Daobin Zhang

Generalized predictive control of a vision-based tracking system using kalman filtering technique

By lack of measurements of the tool position in space, the control loop of a robot does not close around the end effector. Computer vision provides a means for closed-loop control of the robot tool position but also introduces significant time delay and measurement noise. A state estimator and a dynamic visual controller are then necessary for fast response and desired performance. The Generalized Predictive Control (GPC) algorithm is considered in this paper for a vision-guided robot tracking system with the help of the Kalman filter and robust control has been achieved.

Multispectral edge detection in remote sensing images

The research work reported in this paper deals with edge information derived from the individual bands of multispectral satellite images. The combination of this edge information in a meaningful way should lead to an edge image that is more complete than those extracted from any single band. For the extraction of edges from an individual band, three commonly used approximations of the first derivative of a function f(x,y) were tested. Gradient based edge detectors were selected because methods based on vector algebra can be considered to combine them. Furthermore single pixel wide edges can be obtained using both the magnitude and orientation information. As a first method to combine edges from individual planes the algebraic vector sum is evaluated. A second method consists of computing new edge components from the rms values of the corresponding single band components. A third way to combine edges is to take as magnitude the largest magnitude found in any of the bands. The corresponding orientation value is used as the multispectral edge orientation. The last method implements the only correct formulation of the multispectral gradient according to Di Zenzo which is based on the tensor gradient associated with a vector field. As shown in the paper, the last two methods produce acceptable and comparable edge combinations.

Classification of remote sensing images with the aid of Gibbs distribution

In a maximum likelihood classification (MLC) of a satellite image it is explicitly assumed that the spectral properties of one pixel are independent of the properties of all other pixels. As a result, the MLC is unable to distinguish the pixels which come from different land-cover classes but have the same spectral properties and the result is usually a snow-like map. On the other hand, data in the field of remote sensing often appear in the form of distinct parcels and all the pixels in a specific parcel are assumed to come from a single land-cover class. Therefore, there must exist some spatial continuity between adjacent pixels. This property must be of great importance and should be taken into account in the process of land-cover classification. This paper proposes to make use of the theory of the Markov random field (MRF) and the Gibbs distribution for imposing the spatial continuity either by making use of the joint distribution of the Gibbs distribution and the conventional multinormal distribution or by using the Gibbs distribution separately as a postprocessing procedure to the MLC. While the application of the joint distribution and the Gibbs distribution may result in different classifications, experiments show that very significant improvement can be achieved with at least one of these models.