论文部分内容阅读
介绍了基于机器视觉的储粮害虫智能检测系统软件部分各主要环节的具体实现。该系统运用图像差分法及自适应图像增强法提高粮虫样本图像的质量 ,利用改进的直方图阈值将粮虫从背景中分割开来 ,并运用数学形态学处理法进行了滤波。以提取出的粮虫面积、周长、复杂度为特征 ,运用基于模糊决策的分类器对粮仓中常见的 9种、7类害虫进行了分类 ,识别正确率达到 95 .2 %。
This paper introduces the concrete realization of the main parts of the software part of intelligent detection system of stored grain pests based on machine vision. The system uses the image difference method and the adaptive image enhancement method to improve the quality of the image of the grain insects, divides the grain insects from the background by using the improved histogram threshold, and applies the mathematical morphological processing method to carry out the filtering. Based on the area, the circumference and the complexity of the extracted grain insects, 9 kinds and 7 kinds of pests commonly found in grain silos were classified by using the fuzzy decision-making classifier. The recognition rate was 95.2%.