Not too long ago, when someone in the surveying & mapping community talked about a project being accurate, it meant that the info being gathered was accurate to within two to three inches. Survey-grade accuracy was the expectation in various contexts and at multiple levels. However, drones have shown professionals across the industry that gathering such detailed info isn’t always necessary, and also provided the means to gather that less accurate info.

All of this has changed the acceptable level of accuracy that needs to be collected on a given project, but it’s still a matter of context. Anyone looking to survey clean stockpile yards of relatively fine material will want and need a different level of accuracy than someone working on a project with topographic contours. Problems arise when stakeholders want to gather less accurate info when the project calls for a greater level of detail. On the other hand, gathering survey-grade accuracy when it isn’t necessary can create unnecessary complication. Regardless of the specifics, accuracy needs to be a product-driven requirement.

What sort of tools can a surveyor use to gather info that isn’t survey-grade accurate? How can the requirements around accuracy be best explained to stakeholders? In what ways will gathering info that isn’t accurate enough impact an overall project? These are just some of the questions surveyors need to consider as they explore what accuracy can and should mean on a given project.

collecting-control-points

Surveyors and Stakeholders

Many stakeholders and consumers of data do not have a great understanding of accuracy. Accuracy is often a topic that had not crossed their mind when they thought about creating a map or measuring a volume. Terms like relative accuracy and absolute accuracy might not be especially familiar to them, and they might think of data in terms of being within a certain distance of reality rather than something far more specific.

Whether or not the project in question can work with a lower level of accuracy is a question surveyors need to work through with a client, and they need to be prepared to offer options if that lower level of accuracy is acceptable. It’s something that Lewis Graham, President/CTO of GeoCue Group is looking to provide with the Bring Your Own Drone (BYOD) Mapping Kit.

“There are a lot of competent spatial data providers (for example, land surveyors) who want to add drone mapping technology to their equipment and services complement,” Graham said. “Imagine the value added to a survey if you can provide an up-to-date ortho mosaic! However, most folks do not want to jump into this often risky business by investing in a very high end, survey grade drone. They want to go low cost, see if the business is going to pan out and then move up as they gain experience and customers.”

Being able to meet the needs of customers who have less rigorous accuracy needs is a critical consideration. Stakeholders who realize they don’t need survey-grade accuracy are typically not going to be willing to pay for it. Many surveyors have tremendous measurement science knowledge, but they may not be experienced with how to create accurate images and point clouds using a low cost, non-metric camera on a flying platform. The BYOD Kit offers that exact service.

That said, accuracy has been and continues to be a product-driven requirement, which means gathering a lower level of accuracy is not always the best choice. Drones without direct geopositioning represent the means to gather less accurate info, but doing so can present problems further down the production lifecycle when survey-grade accuracy is needed.

 

Misconceptions and Challenges

The proliferation of drone technology has meant that just about anyone can purchase a drone and send it into the air to collect data. Additionally, some professional surveyors have been anxious to take advantage of the capabilities drones are able to provide even if they don’t yet have a complete understanding of how to effectively utilize the technology. Problems can arise if data are collected without a full consideration of the products to be created from the collection. Data collected without proper control will not meet survey accuracy requirements. On the other hand, paying for survey grade data for mapping grade questions may waste time and money.

Consider the classic “borrow pit” example that Graham mentioned. This is an excavation where dirt is being purchased for use at another site. The borrow pit owner decides to use point cloud modeling technology to compute the amount of dirt that has been sold. This computation is performed by creating before and after point clouds. A volumetric difference between the clouds is then computed. If the two clouds are not very accurately matched to the same vertical reference point then the difference will not reflect what has been sold.

For example, each 1-inch of vertical error creates a volumetric error of about 134 cubic yards per acre. Graham has seen end users buy a drone and a point cloud generation application and do this computation with no knowledge of referencing. They are then quite surprised at the huge difference between the number of truckloads of dirt taken from the site and their volume computation. There are obviously huge budgetary ramifications for these kinds of mistakes. The problem is their technique but it is often the entire technology that is blamed.

Even simple stockpile volume computations where only relative accuracy is needed require care. One of the major issues that affects relative accuracy is proper calibration of the camera being used for data collection. This is an aspect of the overall workflow that is frequently overlooked. Bad calibration leads to scale errors.

