Human Vision for Landscape Assessment Dr. Yuhan Shao | Prof. Eckart - - PowerPoint PPT Presentation

Improved Photographic Representation of Human Vision for Landscape Assessment

DLA 2015 | Dr. Yuhan Shao | The University of Sheffield

• Dr. Yuhan Shao | Prof. Eckart Lange | Dr. Kevin Thwaites

Purpose of the Study

• The main objective of this study is to develop a photo based interview process

that offers an improved representation of real human vision.

• The process is composed of two main steps:
• the first step is to produce panoramic photos that replicate real horizontal as well as

vertical human vision.

• These panoramic photos are then used in the second step to interview participants in
• rder to identify local characteristic elements of the study site in Yantai, China.
• The results revealed that an interview using 180˚ panoramic photos offers improved

representation of human vision and helps to pinpoint and understand people’s perceptions of their surrounding landscape.

DLA 2015 | Dr. Yuhan Shao | The University of Sheffield

Problem

• Using images in landscape study to assess the quality of natural landscape (Linton,

1968).

• The use of photographs as proxies to simulate landscape has been further developed

by various researchers since the 1970s (Zube et al., 1974, Dunn, 1976, Shafer and Brush, 1977, Nassauer, 1983, Zube et al., 1987) and there have also been extensive empirical studies conducted using photo-realistic visualisation to represent the real environment in recent years e.g. (Daniel and Meitner, 2001, Lange, 2001, Williams et al., 2007).

• In landscape research, photos or images have been extensively used to gather

people’s opinions on certain landscapes or to assess and compare between different landscapes.

• However, there are gaps in terms of what the photos represent and what we

truly see. And human vision is a more dynamic process while the photo only represents a point in time with limited ability to change.

DLA 2015 | Dr. Yuhan Shao | The University of Sheffield

Available Technology

• To shorten the gaps between real human vision and the landscape representation.

Computer based landscape modeling techniques have been introduced as a tool for landscape research in the last decade. Eg. 3DMax, Sketch Up and GIS, etc.

• Using

3D visualisation techniques to gather sufficient information from participants’ interviews.

• But these techniques focus more on modeling than replicating the real

experience; therefore a convenient and cost effective technique for showing all visual landscape is still missing (Lange, 2001).

• Among all technologies available, consider time limitation and cost limitation and

convenient issue, the current most efficient way is shooting peripheral dimension photos use modern cameras (Danahy, 2001 Maclennan and Envision, 2006).

DLA 2015 | Dr. Yuhan Shao | The University of Sheffield

Limitation of Photo Representation

• Vision is composed of both foveal & peripheral visions. It is a dynamic process when

locating the surrounding environment. (Hilgard et al., 1975)

• Foveal vision: focus on only one object & ignore everything else around it. (Antrop,

2007);

• Binocular vision: is vision in which creatures having two eyes use together (Fahle,

1987).

• Peripheral vision (FOV): binocular vision plus eyeball movement which mean takes

in everything around us. (Ernst et al., 2000).

• Many researches suggested that peripheral vision improves our ability to make

in-depth judgments (Watanabe and Matsuoka, 1999).

DLA 2015 | Dr. Yuhan Shao | The University of Sheffield

Limitation of Photo Representation

• Gibson, 1979 and Barfield et al. 1995 indicated the vertical human FOV is

approximately 135˚.

• Standard modern digital camera would have a smaller vertical angle (60˚).
• Ulrich (1981) pointed out that when shooting distance is limited, panoramic photos

would have the limitation of capturing the vertical view.

DLA 2015 | Dr. Yuhan Shao | The University of Sheffield

To represent the dynamic and peripheral vision of the human eye and focusing with foveal vision on objects of attention is the main obstacle in this study. Based

• n this point a panoramic photo is suit for the study.
• Zube et al. (1974) suggest panoramic wide-angle photo is valid landscape

simulation media.

• Sevenant and Antrop (2011) support panoramic photo is efficient for

representing landscape. The obstacle between human vision and photos was discussed from both horizontal and vertical angle view. Therefore two issues are addressed in this research:

• 1. Which horizontal angle should better simulate people’s peripheral vision?
• 2. How can panoramic photos be combined to represent people’s vertical

viewing angle?

Research Question

DLA 2015 | Dr. Yuhan Shao | The University of Sheffield

Solution

Case studies were carried out to compare the results between different panoramic photo sets on both horizontal and vertical representation.

• Based on the literature review, researcher chose 3 different degrees to testify:
• 120˚ binocular vision (Zube et al., 1974)
• 180˚ FOV Daniel and (Boster, 1976)
• 360˚ (Englar and Consolati, 1990)

The photos were taken along the center of the street using a Nikon Coolpix P510 Digital Camera with zoom range from 24mm to 1000mm (35mm equivalent focal lengths). The camera was attached to a tripod at eye level. It has been suggested that between 12 to 20 participants for one interview group is an appropriate size (Guest et al., 2006, Creswell, 1999, Sandelowski, 1995). Hence a sample size of 15 was chosen for each photo set. So that there are total 150 participants, for different angle sets from both horizontal and vertical angle and between onsite and offsite groups.

DLA 2015 | Dr. Yuhan Shao | The University of Sheffield

Horizontal Angle

The historical boundary wall and other historic buildings were most identified. For these historical memories deeply embedded and being viewed as the “symbol of the site”. Plants and unique shaped streetlights were also selected due to their design characteristics.

• 120˚ V.S. 180˚ V.S. 360˚

DLA 2015 | Dr. Yuhan Shao | The University of Sheffield

Vertical Combination

Nassauer (1983) pointed out: when the shooting distance is long enough, one panoramic photo would be efficient to capture the field to provide accurate field experience, e.g. at the coast and towards the sea. When there are considerable amount of high elements in the sight, it would be the best to take two 180˚ panoramic photos,

• ne

capture the bottom part of the sight, the other capture the top part, and then combine them together to reproduce the entire sight on photo.

DLA 2015 | Dr. Yuhan Shao | The University of Sheffield

Limitation

The vertical shooting angle of the camera to mimic human vision for the improved photo based interview process remains a major limitation.

• In a situation where the building’s height is much higher than the shooting

range the entire angle cannot be captured.

• For example, in New York Times Square, where the shooting distance is very

limited and the buildings are extremely high, the vertical combination technique would not work. However, there is no technique available that could tackle such limitation effectively.

DLA 2015 | Dr. Yuhan Shao | The University of Sheffield

Thank You

DLA 2015 | Dr. Yuhan Shao | The University of Sheffield

• Dr. Yuhan Shao | shaoyuhan7@hotmail.com