Photogrammetry-The science of quantitative analysis of measurements

from photographs

Photos•Photons•Light

Gamma

•To Draw

Metron•Metric•To Measure

Photogrammetry

What’s the primary objective???

• The primary objective of the technique is to derive precise coordinates of a point.

How this is done???

• This is done by viewing the area from two different angles, thereby recreating the same conditions as it existed at the time of photography.

Concept Of Photogrammetry

How it is done?Mathematically intersecting converging

lines in space, the preciselocation of the point can be determined.

Exterior Orientation

Image Matching

Principals of PhotogrammetryInformation needed to determine 3D locations:1. The location of each camera’s perspective centre.2. The orientation (rotation) of each camera about its perspective centre.3. The location of the point on each image sensor.This is the essence of photogrammetry!

WHY Close Range PHOTOGRAMMETRY???

• Very precise • Time effective• Cost effective• Based on well established and tested

Algorithms• Corrects all sorts of distortions• Wider Scope of Applications

Aim Development of Non-metric Close Range Photogrammetric Applications

Objectives

• Making accurate 3-D models– Can be used for Accurate Measurements

• Making Textured models – Can be used as a final product(Photo-Realistic)

• Making Dense Surface Models– Can be used for Meshing, Contour Maps

Methodology

Camera Calibration

Data Acquizition

Stero All Directional

Calibration

• Camera Calibration is the process of determining the characteristics of a camera so it can be used as a measurement device.

• Characteristics include – focal length,– imaging scale,– image center and – lens distortion.

The Process…

• Deciding on Camera and the Focal length.• Taking Photographs– 12 Photographs of the Calibration Sheet each from

Different Angles (Geometry).• Importing the Images to Photomodeler Pro.• Invoking the Automatic Camera Calibration

Process.

12 Photos

The Outcome

• Focal length – The distance between the optical center of the lens and

where the optical axis intersects the image plane. • Imaging scale

– CCD Format Size • Principal point

– where the optical axis of the lens intersects the photograph• Distortion Parameters

– Radial Distortions– Decentric Distortions

The Parameters window in Photomodeler screenshot

Data AcquisitionPlanning the measurement Project

Involves selecting the number and locations of camera positions for taking the photographs.

Data Acquisition

All Directional PhotographyPhotography from all directions45% change gives good overlap

Data Preparation

1. Feature Marking :

– Involves marking Target Features on Photographs

– Target Features can be formed using Points, curves, Edges, Cylinders or Shapes.

– Higher level features such as lines and surfaces are then built on these marked features.

Screenshot of Feature Marking

Orientation

• Done by Referencing the Photographs– The points occurring in two are more photographs

are referenced – Similarly Edges or Curves are also referenced

The Process is same as that of “geo-Referencing” the satellite image. The Difference here is no Toposheet or Already Referenced Photograph is Used.

A Screenshot of Process of orientation

Idealisation

• An ideal camera has – no lens distortion, – square pixels, and – a centered principal point.

• The images in the project are Resampled and • The photographic marks are shifted to match

the new idealized images.

Distorted image Ideal image

Data Processing

• Point cloud • generation

• Meshing• Surface

• generation• Texturing

• Data • Processing

• Point Cloud Generation1. For each pair of photographs, overlap is checked,

a) align the image rows along the epi-polar lines b) to reduce the image to the region of interest

2. The algorithm searches along a row of the destination image using an N x N patch of imagery from the source image.

3. Where it finds a good match it records it.

4.    All the matches are then optimized for the best overall fit, throwing out bad and weak matches

5.    A sub pixel refinement is carried out for the matches.

6.    The matched positions are used to create 3D points using camera station information.

• Meshing – Equivalent to Triangulated Irregular Network

Surface Generation

Texturing

• Block Adjustment is the process of defining the mathematical relationship between the images contained within a block, the camera or sensor model, and the ground. Once the relationship has been defined, accurate imagery can be created.

Exporting the Data

• The Resulting 3D Data can be exported to CAD or any other Graphics Program for further application specific processing. Various file formats are available in Photomodeler Pro like DXF, 3DS, OBJ, VRML, IGES, 3DM, STL or RAW files.