基于多参数信息融合的种子活力检测方法研究

Seed is the most basic means of production in the agriculture and forestry, the level of seed quality not only affects the role of elite seed but also relates to the harvest of the agriculture, relates to the effectiveness of the forestry construction and relates to the development of the national agriculture and forestry..Seed vigor is an indicator for evaluation of seed quality. Traditional detection methods of seed vigor include the standard germination testing method, the accelerated aging testing method and others, the shortcomings of them are the longer testing period and destructive to the seed. The methods based on new technologies including near infrared spectroscopy analysis, Q2, infrared imaging method and laser bio-speckle detection method have limitations and uncertainties of the information on seed vigor..So this subject plans to combine the academic advantages of forest cultivation of Beijing Forestry University, with several usual kinds of tree seeds as the research objects, capture the laser bio-speckle image, thermal infrared image and visible image, using image processing and pattern recognition technology to set up the multi-parameter model of seed vigor, explore the mechanism of information fusion technology on testing seed vigor to establish a rapid, non-destructive and accurate detection method. .This subject will lay the theory and practice foundation for the subsequent development of rapid seed vigor sorting system, play an significant role on fine seed breeding, spread of precise selection and sowing of the crop and also in national strategy implementation of precision agriculture and forestry.

种子是农林生产中最基本的生产资料，种子质量的高低不仅影响良种特性的发挥，而且关系到农业的丰歉，林业建设的成效，关乎国家农林业的发展。.种子活力是评价种子质量的一项重要指标。传统的种子活力检测方法包括标准发芽试验法、加速老化试验法等存在着周期较长，对种子本身有损害等缺点，而基于新技术的近红外光谱分析法、Q2法、红外热成像法及激光生物散斑检测方法具有信息的局限性与不确定性。.因此本课题拟结合我校森林培育学科优势，以几种常见的林木种子为研究对象，采集种子整个萌发过程的激光生物散斑图像、热红外图像与可见光图像信息，用图像处理与模式识别方法，建立多参数种子活力模型，探索信息融合技术在种子活力检测上的应用机理，实现无损、准确、快速的种子活力检测。.本课题的研究为后续快速种子活力分选系统的开发奠定理论与实践基础，对农林作物良种选育、精选精播技术的推广起着重大作用，对国家精准农林业战略实施有着重要意义。

种子是农林生产中最基本的生产资料，种子质量的高低不仅影响良种特性的发挥，而且关系到农业的丰歉，林业建设的成效，关乎国家农林业的发展。.种子活力是评价种子质量的一项重要指标。种子活力检测是保证田间出苗率和生产潜力的必要手段，也是种子产业中必不可少的环节。传统的种子活力检测方法包括标准发芽试验法、加速老化试验法等存在着周期较长，对种子本身有损害等缺点，而基于新技术的近红外光谱分析法、Q2 法、红外热成像法及激光生物散斑检测方法具有信息的局限性与不确定性。.本课题利用我校森林培育学科优势，以栓皮栎、青豌豆、麻皮豌豆等为研究对象，采用人工老化的方法，制备不同活力等级的实验样品，并采集种子萌发过程中的激光生物散斑图像、红外热图像及可见光图像信息。通过图像处理方法，得到散斑强度变化曲线、种子表面温度变化曲线及外观变化数据。把这些数据进行融合，建立多参数种子活力模型，用以检测种子活力等级，从生物角度分析并解释其原理。实验结果显示，栓皮栎种子、青豌豆种子、麻皮豌豆种子活力检测的准确度分别可达93.8%、91.5%、89.4%，达到了无损、准确、快速的种子活力检测的目的。.本课题的研究为后续快速种子活力分选系统的开发奠定理论与实践基础，对农林作物良种选育、精选精播技术的推广起着重大作用，对国家精准农林业战略实施有着重要意义。