King Abdul-Aziz University

Faculty of Architecture & Planning

Geomatics Department

Photogrammetry
Scientific Paper

Abdulwahab Fadelelahi
Osaid Fida

December 24th 2020

Photogrammetry Project:

Summary………….………………………………………………………………………………………………………………………………………..……………………3

Introduction………………….…………………………………………………………………………………………………………………………………………………3

Methodology………..…………………………………………….……………………………………………………………………………………………………………4

Alignment……….………………………………………………………………………………………………………………………………………………4

Point Dense Cloud………………..…………………………………….……………………………………………………………………………….5

Meshing………………………….……………….…………………………………………………….…………………………..……………………………5

Texturing………………………….…………………….……………………………………………………………………………………………………….6

Building DEM………………………………………………………………..………………………………………………………………….…….………7

Building Orthomosiac……………………………………………………………………………………………………………………………..….8

Discussion………………………………………………………….……………………………………………..…………………………………………………………….8

Conclusion………………………………………..…………………………………………………………………………………………………………………………….9

References…………………………………………………………………………………………………………………………………………………………………….10
Summary:

Photogrammetry is the science and technology of obtaining reliable information about

physical objects and the environment through the process of recording, measuring and

interpreting photographic images and patterns of electromagnetic radiant imagery and other

phenomena. Strengths of aerial photogrammetry include; Quick and cost-effective collection

of data, and the ability to capture data in remote, unsafe or difficult to access locations,

lowering safety risks.

Introduction:

Studies that include digital image processing have become very popular over the past

decades. Photographs of an object are captured from different locations/angles. The resulting

‘rays’ (representing the line of sight from the camera to a point on the object), are

mathematically reconstructed to produce 3D coordinates of the point. These 3D coordinates

produce a digital model of the real-world object, allowing precise measurements of features.

This process can generate digital models of real-world objects and measure distances,

heights, areas and volumes and prepare accurate topographic maps.

In the first project data was acquired using a drone, a drone can capture views from

any height and in any direction. It can capture high-resolution photos for a super detailed site

plan. Meanwhile in the second project, capturing the images was performed from close range.

Close-range photogrammetry is meant to be when the distance (range) from the camera to the

object of interest can be from several feet to about 1,000 feet (about 1 to 300 meters). It can

be used in many fields like , biomechanics, chemistry, biology, archaeology, architecture,

automotive, and aerospace, as well as accident reconstruction.

Firstly, the methodology starts with gathering the data. In the first project data was

collected using drone (AscTecFalcon8UAV), an aircraft without a human pilot on board,

which include a UAV, a ground-based controller, and a system of communications between

the two. It can be used to capture Aerial Images. As for the close-range project data was

collected using a personal camera (HUAWEI YAL-L21). The photos are loaded into Agisoft

Metaphase software. Photos are then aligned and converted into a sparse point cloud model.

the program calculates depth information for each camera to be combined into a single dense

point cloud. After dense point cloud has been reconstructed it is possible to generate

polygonal mesh model based on the dense cloud data. The software will wrap the texture

around your model. Then for aerial photographs only, a Digital elevation model can be

generated based on the dense cloud or mesh model. Finally, based on the DEM values, an

orthomosiac photo is produced.

Alignment:

Photo alignment finds matching points between overlapping images, estimates camera

position for each photo and builds sparse point cloud model. You will see a series of dots on

your screen in the rough outline of what you are trying to model.
Dense Cloud:

Based on the estimated camera positions the program calculates depth information for

each camera to be combined into a single dense point cloud. Once this stage is complete, it

will look like you have a completed model, but if you zoom in, you will see thousands of

individual points.

Meshing:

After dense point cloud has been reconstructed it is possible to generate polygonal

mesh model based on the dense cloud data. This final step to generate the base model of your

object. At the end of this process, you will have a 3D model. Height field surface type was

chosen for the drone captured images because it is selected for aerial photography processing

as it requires lower amount of memory. It is optimized for modelling of planar surfaces, such

as terrains.
Texturing:

This option creates a texture based on an analysis of the loaded photographs. Once the

texture is complete, the software will wrap it around your model, adding a much higher level

of detail than the previous model coloring. This is the final step in the close-range workflow.
DEM:

This process is applied on aerial photographs. Digital elevation model can be

generated based on the dense cloud or mesh model. DEM is frequently used to refer to any

digital representation of a topographic surface (raster), and also to determine terrain attributes

such as elevation at any point, slope and aspect. Each pixel in the raster has an elevation (z

value) according WGS84 stored in it. The values of elevations are taken by interpolated pixel

values.
Orthomosiac:

An orthomosaic photo is an aerial photograph geometrically corrected such that the

scale is uniform. The orthomosiac takes the value of the DEM. First the orthomosaic will take

the position from DEM (elevation value), and then the color (RGB value) from the aerial

photo, then the orthomosiac is produced.

Disscussion:

3D rendering takes photogrammetry to the next level. Over time, a number of

different rendering techniques have been developed. Nevertheless, the goal of every render is

to capture an image based on how light hits objects, just like in real life, which is the most

intersting part about it. However, the process is complex and very time consuming, especially
that the number of depth points in the drone taken images project were more than one

million, the process took more than 11 hours which is really the downside of 3D modeling.

But that doesn’t cancel out the fact that the results of 3D model can be very promising and

helpful in many fields. As for the close range project, the process was much smoother and

faster because of the smaller area coverd. The tricky part was taking the photos which is the

most critical phase of the workflow, any blunders in this procedure can affect the whole

project. Photos must be taken carefully and percisley.

Conclusion & Recommendations:

Eventually we were able to learn the strengths and limitations of photogrammetry,

and how powerful 3D modelling is. We’ve also experienced how aerial and terrestrial

photogrammetry works and how are they different from each other. This project has also

taught us many new things like how Mesh surface is built on polygons based on the point

cloud, how photo alignment estimates the camera locations and finds matched points between

overlapped images, and also why Orthomosiac is only generated after building the DEM.

Some recommendations would be:

• Understanding the camera’s interior geometry

• Understanding the shape and size of the area or object before capturing

• Captured target must have good visibility in the photo frame

• Agisoft Metashape User Manual: https://www.agisoft.com/pdf/metashape-