A 1.890Mb PDF of this article as it appeared in the magazine complete with images is available by clicking HERE
I was recently giving a talk on mine site mapping using small Unmanned Aerial Systems (sUAS) where I was sort of verbally attacked about licensure. It is bad enough that the FAA requires a pilot license to fly a model airplane and now this person in the audience was indignant that I was not a big advocate of also requiring a professional land surveyor (PLS) license.
Recently we have been doing a bit of service work ourselves with sUAS. Most of this is with customers who will eventually internalize data collection and processing by acquiring technology from our company. Several of these companies have asked us fill the role of an independent mapping company for their annual stockpile surveys. This has led to communication between our company and accounting firms. The accounting firms are very interested in organizing a collection of bona fides from us as an assurance that we know what we are doing. We have provided various and sundry information that has sated their queries. Much of this information has been detailed descriptions of the process and the quality checks that are inserted to assure accurate products. However, I feel that their desires would be fulfilled by simply providing a name associated with a PLS.
In the meantime, we have been honing our processes for creating deliverable products from dense image matching data (see Figure 1, for example). The typical deliverables include stockpile toes, stockpile volumes, high resolution digital contours and, of course, a digital orthophoto mosaic. We have organized a general process for creating these products with a set of QC steps along the way. The interesting twist is that we are, at heart, an engineering software development company so we have the luxury of modifying and tweaking our product (mostly LP360 for sUAS) along the way.
One of the big concerns that we have is consistency of the products. This is a repeat business where a mine site might need a survey every 90 days. How can we assure that this month’s data are consistent with the data collected last quarter? Above all else, how can we assure that the products are accurate (whatever that might mean) with respect to relative measurements or a network? This is really the question being asked by the audit firms although I am not so sure they realize this.
Now I am not a basher at all of credentials. After all, is this not the function of a college degree–an indicator that one has completed a series of courses that renders competence in a particular area? Unfortunately, this set of standardized courses and examinations does not necessarily mean one has the expertise for a specific set of technologies. I have met Professional Engineers (PE) in electrical engineering (my own field) who have never worked on a guidance and control system. Should they be certifying the pointing accuracy of an imaging satellite? Similarly I have met many PLSs who have not done any work in geodesy (certainly a necessary skill set these days when one is doing vertical modeling).
Something that was recognized by W. Edwards Deming (the father of modern quality control) is that the focus must be on the process as much, if not more, than the persons carrying out that process. One needs to design a scientific process with measurement metrics at critical points that are able to detect anomalies. Perhaps I am a bit slow but it occurred to me as I was crafting a letter to an accounting firm that this is certainly the case for our mine site mapping situation.
Consider the simple case of metric accuracy (see Figure 2). We use an on-board Real Time Kinematic (RTK) Global Navigation Satellite System (GNSS) to obtain the camera exterior orientation at each exposure. What is the proper procedure for assuring the EO itself is correct? Well, on the one hand, a photogrammetrist would be trained to recognize, for example, that camera focal length is highly correlated with vertical scale. On the other hand, a PLS (one who was not grandfathered in to licensure prior to GNSS surveying) would know how to interpret an Online Positioning User Service (OPUS) report from the National Geodetic Survey. It is actually rare to find both of these skills in a single person not trained in this particular process.
My point (pun intended) is that we need to quit focusing so much on the bona fides of the individuals involved in complex data collection (after all, there are many of them) but instead focus on the process. If I were purchasing data from a company, this is what I would do. I would not ask questions about the licenses held by specific individuals but rather would ask for a detailed workflow and the QA steps in that flow.
Of course where this approach breaks down is that customers usually do not have the staff with the skill sets necessary to assess the validity of a specialized engineering mapping workflow. Can we address this by using companies specialized in a particular area of expertise to certify processes? Today large mapping projects often employ an independent QA/QC contractor to assure the delivered data meet the delivery specification. I posit that the process itself should also be examined. Examination of the products without regard to the process will work much of the time but not all of the time. It takes a specific configuration of check points to detect a focal length error. This vertical scale error can often slip through a product inspection process.
Of course, I am not offering a solution here. I am simply pointing out that we need to think about the fact that quality products emerge from correct and monitored processes, not from the skill set of folks who might not have even been involved in the process design.
Lewis Graham is the President and CTO of GeoCue Corporation. GeoCue is North America’s largest supplier of LIDAR production and workflow tools and consulting services for airborne and mobile laser scanning.
A 1.890Mb PDF of this article as it appeared in the magazine complete with images is available by clicking HERE