知搜-全球专利检索

  1. 登录
  2. 注册
发明公开
CN110147757A 基于卷积神经网络的输电通道工程车辆辨识方法及系统 无效 - 驳回

基于卷积神经网络的输电通道工程车辆辨识方法及系统
摘要：
本公开提供了一种基于卷积神经网络的输电通道工程车辆辨识方法及系统。其中，基于卷积神经网络的输电通道工程车辆辨识方法，包括获取输电线路监控图像并进行统一像素大小处理；将处理后的输电线路监控图像输入至基于卷积神经网络的工程车辆辨识模型，由工程车辆辨识模型辨识出图像类别；其中，图像类别以图像中包含的工程车辆类别进行分类；将图像类别结果映射到原始输电线路监控图像中，若图像类别结果为含有工程车辆，则工程车辆辨识模型输出标注工程机械边界框位置与种类标签的输电线路监控图像，继续下一张图像的检测；否则，跳过这张图像，继续下一张图像的检测。
摘要（英）：

The invention provides a power transmission channel engineering vehicle identification method and system based on a convolutional neural network. The power transmission channel engineering vehicle identification method based on the convolutional neural network comprises the steps of obtaining a power transmission line monitoring image and carrying out unified pixel size processing; inputting the processed power transmission line monitoring image into an engineering vehicle identification model based on a convolutional neural network, and identifying an image category by the engineering vehicleidentification model, wherein the image categories are classified according to engineering vehicle categories contained in the image; mapping the image category result into an original power transmission line monitoring image; if the image category result is that an engineering vehicle is contained, enabling the engineering vehicle identification model to output a power transmission line monitoring image marked with an engineering machinery boundary frame position and a category label, and continuing detection of a next image; otherwise, skipping the image, and continuing the detection of thenext image.

0/0