基于渐进式分割的蔬菜病虫害识别仿真研究

Simulation Research on Identification of Vegetable Pests and Diseases Based on Progressive Segmentation

针对当前蔬菜病虫害相关成果存在的识别准确率较低等问题,提出基于渐进式分割的蔬菜病虫害识别方法。首先采集蔬菜图像,利用灰度直方图均衡化法增强图像细节,提高图像清晰度。对增强后图像进行二值化处理,实现病斑部位的分离,同时进行边缘检测与细线化操作,提取病斑轮廓。引入渐进式分割法,利用渐近分割方法选择图像区域,通过局部优化完成匹配,直至获取最终的分割结果,实现蔬菜病虫害识别。通过实验证明,所提方法图像处理效果良好,且具有较高的识别准确率。

Due to the low accuracy of discrimination of vegetable pests and diseases,this paper presented a method to discriminate vegetable pests and diseases based on progressive segmentation.Firstly,the vegetable image was collected,and the gray-scale histogram equalization method was used to enhance the image detail and improve the image clarity.Then,the enhanced image was binarized to realize the separation of diseased spots.Meanwhile,the diseased spot contour was extracted through edge detection and thinning operation.The progressive segmentation method was adopted to select the regions in an image.Finally,the matching was completed by local optimization until the segmentation result was obtained.Thus,the discrimination of vegetable diseases and pests was achieved.Experimental results show that the proposed method has a good image processing effect and high recognition accuracy.