Gabor小波理论的植物根系图像边缘检测研究

摘要

The images of plant root that extracted from the study of plant root usually accompany with big noises, and are affected by variation of the light intensity. Because Gabor wavelet is able to filter in different direction and in multiple scales, and determin the characteristics of the grain edge under optimal time and frequency significance, thus it is proposed that using Gabor wavelet for detecting image edge of the plant root. The result of Matlab indicates that the contour extracted by this method is closer to real contour image, so the effectiveness and feasibility of this method are verified.%植物根系中提取出的植物根系图像具有噪声较大、受光照强度变化影响大等特点,利用Gabor小波能够多方向、多尺度滤波及确定时频最佳意义下纹理的边缘特性,提出了运用Gabor小波检测植物根系图像边缘的方法.Matlab仿真结果表明,采用该方法提取的轮廓更接近于真实轮廓图像,验证了该方法的有效性和可行性.