From the Publisher: 
The accessible presentation of this book gives both a general view of the entire computer vision enterprise and also offers sufficient detail to be able to build useful applications. Users learn techniques that have proven to be useful by first-hand experience and a wide range of mathematical methods. A CD-ROM with every copy of the text contains source code for programming practice, color images, and illustrative movies. Comprehensive and up-to-date, this book includes essential topics that either reflect practical significance or are of theoretical importance. Topics are discussed in substantial and increasing depth. Application surveys describe numerous important application areas such as image based rendering and digital libraries. Many important algorithms broken down and illustrated in pseudo code. Appropriate for use by engineers as a comprehensive reference to the computer vision enterprise.

Computer vision : a modern approach = 计算机视觉 : 一种现代的方法

A scheme is developed for classifying the types of motion perceived by a humanlike robot. It is assumed that the robot receives visual images of the scene using a perspective system model. Equations, theorems, concepts, clues, etc., relating the objects, their positions, and their motion to their images on the focal plane are presented. >

http://ieeexplore.ieee.org/iel2/815/3148/00100651.pdf

Robot vision

The Quest for Consciousness

In this paper, we present a real-time obstacle detection system for the mobility improvement for the visually impaired using a handheld Smartphone Though there are many existing assistants for the visually impaired, there is not a single one that is low cost, ultra-portable, non-intrusive and able to detect the low-height objects on the floor This paper proposes a system to detect any objects attached to the floor regardless of their height Unlike some existing systems where only histogram or edge information is used, the proposed systemcombines both cues and overcomes some limitations of existing systems The obstacles on the floor in front of the user can be reliably detected in real time using the proposed system implemented on a Smartphone The proposed system has been tested in different types of floor conditions and a field trial on five blind participants has been conducted The experimental results demonstrate its reliability in comparison to existing systems

/pdf/a-smartphone-based-obstacle-sensor-for-the-visually-impaired-vaibmkrj9h.pdf

A smartphone-based obstacle sensor for the visually impaired

Aerial photography obtained by UAVs (Unmanned Aerial Vehicles) is an emerging market for civil applications. Small UAVs are believed to close gaps in niche markets, such as acquiring airborne image data for remote sensing purposes. Small UAVs will be able to fly at low altitudes, in dangerous environments and over long periods of time. However, the small lightweight constructions of these UAVs lead to new problems, such as higher agility leading to more susceptibility to turbulence and limitations in space and payload for sensor systems. This research investigates the use of low-cost fisheye lenses to overcome such problems which theoretically makes the airborne imaging less sensitive to turbulence. The fisheye lens has the benet of a large observation area (large field of view) and doesn't add additional weight to the aircraft, like traditional mechanical stabilizing systems. This research presents the implementation of a fisheye lens for aerial photography and mapping purposes, including theoretical background of fisheye lenses. Based on the unique feature of the distortion being a function of the viewing angle, methods used to derive the fisheye lens distortion are presented. The lens distortion is used to rectify the fisheye images before these images can be used in aerial photography. A detailed investigation into the inner orientation of the camera and inertial sensor is given, as well as the registration of airborne collected images. It was found that the attitude estimation is critical towards accurate mapping using low quality sensors. A loosely coupled EKF filter applied to the GPS and inertial sensor data estimated the attitude to an accuracy of 3-5° (1-sigma) using low-cost sensors typically found in small UAVs. However, the use of image stitching techniques may improve the outcome. On the other hand, lens distortion caused by the fisheye lens can be addressed by rectification techniques and removed to a sub-pixel level. Results of the process present image sequences gathered from a piloted aircraft demonstrating the achieved performance and potential applications towards UAVs. Further, an unforeseen issue with a vibrating part in the lens lead to the need for vibration compensation. The vibration could be estimated to ±1 pixel in 75% of the cases by applying an extended Hough transform to the fisheye images.

