dc.description.abstract | Electronic nose is aimed to simulate the human sense of smell. The major advantage of the electronic nose is its fast detection ability. This study uses the electronic nose Cyranose® 320 (Smiths Detection, Pasadena, USA, currently produced by IOS, Baldwin Park, USA), a handheld chemical vapor sensing device. It is composed of three parts: Sensors, signal processing and pattern recognition units. E-nose is widely applied by scientists now, mostly using in environmental monitoring, food storage inspection and microbial identification.Under this study we found that it can detect one sweet oranges infected by green mold among six sweet oranges in a PE bag, showing its high accuracy. We have tested out that when the E-nose is operated with 13 selected sensors, its accuracy is better than operated with all 32 sensors. Using this E-nose, the strawberry infected with gray mold can also be detected generally. Howeven we found the training is notsatisfactory,cansing a lower rate of accurate detection.Similarly, when the E-nose is operated with only 8 sensors,itsaccuracy is better than that operated with the 32 sensors. Phellinus noxius fungi growing on potato dextrose agar (PDA) were also dectected for their volatile gases by the E-nose. Results show that it can identify the smell of brown root rot fungi on PDA after culturing for 3 days. Wheras this E-nose showed no response to the other three fungi induding, Phellinus laevigatus, Colletotrichum gloeosporioides,and Phytophthora infestans.A headspace solid-phase microextraction gas chromatography–mass spectrometry (HS-SPME-GC-MS) was applied to analyze the volatile compounds related with the above three studys. Results showed that when sweet orange infected with green mold, a volatile gas limonene increase its release. When healthy strawberry fruits is infected byBotrytis cinerea for two days, the hexanoic acid has increased greatly. The Phellinus noxius fungi on PDA can release new component as determined to be 2-Butenedinitrile. The field application of electronic nose to detect trees brown root rot disease is initiated also in this study.When the electronic nose is trained by the smell of brown root rot fungion PDA, it can not deteck the brown root rot in diseased wood section and mycelial surface in the field. When it is trained by the smell of diseased woodof banyan roottree, it can not differentiate between the diseased parts and the healthy xylem parts,showing the similler response. When the E-nose is trained by the smell of themycelium surface, it can successfully identifythe other mycelium surface. While four mushroom fungi and oneGanoderma specied showed no response to this mode of E-nose.One Ganodermaspecied found on flamboyantrree,however, showed lower positive respons to the mode of E-nose. It seems to be that some ganoderma may produce similar smell as the brownroot rot fungi mycelium surface. In conclusion, the electron nose is proven to be able to detect many plant and tree diseases. We therefore recommend that it can be used to dectect a variety of storage diseases and field trees brown root rot disease in eraly stage. | en |