BP神经网络在香蕉检测中的应用

Application of BP neural networks in the detection of bananas

为了解决绿色通道验货部门存在的人工方法难以判断车辆中是否夹带违禁物的问题,首先运用梯度法对香蕉的回波图像进行特征提取,然后结合BP神经网络处理非线性问题的优势,提出了将BP网络应用于香蕉的雷达回波图像识别和分类的方法.通过在河南高速安阳服务区绿色通道验货部门的实际应用,表明该方法有较好的识别和分类效果.

In order to solve the problem that it is difficult to use manual methods to detect whether there is prohibited goods in the vehicle,this paper puts forward a discernment and classification method that neural network is applied to study the discernment and classification of the banana radar echo.First,gradient was used to extract feature from collected radar echo image of banana,neural network was used to deal with non-linear problem.Practical application in the Anyang Green Channel inspection department of Henan freeway shows this method has a good discernment and classification performance.