Papers by Keyword: Shape Distribution

Abstract: Efficacy of existing designs grinding wheels is relatively low and one of the main reasons for this situation is the lack of control over the shape of grains included in the composition of these tools and their employees cutting elements. Standard wheels are composed of grains with an arbitrary, varying in the range from isometric to the needle-like species. When only part of the grain has a favorable geometry for cutting and participates fully in aggregate nom cutting process. Arranging grain shape and as a consequence, their geometry can increase the efficiency of each individual grain and thus improve the performance of grinding wheels as a whole. The study on the establishment of the variety of forms of grains for abrasives Russian and German production made standard way - by ebb abrasive ingots, their crushing and screening a number of particle size fractions. Studies have shown that each has its abrasive distribution pattern in the manner that depends on the brand of abrasive characteristics of technology and the manufacture of abrasive grit. The interrelation of grain shape to their geometry as derivation of mathematical models. A pilot batch of grinding wheels with controlled grain shape, are prepared by separating the initial mass of abrasive on a number of fractions with the same shape of grains using the vibratory separator. Experimental wheels were tested on the operating processing bearing rings, and as compared with the following instruments used an arbitrary shape of grains. Found that a differentiated approach to the choice of grain shape allows for a increase resistance grinding wheels in 1,5-3,9 times , reduce roughness of machined surfaces in 1.2 - 3.2 times, roundness - 1.3 times, waviness - 2.3 times the wear sized - 2.0 - 3.0 times.

Abstract: Shape Distribution (SD) describes the global feature of 3D CAD model which can be used to distinguish one model from others. But sometimes, different models may have the similar SD features. In this case SD is not suitable for 3D CAD model retrieval. Although two models have the same SD features, their oriented bounding box (OBB) and normals distribution may differ from each other. In this paper, a new algorithm based on OBB and normals distribution is proposed to tackle this problem. SD feature is used for initial retrieval, then OBB and normals distribution are used for fine retrieval. Experimental results indicate the efficiency and feasibility of the proposed method.

Abstract: Existing 3D model retrieval techniques focus on the global feature more than local detailed feature are not applicable to 3D CAD models. This paper combines the local feature and global feature to satisfy the needs of mechanical design. We first analyze the limitations of Shape Description (SD) algorithm. Then propose an automatic feature extraction technology based on the local characteristics of the curvature distribution. Next we fuse the extracted feature using a novel method. Lastly, we design an improved K-nearest neighbor algorithm to retrieve models. Experimental results indicate the efficiency and feasibility of the proposed method.

Abstract: Shape distribution is considered as a kind of art state 3D model retrieval algorithm due to its simplicity, robustness, and not need model pretreatment. Its disadvantage is that the retrieval precision is not high enough. Despite the introduction of five kinds of shape functions, each of which can not sufficiently expresse the 3D models. This paper uses the D2 distribution, the improved D1 distribution and the total surface area method to retrieve 3D models respectively. Then give weights to each method after unitary them. Set fuzzy rules are set to decide the fuzzy weights of three methods according to the mean proportion and variance proportion, Experimental results show that this algorithm can improve the retrieval results significantly.

Abstract: To accelerate the development of new product, one important method is to use the information of the existed products. In this paper, Fuzzy C-Means clustering method (FCM) is proposed to retrieve the desired model in the given model database. In the retrieval process, first, the shape distribution histograms of the model are established and transformed into the proper format of FCM. Then, the shape distribution histograms of the model are served as inputs of FCM and classified into several groups by this algorithm. Last, all the models that belong to the classification of the query model are returned as the retrieval results. To validate the proposed method, a case study is presented. The results show that the result of proposed method is better than that of Shape Distribution. Thus, this method is suitable for 3D mechanical models retrieval.

Abstract: 3D models play an important role in many applications, so there is an urgent need for an effective content based 3D model retrieval system. A variety of 3D model retrieval methods have been proposed in recent years. Shape distributions show superiority over other methods due to ease of computation and invariance to Euclidean motion, but there is poor retrieval performance for the loss in information. This paper introduces two model-partitioning methods to improve shape distributions, in which the two enhanced descriptors are combined with a fuzzy feedback method. Experimental results show that the proposed methods can achieve better retrieval performance.

Abstract: Shape Distribution (3DSD) and Radius Angle Histogram (RAH) are useful methods for retrieving 3D model in mechanical engineering. Through these methods have advantages such as fast speeds and simple operations, the retrieval precision are not very high enough. To improve the retrieval precision, a new method named combined histograms which integrates the advantages of 3DSD and RAH is proposed. This method makes use of the information both of shape and surface of the models to be retrieved. In the retrieval process, the shape histogram and the radius angle histogram of the retrieved model are first extracted. Then, the combined histograms of the model are established by integrating the shape histogram and the radius angle histogram. To validate the proposed method, an experiment is given. The experiment results show that the proposed method has higher retrieval precision than that of 3DSD and RAH and is suitable for mechanical model design.