基于机器视觉的新生仔猪目标识别方法研究与实现
发布时间：2019-05-29 23:00
【摘要】：[目的]新生仔猪目标检测是母猪分娩监测的关键环节。[方法]通过自制图像采集器采集母猪分娩视频图像,机器视觉系统获取分娩图像信息,选取Canny算子对图像进行边缘检测,采用Otsu算法对图像进行二值变换,应用滑动平均算法和形态学开运算对二值图像滤波消噪,提取图像最大连通域,利用团序列检测算法对母猪目标进行分割,对分割后区域进行仔猪目标识别。[结果]试验结果表明,团序列检测算法能够准确分割出母猪目标,检测仔猪目标的正确率达到95.5%。[结论]提出一种能够有效识别新生仔猪目标的方法,为仔猪的出生预警提供了技术支撑。
[Abstract]:[objective] Target detection of newborn piglets is the key link of sows delivery monitoring. [methods] the video images of sow delivery were collected by self-made image collector, the delivery image information was obtained by machine vision system, the edge of the image was detected by Canny operator, and the binary transformation of the image was carried out by Otsu algorithm. The moving average algorithm and morphological open operation are used to filter and Denoise the binary image, extract the most Dalian domain of the image, segment the sows' target by using the cluster sequence detection algorithm, and recognize the target of piglets in the segmented area. [results] the experimental results show that the cluster sequence detection algorithm can accurately segment sows, and the correct rate of detection of piglets is 95.5%. [conclusion] A method which can effectively identify the target of newborn piglets is proposed, which provides technical support for birth early warning of piglets.
【作者单位】： 南京农业大学工学院;
【基金】：国家自然科学基金项目(61503187) 江苏省农机三新工程项目(NJ2016-10) 江苏省产学研合作前瞻性联合创新资金项目(BY2014128-01)
【分类号】：S818.9;S828
本文编号：2488297
[Abstract]:[objective] Target detection of newborn piglets is the key link of sows delivery monitoring. [methods] the video images of sow delivery were collected by self-made image collector, the delivery image information was obtained by machine vision system, the edge of the image was detected by Canny operator, and the binary transformation of the image was carried out by Otsu algorithm. The moving average algorithm and morphological open operation are used to filter and Denoise the binary image, extract the most Dalian domain of the image, segment the sows' target by using the cluster sequence detection algorithm, and recognize the target of piglets in the segmented area. [results] the experimental results show that the cluster sequence detection algorithm can accurately segment sows, and the correct rate of detection of piglets is 95.5%. [conclusion] A method which can effectively identify the target of newborn piglets is proposed, which provides technical support for birth early warning of piglets.
【作者单位】： 南京农业大学工学院;
【基金】：国家自然科学基金项目(61503187) 江苏省农机三新工程项目(NJ2016-10) 江苏省产学研合作前瞻性联合创新资金项目(BY2014128-01)
【分类号】：S818.9;S828
【相似文献】
相关博士学位论文 前1条
1 郁志宏;基于机器视觉的种蛋筛选及孵化成活性检测研究[D];内蒙古农业大学;2006年
相关硕士学位论文 前3条
1 刘卫民;基于机器视觉的奶牛体尺参数测量研究[D];山东农业大学;2016年
2 曹鹏程;基于机器视觉和力学特性肉鸡体质健康模型构建[D];新疆农业大学;2016年
3 邓海霞;基于机器视觉的群体禽蛋大小检测方法研究[D];华中农业大学;2006年
，本文编号：2488297
本文链接：https://www.wllwen.com/yixuelunwen/dongwuyixue/2488297.html