Investigation of fisheye lenses for small UAV aerial photography

One of the most important abilities for a mobile robot is detecting obstacles in order to avoid collisions. Building a map of these obstacles is the next logical step. Most robots to date have used sensors such as passive or active infrared, sonar or laser range finders to locate obstacles in their path. In contrast, this work uses a single colour camera as the only sensor, and consequently the robot must obtain range information from the camera images. We propose simple methods for determining the range to the nearest obstacle in any direction in the robot’s field of view, referred to as the Radial Obstacle Profile. The ROP can then be used to determine the amount of rotation between two successive images, which is important for constructing a 360o view of the surrounding environment as part of map construction.

Monocular Vision as a Range Sensor

For robots to operate in human environments they must be able to make their own maps because it is unrealistic to expect a user to enter a map into the robot’s memory; existing floorplans are often incorrect; and human environments tend to change. Traditionally robots have used sonar, infra-red or laser range finders to perform the mapping task. Digital cameras have become very cheap in recent years and they have opened up new possibilities as a sensor for robot perception. Any robot that must interact with humans can reasonably be expected to have a camera for tasks such as face recognition, so it makes sense to also use the camera for navigation. Cameras have advantages over other sensors such as colour information (not available with any other sensor), better immunity to noise (compared to sonar), and not being restricted to operating in a plane (like laser range finders). However, there are disadvantages too, with the principal one being the effect of perspective. This research investigated ways to use a single colour camera as a range sensor to guide an autonomous robot and allow it to build a map of its environment, a process referred to as Simultaneous Localization and Mapping (SLAM). An experimental system was built using a robot controlled via a wireless network connection. Using the on-board camera as the only sensor, the robot successfully explored and mapped indoor office environments. The quality of the resulting maps is comparable to those that have been reported in the literature for sonar or infra-red sensors. Although the maps are not as accurate as ones created with a laser range finder, the solution using a camera is significantly cheaper and is more appropriate for toys and early domestic robots.

Mapping of indoor environments by robots using low-cost vision sensors

Mapping algorithms commonly use "radial sweeps" of the surrounding environment as input. Producing a sweep is a challenging task for a robot using only vision. With no odometers to measure turn angles, a vision-based robot must have another method to verify rotations. In this paper we propose using the radial obstacle profile (ROP), which gives the radial distance to the nearest obstacle in any direction in the robot's field of view. By matching the ROPs before and after a turn, the robot should be able to verify that the expected angle of rotation matches the actual angle. Combining successive ROPs then produces a radial sweep.

/pdf/vision-based-pirouettes-using-radial-obstacle-profile-42man5z6d2.pdf

Vision-based pirouettes using radial obstacle profile

Two key objectives of robot vision are autonomous navigation and mapping. Digital cameras have become relatively cheap in recent years and have appeared in a variety of consumer devices, such as PDAs (Personal Digital Assistants). It therefore makes sense to try to build systems that use vision as their primary input instead of the more traditional sonar and infrared sensors that have been used in the past. The camera can also be used for a range of other tasks. Service robots and toys typically operate in an indoor environment but rarely have a map of their environment when they are first turned on. This paper therefore addresses the problem of vision-based mapping where a camera has deliberately been chosen as the only sensor.

/pdf/early-results-in-vision-based-map-building-2t92oe4cee.pdf

Early Results in Vision-based Map Building

Mobile robots operating in unknown environments need to build maps. To do so they must have an exploration algorithm to plan a path. This algorithm should guarantee that the whole of the environment, or at least some designated area, will be mapped. The path should also be optimal in some sense and not simply a "random walk" which is clearly inefficient. When multiple robots are involved, the algorithm also needs to take advantage of the fact that the robots can share the task. In this paper we discuss a modification to the well-known distance transform that satisfies these requirements.

/pdf/directed-exploration-using-a-modified-distance-transform-1z3h4c66il.pdf

Trevor Taylor

Papers

Monocular Vision as a Range Sensor

Mapping of indoor environments by robots using low-cost vision sensors

Vision-based pirouettes using radial obstacle profile

Early Results in Vision-based Map Building

Directed Exploration Using a Modified Distance Transform