摘要
为了提高钢筋的计数准确率和效率,综合运用图像处理技术和神经网络技术,实现对钢筋的识别和计数。对获取的钢筋原始图像进行数字图像处理,得到感兴趣的部分即钢筋的轮廓;计算单根钢筋轮廓的宽度、高度、面积和打捆钢筋的总面积4个特征量;将这4个特征量作为神经网络的输入,训练网络识别钢筋并计数。仿真实验验证了这种方法的可行性和有效性。
In order to improve the counting accuracy and efficiency, the image processing techniques and neural network technology are integrated used to realize the identification and counting of the steel bar. The obtained original image of the steel bar is image processed digitally to obtain the interested part which is the the outline of reinforcement, calculate the width, height, area of a single steel bar contour and the total area of four bundled steel bar. These four characteristic quantities act as neural network input, the network is trained to recognize and count reinforcement. Simulation results verify the feasibility and effectiveness of this method.
出处
《传感器与微系统》
CSCD
北大核心
2010年第8期44-47,共4页
Transducer and Microsystem Technologies
关键词
图象处理
特征提取
神经网络
模式识别
钢筋计数
image processing
feature extraction
neural network
pattern recognition
reinforcement count