Pix4D Demo

Introduction:

This lab demonstration was designed to introduce students with Pix4DMapper, a program that transforms a series of aerial imagery into a single compiled mosaic capable of rendering both 2D and 3D surface models. 

An online Pix4D Software Manual was provided to familiarize students with the capabilities and limitations of the program's functions. Through exploration of this manual, students were prompted to answer the following questions related to the software:

-What is the overlap needed for Pix4D to process imagery?

        70% Frontal, 60% Side

-What if the user is flying over sand/snow, or uniform fields?

        More overlapped imagery is needed in these instance: 85% Frontal, 70% Side

-What is Rapid Check?
        Rapid Check allows processing to be completed at a much quicker pace, but with a lower accuracy and resulting resolution.

-Can Pix4D process multiple flights? What does the pilot need to maintain if so?
        The software IS able to process multiple flights, provided that there is a thorough amount of overlap between photos and that the photographs were taken under the same conditions (same relative time of day, weather, ground layout, etc.).
-Can Pix4D process oblique images? What type of data do you need if so?
        Oblique images CAN be processed, provided there is enough overlap in and between datasets. 
-Are GCPs necessary for Pix4D? When are they highly recommended?
        GCP's are NOT necessary for Pix4D. They are however highly recommended during tunnel reconstruciton. 
-What is the quality report?
        The quality report works to identify errors within the processing of the given data imagery. 

Methods: 

Following an in class demo, students began by creating a new project in the program, importing the series of 68 provided images titled "Flight 1", and undergoing initial processing before finally completing the processing with Point Cloud and Mesh processing and DSM, Orthomosaic and Index processing.

The results from the initial processing stage generated a quality report (Figure 1) and a single mosaic (Figure 2) that could be used for 3D rendering after completion of the final two processing features (Figure 3). All 68 of the photos in the series were successfully processed.

Figure 1: Quality Report

Figure 2: Flight 1 Mosaic of Litchfeild Mine, Eau Claire, WI

Figure 3: 3D Rendering of Mine Site

Upon the successful completion of 3D rendering, students were able to use the program to produce a fly-by video visual of the 3D terrain surface examined in Figure 3. The resulting video revealing a glimpse of the 3D model from each of the four directional perspectives has been provided in Video 1 below:

Video 1: Flyover of Mine Site

Data Discussion: 

The overall results from the processing of the original 68 aerial photos proved to be successful in stringing together a singular mosaic which was later used for 3D rendering. Contributing largely to the success of this data processing was the high amount of overlapping images used in the data set. The diagram in Figure 4 illustrates where these areas of overlap occurred, as well as the degree to which overlap occurred. The bright green area occupying the majority of the figure is representative of the areas with five or more overlapping images for every pixel. Contrastingly, the red areas at the fringes of the study area depict a scarcity in image overlap per pixel, probably because it lies outside of the mining site.

                   Figure 4: Overlap

Completion of the second two processing stages resulted in a total of two rasters, one as a DEM, the other a Mosaic, as depicted in the map below. The DEM works to symbolize the elevation of the terrain in the area of interest while the mosaic provides the actual visual frame work. Areas with high elevation are pictured in red on the DEM image while areas of low elevation are pictured in blue.

Final Critique:

In all, the results of this lab are meant to demonstrate the functions and capabilities of Pix4DMapper as well as the potential benefits its use. While this demonstration only revealed a small fraction of the software's potential, it is evident that this program could prove powerful in the geospatial world, analyzing everything ranging from land use and agriculture to architectural and structural projects in urban planning.