Model-based objact recognition has attracted considerable attention in the vision community over the last twenty years. This paper prasents a model -based two dimensional object recognition system using feature indexing under a linear transformation through geometric hashing. The objects are represented by their dominant invariant feature under a linear transformation in the normalised orthogonal coordinate frame. The goal of this paper is to solve the 2D recognition in the industrial enviornment in an efficient awy and reduce the net computational overhead. The proposed algorithm is also able to identify a partially occluded object in cluttered enviornment. The paper shows how the symmetry of the object can be exploited to reduce the model data storage to the speed up the matching process. The matching is done through its feature set tuples. The feature set includes both the local and global features of the objects, which helps to recognise the object more accurately.