Abstract | ||
---|---|---|
In the production line, it is difficult to classify the tiles because there are a large number of tile types with a very similar appearance and the tiles will be rotated in the production process. Texture is regarded as an important feature of tiles. Hu invariant moments are features with invariant to rotation of objects. Inspired by processing mechanism in the biological visual system, a multiple visual pathway model is proposed to fuse multifeatures from colour channels and Hu invariant moments. The colour image is transformed to HSI colour space. The textures in three channels of HSI are analyzed. It is found that the texture of I channel is better than other channels. In the tile recognition process, the problem of rotation cannot be solved by texture features. The Hu invariant moments are used to solve the problem of rotation. The I co-occurrence matrix and the Hu invariant moments are combined in series using a multi-channel model. The fused features are used to train SVM to identify tiles. In this paper, the recognition rate of the fusion feature method is 20% higher than that of the methods based on color or texture features. The recognition rate of the proposed method has reached 94.18%. |
Year | DOI | Venue |
---|---|---|
2017 | 10.1109/CISP-BMEI.2017.8301970 | 2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics (CISP-BMEI) |
Keywords | Field | DocType |
texture features,HSI,I co-occurrence matrix,Hu invariant moments,fusion feature,SVM | Computer vision,Pattern recognition,Visualization,Computer science,Matrix (mathematics),Support vector machine,Fusion,Feature extraction,Artificial intelligence,Invariant (mathematics),Tile,Brightness | Conference |
ISBN | Citations | PageRank |
978-1-5386-1938-4 | 0 | 0.34 |
References | Authors | |
2 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Yanan Zhang | 1 | 9 | 6.92 |
QingXiang Wu | 2 | 44 | 12.42 |
Yao Xiao | 3 | 13 | 7.62 |
Caiyun Wu | 4 | 16 | 8.46 |
Peng-Fei Li | 5 | 56 | 20.94 |
Caiyou Yuan | 6 | 1 | 0.69 |