利索-全球专利检索

  1. 登录
  2. 注册
发明公开
CN106778511A 一种马尾松多尺度智能识别的方法及装置 无效 - 驳回

一种马尾松多尺度智能识别的方法及装置
摘要：
本发明实施例公开了一种马尾松多尺度智能识别的方法及装置。该方法包括：基于频域的谱残余方法分割包含马尾松的显著性区域；获取预设气候条件下和预设植株大小条件下的马尾松的图片，对所述马尾松进行形态建模；在显著性区域分割的基础上，根据所述形态模型对所述马尾松进行目标预测；基于马尾松的光谱统计特性，对目标预测的结果进行处理，剔除虚警，获取最终的马尾松识别结果。从而使得该模型可以识别出不同植株大小等多种条件下的马尾松，鲁棒性强，通用性好；该模型可以在不同尺度条件下实现马尾松的智能识别；该模型可以实现马尾松的自动识别。
摘要（英）：

The embodiment of the invention discloses a multi-scale intelligent identification method and device for pinus massoniana. The method comprises the steps that a significant region containing the pinus massoniana is partitioned based on a frequency domain spectrum remnant method; a picture of the pinus massoniana under preset climatic conditions and preset plant size conditions is acquired, and morphological modeling is performed on the pinus massoniana; on the basis of partitioning of the significant region, target prediction is performed on the pinus massoniana according to the morphological model; based on spectrum statistical properties of the pinus massoniana, the target prediction result is processed, a false alarm is removed, and a final pinus massoniana identification result is acquired. Therefore, intelligent identification of the pinus massoniana can be realized under different scale conditions through the model, and the model can be used for realizing automatic identification of the pinus massoniana.

0/0