关键词: 土壤水分, 水稻, 视觉检测技术, 籽粒形态, 粒重

Abstract: Based on machine vision detection technology, with Shanyou 63 (a representative common hybrid rice) and Guodao 6 (a super hybrid rice) as experiment materials, designed different water treatment from beginning to 30 days after panicle initiation and determined soil moisture by W.E.T Sensor (ML2x type, made in USA) and electronic balance (made in Shanghai, China), the grain morphological traits of full panicle and different position within a rice panicle under different water condition was studied, the relationship among these traits and their relation to grain weight was analyzed. The results indicated that the response of grain length, projective area, and 1000-grain weight to soil moisture was more obvious, the change of GRWA (ratio of grain weight to area) was caused by the effect of water deficit on grain area and weight, the response of ratio of grain length to width manifested complexity in a certain extent, and that of grain width to water was different in Shanyou 63 and Guodao 6. The grain morphologi-cal traits, yield and its components showed a drop trend of different extent with enhanced degree of water deficit, these grain morphological traits and grain weight reduced observably while controlling soil moisture from beginning to 30 days after panicle initiation, the effect of water deficit was most significant during 0–30 d control water, secondary, during 10–20 d heavy or middle water deficit and 0–10 d or 20–30 d heavy water deficit. The expressive laws of the effect of controlling water at different stage on grain morphological traits and weight of different position within a panicle was basically inosculated with the order of panicle initiation, growth and development. Among these morphological traits (except for width) all showed significant positive correla-tion at 0.01 probability level, while the relationship of grain width with other morphological traits was significant negative corre-lation at 0.01 or 0.05, respectively. The correlativity existed significantly between grain weight and grain length, projective area and ratio of grain length to width at 0.01 or 0.05, respectively. Therefore, the three traits can be used as grain morphological effec-tive index to appraise drought-resistance in rice.

Key words: Soil water, Rice (Oryza sativa L.), Vision detection technology, Grain morphology, Grain weight