论文部分内容阅读
采用数字图像处理技术实现了对玉米种子表面裂纹的识别和检测。选择冷阴极荧光灯(CCFL)设计了图像采集的光照环境,建立了玉米种子图像的采集系统,然后针对玉米种子图像提出了一种基于籽粒形态学特征的表面裂纹检测方法。该方法采用水平和垂直边缘检测算子处理得到裂纹、种子边界和噪声等边缘信息;然后通过玉米籽粒的形态特征寻找其尖端位置,并使用图像代数运算的方法去除大部分非裂纹信息;最后根据裂纹的长度和位置特征提取得到裂纹,并计算裂纹的绝对长度和相对长度。对农大4967和农大3138两个品种的玉米分别选取裂纹粒和无裂纹粒各50粒进行图像识别,试验结果表明:识别准确率分别为94%和90%,基本满足玉米种子表面裂纹检测的精度要求。
The use of digital image processing technology to achieve the identification of corn seed surface cracks and detection. Light cold cathode fluorescent lamp (CCFL) was chosen to design the light environment of image acquisition, and the acquisition system of corn seed image was established. Then, a method of surface crack detection based on grain morphology was proposed for corn seed images. The method uses the horizontal and vertical edge detection operator to get the edge information such as crack, seed boundary and noise, and then looks for the tip position by the morphological characteristics of corn grain and removes the most non-crack information by using image algebraic operation. Finally, Crack length and location feature extraction crack, and calculate the absolute crack length and relative length. The results showed that the accuracy of identification was 94% and 90%, respectively, which basically met the precision of detecting the surface crack of corn seeds Claim.