基于改进算法的储粮害虫图像识别预处理

摘要

引入图像预处理及模式识别技术，实现储粮害虫图像的预处理。结合害虫图像的特征，对传统的灰度化方法进行改进，以HSI的变换方法对图像进行灰度化处理；通过直方图均衡化调整图像的灰度间距，提升图像的对比度；对传统的滤波方法进行优化，以方向滤波算法保护图像边缘，滤除图像噪声；以FCM分割算法进行图像分割，获得害虫图像的主要特征，从而实现了对目标图像的去噪、增强，为储粮害虫进一步的智能化识别与处理打下基础。%The image pretreatment of pests of grain stored insect was realized by introducing image preprocessing and pattern recognition technology. Combining with the feature of insect images, the traditional gray processing was improved to transform method of HSI to gray scale image processing. The image gray level spacing was adjusted through histogram equalization to enhance the contrast grade of image. The traditional filtering methods was optimized to direction filtering algorithm protecting the image edge. The image noise was filtered. With FCM segmentation algorithm for image segmentation, the main characteris-tics of the pest images were acquired. This, the target image denoising and enhancement was realized. A good foundation for intellectualization and processing of pests in further recognition was laid.