Font Size: a A A

Study And Process Optimization On Flat-granular Object With Real-time Separating Characteristic

Posted on:2008-03-06Degree:MasterType:Thesis
Country:ChinaCandidate:Y ZhaoFull Text:PDF
GTID:2178360218450489Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
Nowadays, more and more industry manufactories have adopted computer technology to assistant production, and improve their efficiency obviously. In the process of computerizing, separating problem has been one of the most difficult application problems. Thus, this paper aims at actualizing an automatic system which can separate watermelon seeds. Discuss emphatically about introduction and design of target system, optimization of identifies software and research of the multi-connectivity algorithm. Finally, actualize an effective, accurate, steady and flexibly identify prototype system. The primary research and implementation in the paper include following points:1.As watermelon seeds considered, we avoid extra identification computation of the identify system through analysis non-inferior solution (locally best) of flat-granular object's feature space;2.Optimize the process of identification, found the inner relationship between dealing sequence and average efficiency through the statistical data. After the selection of feature sets, the system can optimize the dealing sequence and choose the best to decrease the average time of identification;3.Based on the selection of identify approach and characters, propose the frame of the software to make sure it can satisfy the request of real-time separating;4.Make deeply research on the multi-connectivity algorithm and propose some improvement;5.Introduce some optimize techniques to make the system more efficient and steady.The improved multi-connectivity algorithm can advance the efficiency of the abstraction of pictures'features and can be easily applied to other similar system. The capability and the efficiency of our system properly satisfy the practice.
Keywords/Search Tags:Particle objects, Image identification, Feature space, Process optimization, Software design, Multi-connectivity
PDF Full Text Request
Related items