2019 NYSAPLS Conf> Fundamentals of Photogrammetry for Land Surveyors

Transcription

1 2019 NYSAPLS Conf> Fundamentals of Photogrammetry for Land Surveyors George Southard GSKS Associates LLC

2 Introduction George Southard: Master s Degree in Photogrammetry and Cartography 40 years working in the mapping industry Owner GSKS Associates LLC Consulting for the Geomatics Profession

3 data processing using the science of Photogrammetry

4 Definitions: Remote Photogrammetry The science or art of obtaining reliable measurements by means of photographs. Photogrammetry is the art, science, and technology of obtaining reliable information about physical objects and the environment through the processes of recording, measuring, and interpreting photographic images. (ASPRS, 1980)

5 Definitions: Analog Photogrammetry Using optical/mechanical/electrical instruments, to perform measurements from images printed on paper, film or glass, thus creating stereographic views of the image space for the purpose of 3D measurements. Analytical Photogrammetry Using equipment similar to those used in analog photogrammetry but which have computer/electronic components added for more efficient measurement of photographic images. Computer software programs are also a key element in these operations. Digital or Softcopy Photogrammetry The performance of photogrammetric operations using digital rather than hardcopy images. This work is accomplished primarily using computer, monitors, and a specially designed mouse along with sophisticated software.

6 Key Technologies and Terms Photogrammerty The Science of making maps from stereo imagery o Stereo (3D) Imaging o Ground Control (reference points, GPS, GNSS) o Aero-Triangulation (georeferencing) o Stereo Compilation o Topographic Mapping o Planimetric Mapping o 3D Point Clouds o Orthophotography

7 History

8 A brief history of Photogrammetry Origins of Remote Sensing Remote sensing began with aerial photography First photographs taken in 1839

9 A brief history of Photogrammetry Paris by Nadar, circa Gasper Felix Tournachon "Nadar" takes photograph of village of Petit Bicetre in France from a balloon.

10 A brief history of Photogrammetry City of Boston by Black and King (1860), from hot air balloon

11 A brief history of Photogrammetry

12 A brief history of Photogrammetry Major developments in aerial photography WW1

13 A brief history of Photogrammetry After the war the technology was in place to begin large scale aerial surveys

14 Foundational Principles

15 Photogrammetric Fundamentals First Assumption: the photo image is a flat planar surface

16 Photogrammetric Fundamentals Second Assumption: There are planar distortions in all photo images: - Distortions come from two sources 1) the camera platten for film or the CCD platten for digital images 2) the camera lens(s)

17 Known constants and variables..

18 Photo Orientation Six positions of orientation are needed to georeference each photo Ω Omega Yaw Φ Phi Pitch Κ Kappa - Roll x Longitude y Latitude Z - Elevation

19 Aerial Photography -Stereo pair Over lap about 60%

20 Precisely controlled image capture.. Over lap about 60%

21 Types of Photogrametry

22 Photogrammetric Types Photogrammetric Types from Applications Point of View (d is distance from camera to object) Close Range Photogrammetry Aerial Photogrammetry Space Photogrammetry d<50 m 50m>d<15km d = 300 km+

23 Close Range (terrestrial) Photogrammetry

24 Close Range (terrestrial) Photogrammetry

25 Aerial Photogrammetry

26 Space Photogrammetry Extraterrestrial pictures taken from spacebased cameras

27 Photogrametric Imagery

28 Introduction Types of Images Panchromatic, Black & White, Grayscale Color - Red Green Blue (RGB) Multispectral (RGB + Infrared) Hyperspectral

29 Panchromatic Image Sensitive to light in the nm range

30 Black and white Image

31 Grayscale Image

32 False Color composite image

33 True Color composite image

34 Types of photographs (categorized by tilt) Vertical - camera axis as nearly vertical as Possible Oblique - camera axis intentionally tilted Low Oblique High Oblique

35 Types of photographs (categorized by tilt)

36 Vertical - Aerial Photo Mainly used for mapping

37 Low Oblique Aerial Photo Seldom used for mapping Low oblique (no horizon)

38 High Oblique Aerial Photo Horizon line in the photo Typically used for 3D city modeling

39 Aerial images are not maps! Characteristic of a Map vs a Photo Maps are based on parallel projection while photo has central projection Maps have a unique scale. Photo scale varies depending on terrain relief and degree of radial distortion

40 Image Acquisition for photogrammetric mapping

41 Image Acquisition

42 Precisely controlled image capture..

43 Precisely controlled image capture..

44 Image Acquisition Photos taken in parallel flight strips

45 Image Acquisition Each successive photograph overlap previous photo

46 Ground Control

47 Types of Ground Control Points

48 Types of Stereo Model control layout Full Stereo Model Control Stereo Model Control for Aero-triangulation

49 Types of Stereo Model control layout Full Stereo Model Control with Aero-triangulation

50 Ground Control Point Planning

51 Types of Stereo Model control layout Aero-Triangulation - (georeferencing all images for a unified block of ground control positions and tie points)

