摘要
提出一种基于稀疏表征特征的果蔬识别算法,首先采用图像获取装置获取果蔬图片,对果蔬图片进行预处理,得到归一化果蔬图片,提取果蔬图片的5尺度8方向的Gabor小波特征,采用主频量分析算法对其进行降维处理,构建稀疏表征识别分类模型,并对提取到的特征进行分类识别,最终得到识别结果。仿真证明文章研究的算法分类正确率达96%左右,误识率低,运算速度快,更适合用来对果蔬进行分类识别,具有较强的实用性。
In order to further improve the recognition rate of fruits and vegetables products, a classification algorithm based on Sparse Representation Classification（SRC） is presented in this paper. Firstly, the fruits and vegetables images are acquired through image acquisition device, and then those images are preprocessed to be normalized images. After that, through 5 scale and 8 direction Gabor convolution and PCA preproeessing, classification feature is extracted from images. At last, SRC model is used to classify fruits and vegetables products. Simulation results show that our method with 96% of recognition rate and low speed is more suitable for fruits and vegetables products classification.
出处
《信息化研究》
2013年第6期15-18,共4页
INFORMATIZATION RESEARCH
基金
国家自然科学基金项目(No.61231002
No.51075068)
关键词
果蔬识别
稀疏表征
主成分分析(PCA)
fruits and vegetable classification
sparse representation-based classification
principle component analysis
