Editors Note: This article is the first of three in a series on UAS planning. This first article from Cheung and Bealo explores practical principles for making a software decision.
The criteria for selecting an appropriate unmanned aircraft systems (UAS) data processing software is very similar to picking other data processing software packages. Aside from the obvious factors of cost and functionalities, we will apply Damsgaard and Karlsbjerg (2010)’s Seven Principles for Selecting Software Packages[i] to highlight other considerations that should be accounted for in choosing an appropriate software solution.
The cost of UAS data processing software varies a great deal, from free, open-source programs (e.g. OpenDroneMap) to proprietary software packages (e.g. Pix4Dmapper, PhotoScan) that may cost thousands of dollars. Keep in mind that different software packages may have special licensing options (i.e. educational) that may significantly reduce the software cost. In addition, many UAS data processing software also may offer monthly (e.g. DroneDeploy, Pix4Dmapper) or point-based (e.g. Maps Made Easy) subscription, which may also reduce cost.
Although some software packages may automate the entire process and require less user inputs than others, the data processing workflow and basic capabilities between different software packages are comparable. The general workflow in UAS data processing involves adding and aligning photos, and creating point clouds and meshes, which permits creation of a variety of outputs such as 3D models, orthomosaics, contour lines, and digital elevation models. Some software may offer additional outputs such as various plant index calculations (e.g. Normalized Difference Vegetation Index, Soil Adjusted Vegetation Index) that are traditionally done in third party remote sensing software. Understanding the assumptions and limitations of these indices is essential to avoid misinterpretation. By applying the conceptual framework proposed by Damsgaard and Karlsbjerg (2010) in their qualitative study of software acquisition, we examined the issues of user network and support, standardization and compatibility, and long term planning as they relate to UAS data processing software.
User Network and Support
When a user decides to purchase a software package, the user is essentially joining a network of other users and developers of the software. It is crucial that this network is of a sufficient size, and consists of adequate number of end-users and developers who can configure and support the software. Since UAS hardware (e.g. airframe and sensor) are constantly evolving, UAS software, which may be used for mission planning or data processing, should have a critical mass of users and developers capable of making critical updates in response to changes in the hardware and user needs.
Standardization and Compatibility
Standardization of the user interface and output are two important considerations in making software purchases. Standardization of the user interface provides users with a familiar interface, reducing the time required for user training. A popular example of user interface standardization is the move from standard tool bars towards a “ribbon” interface, which consists of multiple tabs, each filled with large graphical icons. To the authors’ knowledge and experience with currently available UAS data processing software, there does not seem to be a concerted effort at user interface standardization at the moment. One exception to this is Esri’s Drone2Map, which adopted the “ribbon” interface seen in their other mapping products and other popular software such as Microsoft Office and AutoCAD. Guaranteed compatibility regarding output is another important consideration by the users and other third-party software. A sample of some common standard output formats for a UAS workflow are listed below:
- Point clouds: .las
- 3D models: .obj, .stl
- Digital elevation models: .tiff
- Orthomosaics: .tiff, .jpeg, .kmz
Display, analysis, or 3D printing of the output within a third-party software (e.g. Google Earth, MatterControl) requires standardized output. Some software packages also provide users with the ability to build custom tools or automate workflows, which can save users valuable time and expand the software’s capabilities. However, users should take into consideration that the customizations and tools may not be supported in future releases of the software, potentially resulting in the loss of time and resources initially expended to implement the tools or customizations.
Long Term Planning
While it may be difficult to predict the viability of UAS data in various industries, organizations should deliberate on their visions for UAS in their workflows in order to decide on an efficient UAS data processing solution. For example, while an organization with constant need for UAS data and processing capabilities (e.g. weekly construction site inspections, daily crop surveys) may opt for a software with the perpetual licensing option (e.g. PhotoScan), another organization with infrequent UAS data needs may be best served by a software that is based on a point system (e.g. Maps Made Easy). UAS software needs to be constantly updated to be compatible with new sensors or aircraft models. Researching the history and probable trajectory of a software company to ensure that the software has a critical mass of users and developers to support future developments and upgrades is essential.
Users should also be aware that each UAS data processing software has different hardware requirements, possibly making it cost-prohibitive to switch once the investment has been committed. We considered cloud-based software packages such as DroneDeploy with monthly subscription packages to be on the low end of investment spectrum, because all the processing tasks are done on a remote server. Users are not required to maintain any hardware or software (aside from the computer and internet connection needed to upload UAS data and view/export the finished products). On the high end of investment spectrum are costly software packages that are installed on local user machines or servers, requiring sizable financial investments to acquire the software and potential hardware upgrades (e.g. graphics card, computer memory, processor) to perform the intensive UAS data processing tasks.
Given the ubiquitous nature of UAS technology, information derived from UAS are enabling many industries to perform routine tasks at a level of detail and speed that is unprecedented. To efficiently transform UAS data into useful information, choosing the proper UAS data processing software is a cornerstone to a successful workflow. Several principles for selecting software packages should be considered in addition to the obvious factors of cost and capability. Rather than viewing the number of considerations listed in this article as an invitation to wait until a seemingly “perfect” software comes along, internalize the considerations to inform a purchasing decision, and become a part of the network that designs the appropriate UAS data processing software for your industry.
Work reported in this article are supported by the National Science Foundation under Grant ATE #1700552. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.
Damsgaard, Jan, and Jan Karlsbjerg. “Seven Principles for Selecting Software Packages.” Communications of the ACM, vol. 53, no. 8, Jan. 2010, pp. 63–71., doi:10.1145/1787234.1787252.