“There is a lot of misconception about doing differential computations,” Graham continued. “These are things such as volumetrics or height where you are not concerned with the location of objects with respect to a reference network – just the length or area or volume. Many practitioners new to the art believe that accuracy does not need to be considered in these cases. However, it is analogous to measuring a board with an incorrect tape measure. Someone asks you to build a box that is exactly 3 feet on edge. If your tape measure is off by 8%, the box looks great from your perspective but the customer, who is using her own accurate ruler, measures a box that is 33 inches on edge. An embarrassing mistake!”

 

A Product Driven Requirement

So what’s the best way to avoid such issues and determine the level of accuracy needed on a given project? That’s a simple question that used to have a simple answer, but the realization across the industry that survey-grade accuracy isn’t always necessary has complicated things. Accuracy has to be a product-driven requirement, but determining those requirements might need to be a wholly separate endeavor.

The person requesting a mapping service needs to articulate the products they desire,” Graham mentioned. “For example, in most of our mine site mapping projects, the customer wants an orthomosaic image of the area, volumetrics of the stockpile area and often 2’ contours of the pit area. The service provider should provide an accuracy proposal to the customer because, first of all, the provider is the expert and secondly, the customer needs accuracy put in terms that are understandable.”

In the example Graham mentioned, the contours will be the driving element of the accuracy plan. As a rule of thumb, the accuracy required to support contours is about 3x the contour interval at 95% confidence. Thus, the requested two feet contours will require this project to have a network vertical accuracy of about 2/3 foot at 95%.

That’s the kind of specificity that can only be determined through a conversation about such details. In general, the customer should discuss the products they need with the service provider. The service provider needs to consult with the customer to ascertain the desired horizontal and vertical spatial reference systems, talk through the nuances of accuracy with the customer and then propose a written plan. That plan doesn’t need to be more than a page that spells out the reference systems and strategy.

Not only does this build up the confidence of the customer, but it can also help when things get complicated. Such things happen when, for example, a customer provides a priori data in an unknown reference system.

 

Accuracy in depth - Topolyst: Stockpiles with toes automatically collected and computed against a predefined, breakline enforced surface model

Topolyst: Stockpiles with toes automatically collected and computed against a predefined, breakline enforced surface model

Incorporating Drone Mapping Technology

We’ve already laid out how and why drones will eventually be just another tool in a surveyor’s toolkit, but the logistics around how and when that will become a reality are less certain. Surveyors in different areas, who focus on different types of work, and deal with projects of various scale all have different levels of understanding around how they can and should be using drones. When you factor in the level of understanding stakeholders may or may not have, it makes that future even less certain.

Nonetheless, it’s clear the technology is here to stay, which means anyone focused on gathering data for survey projects of all types and sizes needs to know where to start and where to go in terms of being able to utilize a drone. That’s the exact need the BYOD Drone kit was designed to address.

“It bundles all of the software needed to do serious mapping with a low cost platform such as the DJI Inspires and Phantoms,” Graham concluded. “Most importantly, it provides bootstrap and continuing training on the various techniques that need to be employed to successfully complete projects. If the user decides to move upscale to, for example, an RTK or LIDAR equipped drone, all of the elements of the BYOD move right along with this progression. They will already have the required exploitation software and the knowledge of how it is used.”

The kit itself includes Agisoft PhotoScan Pro for creating point clouds and GeoCue’s Topolyst for exploiting the data. PhotoScan has a number of tools for evaluating accuracy and performing calibration. Topolyst has a set of analysis tools that allow a user to determine accuracy according to the latest American Society for Photogrammetry and Remote Sensing (ASPRS) horizontal and vertical accuracy standards. This allows a user to deliver quantified accuracy reports to their customers.

Users will learn all of the fundamentals with the BYOD. The training aspect of the system provides a great deal of background with respect to the operational aspects of accurately collecting data. For example, when is control necessary? How much control and how should it be laid? What is camera calibration and how is it achieved for these low cost, non-metric cameras? The experience of engaging in metric mapping via the tools and training provided with the BYOD move directly upscale as users advance their technology.

Drones have changed the expectations for both surveyors and stakeholders, and exactly how that will impact the industry as a whole is a topic that will continue to work itself out over the next few years. That’s why it’s more essential than ever for surveyors to be able to speak to, serve and provide the specific accuracy needs of stakeholders.

 

For a more complete look at the features and capabilities of the Bring Your Own Drone (BYOD) Mapping Kit as well as a section specifically focused on the technical details around what accuracy means, download Drone Mapping Considerations: Mapping with Low Cost Platforms