基于图像识别的奶牛计量瓶自动计量系统的研究与设计

发布时间：2018-03-30 21:06

本文选题：图像识别　切入点：奶量计量瓶　出处：《北方民族大学》2013年硕士论文

【摘要】：在我国奶产业的持续发展和技术上的不断更新中，奶牛产奶量的自动计量是现代化奶牛养殖场中必不可少的环节。我国大型的养殖基地中，多数引进进口设备，以此实现自动挤奶和奶量自动计量的目的，，然而这些进口设备普遍存在着价格昂贵，不便维护等缺点。本项目使用图像识别及检测技术来分析和识别奶量计量瓶中的牛奶液面并根据液面位置计算牛奶重量，从而达到记录个体产奶量的目的；为了实现自动识读奶牛产奶量，本项目采用嵌入式ARM9开发板和数字摄像头采集挤奶台计量瓶刻度表和牛奶液面位置的数字图像，然后通过以下步骤得到最后的计量结果：（1）利用Hough直线检测方法对图像进行纠偏操作；（2）对图像进行灰度化；（3）对灰度化的图像进行滤波，消除图像中的噪声；（4）对滤波后的灰度图像进行直方图处理；（5）使用Canny边缘检测方法检测牛奶液面；（6）根据检测的液面位置计算牛奶的重量。该系统所使用的硬件设备造价低，软件开发环境使用开源的Linux操作系统和OpenCV开源视觉库，节约了开发成本。经现场安装测试表明，该系统可准确测量牛奶液面刻度，计量误差低于百分之三。系统的开发对于改善宁夏奶牛养殖园区工作人员的工作环境、提高奶牛养殖园区工作效率和经济效益、促进宁夏奶业的自动化和信息化将产生深远的影响和巨大的推动作用，具有一定的创新性和实用价值。
[Abstract]:In the continuous development of China's dairy industry and the continuous renewal of technology, the automatic measurement of milk production is an essential link in modern dairy farms. Most of the large breeding bases in China have imported imported equipment. In order to realize the purpose of automatic milking and automatic metering of milk quantity, however, these imported equipments generally have the disadvantages of high price and inconvenient maintenance. This project uses image recognition and detection technology to analyze and identify the milk level in the milk quantity measuring bottle and calculate the milk weight according to the position of the liquid level, so as to achieve the purpose of recording the individual milk production, in order to realize the automatic recognition of the milk yield of the cow, This project adopts embedded ARM9 development board and digital camera to collect digital image of measuring bottle scale and milk liquid level position of milking station. Then, through the following steps, we get the final measurement result: 1) using the Hough line detection method to correct the deviation of the image. (2) graying the image and filtering the grayscale image. Elimination of noise in the image (4) histogram processing of filtered grayscale image (5) Canny edge detection method is used to detect the milk level. (6) the weight of milk is calculated according to the detected liquid level position. The hardware equipment used in the system is low in cost. The software development environment uses open source Linux operating system and OpenCV open source visual library, which saves development cost. The field installation test shows that the system can accurately measure the milk level calibration, and the measurement error is less than 3%. The development of the system can improve the working environment of the staff in Ningxia dairy cattle breeding park. Improving the efficiency and economic benefit of dairy farming park and promoting the automation and information of dairy industry in Ningxia will have a profound impact and a huge role in promoting the dairy industry. It has some innovative and practical value.
【学位授予单位】：北方民族大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：TP391.41

【参考文献】