1. 研究目的与意义
水稻是我国主要的粮食作物,其种植面积十分广泛,单产量最高,总产约占粮食物的40%。
水稻生产最基本的生产资料是水稻种子,稻种的好坏直接影响水稻的产量和质量。
不同年份的稻种,发芽率不同,其水稻产量和质量是不同的。
2. 课题关键问题和重难点
关键问题:(1)图像采集:一般用图像采集卡,将视频/图象经过采样、量化以后转换为数字图象并输入、存储到帧存储器的过程。比较典型的PCI或AGP兼容的捕获卡,可以将图像迅速地传送到计算机存储器进行处理,这是机器视觉系统的前端和信息来源。 (2)图像增强:图像增强是指按特定的需要突出一幅图像的某些信息,同时,削弱或除去某些不需要的信息的处理方法。图像增强的目标是改进图片的质量,这是机器视觉系统的前端和信息来源。(3)照明设计:主要包括三个方面:光源、目标和环境的光反射和传送特性、光源的结构。光源决不仅仅是为了照亮物体,通过有效的光源设计可以令需要检测的特征突出,同时抑制不需要的干扰特征,给后端的图像处理带来极大的便利。而不恰当的照明方案会造成图像亮度不均匀,干扰增加,有效特征与背景难以区分,令图像处理变得极其困难,甚至成为不可能完成的任务。
难点:目前稻米静态米粒外观检测算法有待于进一步改进提高检测精度,减少算法处理时间。例如已有的检测算法对稻米摆放方向要求比较高,实际稻米摆放是随机的,检测算法不能受米粒摆放方向影响。虽然有些算法满足了旋转不变性要求,但算法耗时多,实时性差,限制了实用性;谷物外观品质检测中,绝大多数研究是针对静态的米粒个体,而在线检测采集的图像是动态图像序列,许多相对于静态图像的算法不能满足实时性的要求,不适应动态图像序列的特征提取,如何快速地从动态的米粒群体中提取有效的图像信息是米粒外观品质实时检测中的难点,也是目前研究中的薄弱环节;机器视觉系统以图像作为处理对象,处理信息量大,在对谷物外观品质进行检测时,大多采用串行算法,大大影响了计算机的处理速度,降低了检测分级的效率,所以需要研究并行的实时图像处理算法来提高图像处理速度;目前图像处理多采用灰度图像,某些参数可能无法识别,今后应多采用彩色图像、多频图像处理进行特征识别。
3. 国内外研究现状(文献综述)
机器视觉技术在农业上的应用研究始于20世纪70 年代末期,主要用于植物种类的鉴别、农产品品质检测和分级等[[]]。随着计算机软硬件技术、图像处理技术的迅速发展,它在农业上的应用研究有了很大进展,目前该项研究仍是国际农业工程领域中的热门课题[[]]。采用机器视觉技术对稻米的品质进行检测,不仅能够提高稻米品质的检测效率,而且能够克服主观因素的影响,降低检测误差,使得稻米品质的检测变得更加快速、客观和准确[[]]。
国内研究现况: 品种的鉴定与识别是种子质量检验的一个比较复杂的基础问题,也是众多学者们研究比较多的一个方面。实践证明,运用计算机视觉技术通过提取种子外部形状参数来进行各种品种识别和质量检测是非常有效的。国内对于机器视觉在农产品检测方面的应用研究起步较晚,主要开始于九十年代初,与国外研究还有一定差距,宋韬[[]]等人(应用机器视觉技术选择并获得了个玉米粒形态参数,采用冲量算法建立了一个层前馈神经元网络,实现了在任意放置玉米粒的情况下都能在线自动识别出完好玉米与破损玉米,对粒完整玉米以及粒破损玉米的识別试验显示,正确率为。成芳[[]]根据机器视觉检测杂交水稻种子质量的要求,对单粒、静态稻种图像进行霉变分析识基于数学形态学稻种纹理特征提取识别研究另,比较了提取颜色特征的种方法,研究了基于颜色特征的稻种霉变检测法。黄星弈[[]]对基于机器视觉的稻谷品种的识别技术进行了研究,提出了将图像的颜色特征和形状特征相结合进行识别的方法,通过贝叶斯决策方法设计识别分类器,识别的正确率达到以上。凌云[[]]以稻米、小麦和玉米为检测对象,研究了一套动态外观品质检测系统。该系统由硬件和软件两部分组成,系统硬件由计算机、摄像机、进料机构、驱动机构、出料机构和光源系统等几部分组成。吴继华、刘燕德、欧阳爱国研究了一种基于机器视觉技术的快速低成本实时检测杂交水稻种子品种新系统和方法。该系统包括自动上料机构、光照箱、图像釆集卡、摄像头和自动下料机构等硬件以及图像处理及品种识别软件组成。
国外研究现状:when it comes to machine vision, industrial automation is by far the most common application. still, its important to remember that agriculture has benefited from advances in technology since the invention of the plow. with every change in irrigation and cultivation techniques, it becomes possible to raise healthier crops with larger, more nourishing yields. now, machine vision is being applied to agriculture in increasingly effective and novel ways[[]]. computer and machine vision: theory, algorithms, practicalities(previously entitled machine vision) clearly and systematically presents the basic methodology of computer and machine vision, covering the essential elements of the theory while emphasizing algorithmic and practical design constraints. this fully revised fourth edition has brought in more of the concepts and applications of computer vision, making it a very comprehensive and up-to-date tutorial text suitable for graduate students, researchers and rd engineers working in this vibrant subject[[]]. the team previously developed a similar algorithm for facial memorability. whats notable about the new one, besides the