during my compilation tenure. Later on, we switched over to the analytical instruments. Now, everything has gone digital, and all of those highly precise pieces of machinery and optics are functionally obsolete.\u201d<\/p>\n
These rapid changes in equipment have been disruptive, forcing practitioners to spend time learning new systems. \u201cThose operators who did not change very readily from the mechanical instruments\u2014both the analog and the analytical\u2014to the digital ones, are no longer in the photogrammetric field,\u201d says Yates. Because of this rapid technological change, the key challenge in photogrammetry today, he explains, consists of \u201crecognizing what the latest trends are, learning to understand them, turning them into a commercial product, and earning a return on our investment before\u00a0that<\/i>\u00a0becomes obsolete.\u201d<\/p>\n
Despite all the advances in hardware and software, however, some key aspects of photogrammetry have not yet been fully automated with engineering-grade reliability. For example, software cannot yet reliably differentiate between building edges, utilities, and fence lines or between a small car and a rock boulder. Therefore, human operators are still required to interpret what they see in the 3D stereo model. \u201cThe ultimate goal is to make as much of the whole workflow as automated as possible,\u201d says Yates.<\/p>\n