摘要
以番茄作为研究对象,提出基于区域特征的果实识别技术.识别过程包括灰度变换、区域边缘锐化、区域分割、细化处理等步骤.对70幅图像进行的实验表明,其正确识别率为74.3%.如将该技术应用到类似果实的识别领域,仅需对灰度变换公式进行适当的调整.
A new technology of discriminating fruit on trees is presented based on region feature, in which tomato is set as study subject. The discriminating process includes gray transformation, region edge enhancement, region segmentation et al. Seventy images are tested and 74.3 percent images are correctly technology is applied to the similar fruits discrimination, all need to do is to adjust the formula.
出处
《河南科学》
2008年第3期339-342,共4页
Henan Science
基金
国家自然科学基金项目(60075006)
关键词
图像识别
区域特征
番茄
果实
image recognition
region feature
tomato
fruit
