Applying distance histograms for robust object recognition

Abstract

Purpose -One of the main goals of vision systems is to recognize objects in real world to perform appropriate actions. This implies the ability of handling objects and, moreover, to know the relations between these objects and their environment in what we call scenes. Most of the time, navigation in unknown environments is difficult due to a lack of easily identifiable landmarks. Hence, in this work, some geometric features to identify objects are considered. Firstly, a Markov random field segmentation approach is implemented. Then, the key factor for the recognition is the calculation of the so-called distance histograms, which relate the distances between the border points to the mass center for each object in a scene. Design/methodology/approach -This work, first discusses the features to be analyzed in order to create a reliable database for a proper recognition of the objects in a scene. Then, a robust classification system is designed and finally some experiments are completed to show that the recognition system can be utilized in a real-world operation. Findings -The results of the experiments show that including this distance information improves significantly the final classification process. Originality/value -This paper describes an object recognition scheme, where a set of histograms is included to the features vector. As is shown, the incorporation of this feature improves the robustness of the system and the recognition rate.

Applying triangular fuzzy number for multi‐sensor object recognition

“…Compared with the existing methods, the proposed method in this paper, which has no demand for the basic probability assignment, membership function, distribution type, and the prior probability (Cameron and Durrant‐Whyte, 1992; Shao et al , 1996), determines objectivel the attribute weights by optimization programming model, while the attribute weights are artificially given in Chen and Hu (2006) and Shao et al (1996). Arques et al (2007) did not consider the different weights of different characteristic indexes. In addition, the data of the characteristic value and observations of sensors are all exact real numbers in existing literature, whereas this paper deals with the triangular fuzzy number, which is the most difference between the existing literature and this paper.…”

“…There are many different fusion methods developed in these years, e.g. methods based on Shafer‐Dempster evidence theory (Begler, 1987; Yan et al , 2007; Chen and Que, 2006), methods based on fuzzy theory (Odeberg, 1989, 1993; Chen and Hu, 2006; Chanussot et al , 1999; Glossas and Aspragathos, 2001; Grimberg and Savin, 2000), Bayesian approach (Cameron and Durrant‐Whyte, 1992), maximum likelihood (Shao et al , 1996), binary labeling (Windeatt and Ghaderi, 2001), dual hierarchical models (Saitou et al , 2007), multiplex complex amplitude (Yoshikawa and Ii, 2007), distance histograms (Arques et al , 2007), etc. Though when the sensor readings may be described by real numbers, these fusion methods all work well, methods based on Shafer‐Dempster evidence and fuzzy theories excessively depend on the selection of basic probability assignment and the membership function.…”