關(guān)于舉辦“Machine Learning and Computer Vision applications in Proximal Soil Sensing” 學(xué)術(shù)報(bào)告的通知

報(bào)告題目：Machine Learning and Computer Vision applications in Proximal Soil Sensing

報(bào) 告 人：Asim Biswas博士

報(bào)告時間：2019年12月9日（星期一）下午14：30-15：30

報(bào)告地點(diǎn)：資源環(huán)境學(xué)院307會議室

邀 請 人： 何海龍 副教授

報(bào)告摘要：Characterization and quantification of soil properties are important for the optimum use and management of our soil. Traditional methods for estimating soil properties are time consuming and laborious. In contrast, recent technological developments around proximal soil sensing has been showing great promise to meet the high-resolution spatial and temporal data demand for modern-day precision agriculture. More recently, with the advancements and developments in imaging techniques and computational powers of modern computer and handheld devices to process high resolution images have been gathering interest to characterize soil properties. Often the color and the surface textural characteristics of an image and image pixels is nothing but the presentation of the characteristics of that soil. Developing a relationship between the colors and the image surface textural properties as derived from an image with laboratory-measured soil properties show strong promise of image-based soil characterization. Image processing and various machine learning and computer vision algorithms provide the basis for developing these relationships. This paper brings together some examples from current studies on imaging soil in laboratory and in field conditions using various devices including handheld microscope, cell phone camera and digital camera and various image processing techniques including geostatistical, artificial neural network, support vector machine, wavelet transform to characterize soil texture and organic matter. Design and development of image acquisition systems, collection of soil images, processing and extraction of image parameters and development of models will provide information on the use of machine learning and computer vision applications to develop new proximal soil sensors.

報(bào)告人簡介：

：Asim Biswas博士曾先后在加拿大McGill University和University of Guelph任職助理教授和副教授。主要研究領(lǐng)域?yàn)檗r(nóng)業(yè)水文學(xué)，在包氣帶水文學(xué)、精準(zhǔn)農(nóng)業(yè)、土壤水分、熱及溶質(zhì)遷移和空間變異性等方面的研究非常深入，為量化研究土壤水文的相關(guān)屬性和機(jī)制奠定了基礎(chǔ)。Asim Biswas博士已發(fā)表SCI論文67篇，會議摘要131篇，參編著作1部，專利1項(xiàng)。目前主持科研項(xiàng)目十余項(xiàng)，其中主持項(xiàng)目的經(jīng)費(fèi)136萬美元；在國際會議和研討會上邀請發(fā)言28次，口頭報(bào)告84次，海報(bào)17幅。指導(dǎo)和合作指導(dǎo)研究生15名。先后獲得多向獎勵，包括2010年獲得加拿大土壤科學(xué)學(xué)會(CSSS)Bentley 獎；2010-2012年連續(xù)獲得SSSA年會最佳海報(bào)獎，2011年度加拿大地球物理聯(lián)盟(CGU) 最佳學(xué)生論文獎； 2012年University of Saskatchewan優(yōu)秀博士論文獎；2012年Kirkham 旅游獎；2014年獲美國地球物理聯(lián)盟(AGU)頒發(fā)的Donald L. Turcotte獎；2016年及2018年兩次獲美國土壤科學(xué)學(xué)會土壤物理與水文青年科學(xué)家獎；2018年獲美國土壤科學(xué)學(xué)會S6分會青年學(xué)者獎；2018年獲安大略省研究、創(chuàng)新和科學(xué)部早期研究員獎等。

歡迎有興趣的科教人員及研究生屆時與會交流。

資源環(huán)境學(xué)院

2019年12月6日