自然背景下计算机视觉技术在西红柿采摘中的应用

发布时间：2018-03-17 17:31

本文选题：双目立体视觉　切入点：自动识别　出处：《农机化研究》2017年07期 　论文类型：期刊论文

【摘要】：针对西红柿种植过程中,采摘费时费力的问题,将计算机视觉技术与图像处理和模式识别技术相结合,开发了在自然背景条件下的西红柿自动识别系统。运用双目立体视觉系统来识别红色西红柿,去除原始图像的噪声、转换图像的灰度、对新图像进行分割和目标特征提取,从背景中把西红柿分离出来;同时,针对果实间存在的重叠问题,采用圆形Hough变换算法,提取西红柿的圆心坐标和半径特征,通过获取T=Sqrt(s×l)图像,恢复被遮挡西红柿。利用该技术,采集了不同条件下的100张西红柿照片,其中99张照片被准确识别,识别率达到9 9%以上,能够较好地满足西红柿采摘工作的要求。
[Abstract]:In order to solve the problem of time-consuming and laborious harvesting in tomato planting, the computer vision technology is combined with image processing and pattern recognition technology. An automatic tomato recognition system under natural background is developed, which uses binocular stereo vision system to recognize red tomato, remove noise from original image, convert gray scale of image, segment new image and extract target feature. The tomato is separated from the background, and the circular Hough transform algorithm is used to extract the center coordinates and radius characteristics of the tomato. Using this technique, 100 tomato photos were collected under different conditions, 99 of which were identified accurately, and the recognition rate was over 9 9%, which could meet the requirements of tomato picking.
【作者单位】：周口职业技术学院;
【基金】：河南省高等学校重点科研项目(15B520047)
【分类号】：TP391.41;S225.92

【相似文献】