52 Photogrammetric Instruments

53 Direct Optical Projection Stereo plotters MULTIPLEX MODEL SKETCH

54 Direct Optical Projection Stereo plotters Kelsh Optical/ Mechanical Stereoplotter 1930s to 1970s

55 Direct Optical Projection Stereo plotters Wild Heerbrugg A8 Analogue Optical Mechanical Stereoplotter 1960s to 1980s

56 Direct Optical Projection Stereo plotters Wild BC2 analytical stereo-plotter. Analytical Optical/Electronic Stereoplotter 1980s 1990s

57 Softcopy (digital) Stereo plotters Digital Softcopy Stereoplotter 1990s - Present

58 3D Aerial Film Cameras Wild C Wild RC Wild RC30 & Zeiss TOP s

59 Imaging Cameras for Manned Aircraft Large Format RGB and IR 200 MP Medium Format RGB or IR MP Small Format RGB or IR MP

60 Imaging Cameras for Unmanned Aircraft High Quality Photogrammetric Mapping Calibrated Lens Distortion, Mid- Exposure Pulse, Fixed Focal Length Medium Quality Photogrammetric Mapping, Lens characterization, no MEP Imaging only, no photogrammetric mapping, no lens correction

61 Does camera choice effect accuracy? Canon S MP Horizontal RMSE = 6.4 cm Vertical RMSE = 14.0 cm Sony NEX-5 16 MP Horizontal RMSE = 1.3 cm Vertical RMSE = 1.9 cm

62 Sensor Size vs. image noise Camera Sensor Dimensions (mm) Rows x Columns (pixels) Pixel Area (μm 2 ) Canon S x x 3000 (12MP) 3.4 NEX-5R 23.4 x x 3264 (16MP) 28.8 The light collected is proportional to the sensor pixel area. Note that the NEX has 8 ½ times the area of the Canon this is a huge difference! Photon noise varies as the square root of the image signal, so collecting more light results in proportionally less noise, i.e. a higher signal-to-noise ratio is achieved. Higher signal-to-noise means more sensitivity to low-light situations, and broader dynamic range.

63 Image Noise Noise from Cannon S100 images Significant Noise Poor Conformance

64 Image Noise Noise from NEX-5R Images Low Noise High Conformance

65 Focal length calibration Focal length is highly correlated with vertical scale and accuracy Precise focal length cannot be established for zoom lenses, even if the zoom feature is disabled.

66 Other Issues. Consumer cameras do not have a Mid-Exposure Pulse (MEP) o Real Time Kinematic GNSS o One must know the exact correlation of each photo center to the GNSS position at time of exposure o Common practice is to use the camera flash signal to create a MEP with modifications to the electronic circuitry Consumer cameras do not offer stock fixed focus lens options o Requires special lenses (which are expensive or not available for many consumer cameras.)

67 3D LIDAR Scanning Airborne LIDAR: Manned Aircraft Wide Area Mapping khz pulse rate OR Corridor Mapping khz pulse rate

68 Image Processing

69 Photogrammetric Image Processing o Ground Control (reference points, GPS, GNSS) o Aero-Triangulation - (georeferencing all images a unified block and tying the block to ground control positions) o Stereo Compilation (3D extraction of information from the georeferenced block of imagery) o Topographic Mapping o Orthophotography o Planimetric Mapping

70 Photogrammetric Image Processing Contour/topographic map

71 Photogrammetric Image Processing Topographic Map with Planimetric Features

72 Photogrammetric Image Processing o Orthophotography Vertical Photo Mosaic DTM/DSM (3D-view)

73 Photogrammetric Image Processing Orthophotography 3D Ortho Mosaic

74 Land Survey vs. UAS Survey Example UAS Survey GNSS Survey Comments Area 1.5 km km 2 Ground control setup & measurement 1 ¼ hr --- Ground control not required for all applications Setup time 15 min 15 min (per day) Survey time 45 min 30 ½ hr (4 days) Tear-down time 15 min 15 min (per day) Data processing time 4 hrs (2.80 GHz Intel Core i7, 16 GB RAM) --- Data can be processed overnight Total time 6 hr 30 min 32 hr 30 min 5x faster than GNSS Measurement sampling Distance 3.8 cm (at 120 m flight altitude) 15 m Minimum sampling size is 2.4 cm Horizontal accuracy 2 cm 1 cm Vertical accuracy 4 cm 2 cm

75 Topographic Survey Comparison Surface model generated from UAS survey (± 300,000 measurements) Surface model generated from GNSS survey (±1,000 measurements)

76 Questions