fact that it can now perform at near-human levels, is that it uses techniques from deep-learning, a field of artificial intelligence that use systems called neural networks to teach computers to sift through massive amounts of data to find patterns all on their own[[]]. to keep the required number of samples low, the researchers adopted a simplified technique for evaluating hypotheses. suppose that the algorithm has identified three objects from one perspective and four from another. the most mathematically precise way to compare hypotheses would be to consider every possible set of matches between the two groups of objects: the set that matches objects 1, 2, and 3 in the first view to objects 1, 2, and 3 in the second; the set that matches objects 1, 2, and 3 in the first to objects 1, 2, and 4 in the second; the set that matches objects 1, 2, and 3 in the first view to objects 1, 3, and 4 in the second, and so on. in this case, if you include the possibilities that the detector has made an error and that some objects are occluded from some views, that approach would yield 304 different sets of matches[[]]. machine graphics visionis a refereed international journal, published quarterly by thefaculty of applied informatics and mathematics(wzim) of the warsaw university of life sciences - sggw, in cooperation with theassociation for image processing, poland (tpo). machine graphics vision[[]]provides a scientific exchange forum and an authoritative source of information in the field of, in general, pictorial information exchange between computers and their environment, including applications of visual and graphical computer systems. until june 2013 the journal has been maintained by theinstitute of computer scienceof the polish academy of sciences. see the previous version of the web pages of the journal. the journal concentrates on theoretical and computational models underlying computer generated, analysed, or otherwise processed imagery. elected results of the iteration as the binarization threshold, the result of the pretreatment image binaried as figure 4 shown in. after processed, outline of black sugarcane inter node has become more clearer, showing more prominent details of the image, the position of the sugarcane internode is also available to identify; the white part of the black sugarcane stem image is not existed, more to improve the accuracy of the stem and the internode can be correctly identified[[]]. 对江苏产粳米进行内外品质相关性实验分析,研究米粒部分外观特征对米,粒直链淀粉含量和胶稠度的影响,观察外观的分选能否改进稻米的食味品质,为稻米外观分选提供理论依据[[]].
4. 研究方案
方案一:智能传感器和智能相机稻种年份自动识别分选系统设计组成部分:光源、工业ccd、传感器、图像分析处理软件、通讯接口、电脑主机等。方案实施步骤: 第一步:启动检测系统,打开相机、照明光源并调好亮度,图像采集系统准备就绪。 第二步:用光电传感器检测是否有稻种存在,若存在则图像采集系统开始采集信息。 第三步:由电脑上的图像处理软件对采集图像进行分析和检测。 第四部:与已经设置好的参数作比较判断稻种的年份。
方案二:基于pc技术的采用板卡和sdk的稻种年份自动识别分选系统设计
组成部分:图像采集卡、相机、传感器、电脑主机。
5. 工作计划
第 1 周 接受任务书,领会课题含义,按要求查找相关文献资料,列出 阶段实施计划;
第 2 周 阅读相关资料,理解有关内容;
第 3 周 翻译相关英文资料,提出拟完成本课题的方案,写出相关开题报告一份;
