【摘要】 本文提出了基于广义霍夫变换(GHT)的一种改进。通过将参考点设在形状边界上,新方法与传统方法比较,霍夫变换可以减少内存的需要。与此同时,用于寻找峰值的速度也大大提高了。从理论上说,改进后的算法对内存的需要是一个基于形状描述复杂度的函数。越是精确和高级的形状和特征描述,意味着更大的内存空间减少。最后,将改进的GHT用于物体形状识别,取得一些实验性效果。更多还原

【Abstract】 This paper proposes an improvement of Generalized Hough Transformation. Compared with the traditional method, the new method can reduce requirement of storage by setting reference point on shape edge. Meanwhile, the speed of search for peak is accelerated conspicuously. Theoretically, the requirement of storage is a function based on the complexity of the shape description in the new method. The more accurate and higher level the shape is described, the smaller the storage space is needed. In the end, the improved algorithm is put into the object recognition, and obtains some experimental result.更多还原