基于卷积神经网络马铃薯叶片病害识别和病斑检测*     

王林柏  张博  姚竟发  杨志辉  张君  范晓飞 《中国农机化学报》2021,42(11):122

针对自然环境下马铃薯叶片病害识别率低和晚疫病斑定位难的问题,基于大田环境中采集的马铃薯叶片图像,首先对马铃薯叶片病害进行识别,对比AlexNet、VGG16、InceptionV3、ResNet50、MobileNet五种神经网络模型,结果表明InceptionV3模型的识别效果准确率最高,可达98.00%。其次对马铃薯叶片的晚疫病斑进行检测,提出一种改进型的CenterNet-SPP模型,该模型通过特征提取网络获取对象的中心点,再通过中心点回归获得中心点偏移量、目标大小等图像信息,训练后的模型在验证集下的mAP可达90.03%,以F₁为评价值分析对比其它目标检测模型,CenterNet-SPP模型的效果最好,准确率为94.93%,召回率为90.34%,F₁值为92.58%,平均检测一张图像耗时0.10 s。为自然环境下马铃薯叶片病害识别和检测提供较为全面的深度学习算法和模型研究基础。  相似文献   

基于UHPLC-QTOF-MS代谢组学方法筛选奶牛不同妊娠状态时的候选生物标志物     

罗芳  陶金忠 《畜牧兽医学报》2021,52(11):3118-3125

旨在筛选人工授精后第17天奶牛在不同妊娠状态时的候选生物标志物。本研究以宁夏某奶牛场体重为（550±50） kg,体况评分相近的经产（2~3胎次）健康荷斯坦奶牛为试验对象。同期发情后,在人工授精后第17天晨饲前对奶牛进行尾静脉采血,后期采用计步器和B超仪诊断出奶牛的妊娠状态。根据诊断结果将奶牛分为妊娠组（A组,n=12）和未妊娠返情组（B组,n=24）,将这两组血样进行代谢轮廓及代谢物变化分析。主成分分析（PCA）和正交偏最小二乘法-判别分析（OPLS-DA）结果显示,A、B两组血浆代谢轮廓均发生了明显变化,A、B两组之间共检测出8种ROC （Receiver operating characteristic curve）曲线下面积（area under the ROC curve,AUC）＞0.8的差异代谢物。丙氨酸-亮氨酸、甘氨酸、脯氨酸、L-正亮氨酸、DL-苯丙氨酸、肌氨酸、吡咯-2-羧酸和缬氨酸-蛋氨酸有望成为识别妊娠组（A组）和未妊娠返情组（B组）的差异代谢物。综上表明,血浆中的这8种代谢物有望成为奶牛妊娠识别阶段潜在的生物标志物,为配种早期妊娠识别提供新思路。  相似文献   

计算机视觉技术在智慧羊场中的研究进展     

姚州  谭焓  田芳  周勇  章程 《中国饲料》2021,1(7):7-12

智慧羊场具有科学化、标准化管理的优点,是羊场转型发展的趋势。计算机视觉技术是人工智能的核心,可以实现非接触性、自动化、实时监测羊的个体信息。本文总结并分析了计算机视觉在羊场的个体识别方法、行为识别方法、体尺与体重估测、疾病监测的研究现状,并指出计算机视觉在智慧羊场的进一步发展趋势。 [关键词] 计算机视觉|智慧羊场|个体识别|行为识别|体尺与体重估测|疾病监测  相似文献   

上海市犬只狂犬病认定免疫点主要管理做法分析     

沈悦 《中国动物检疫》2021,38(4):67-70

强化犬只狂犬病免疫工作是从源头控制狂犬病的最有效措施。近年来,上海市按照《上海市养犬管理条例》的要求,通过严把免疫资质准入、规范内部管理、落实疫情报告责任、推进免疫质量考核等举措,使犬只狂犬病认定免疫点的布局日趋合理,文明服务更加规范,这已成为上海市民犬只狂犬病免疫的主渠道,免疫数量和质量逐年提升,构筑了良好的免疫屏障。今后在犬只狂犬病认定免疫点管理方面,上海市应强化疫苗源头管理,加大非法免疫惩处力度,合理认定免疫点整体布局,规范引导免疫行为和内部运行。  相似文献   

稻粉虱的发生及其化学防治药剂筛选     

魏琪  何佳春  赖凤香  王渭霞  万品俊  傅强 《植物保护》2023,49(4):284-292

稻粉虱Aleurocybotus indicus David et Subramaniam, 属半翅目粉虱科, 以刺吸式口器为害水稻叶片而导致产量损失, 是我国闽?赣?浙等稻区的一种偶发性水稻害虫, 近20多年已鲜有报道, 但2019年-2021年在中国水稻研究所富阳试验农场稻田暴发?本文描述了稻粉虱的为害症状和形态特征, 并配以原色照片; 克隆得到了稻粉虱线粒体DNA COⅠ基因片段, 发现其与烟粉虱Bemisia tabaci序列一致性最高(85.80%); 田间调查表明, 富阳试验农场及周边稻田稻粉虱均有不同程度的发生, 但前者发生量明显较高(每个分蘖最高可达251头), 推测可能与农场温室内的稻粉虱越冬虫源有关; 利用稻苗浸渍法评估了不同作用机制的7种杀虫剂对稻粉虱的毒力, 发现双丙环虫酯?氯虫苯甲酰胺?氟啶虫胺腈?氟啶虫酰胺和溴氰虫酰胺对稻粉虱成虫均有较好的杀虫活性, 其中前4种药剂推荐中剂量处理后72 h的校正死亡率均可达100%, 其LC50分别为3.028?3.769?6.405 mg/L和7.894 mg/L?本研究可为我国稻粉虱的识别诊断和化学防治提供重要依据?  相似文献   

Current knowledge about the processes of luteolysis and maternal recognition of pregnancy in camelids     

Carolina Paula Bianchi  Maria F. Gallelli  Juan Manuel Herrera  Micaela A. Benavente  Liliana Rossetto  Marcelo Alfredo Aba 《Reproduction in domestic animals》2023,58(1):3-9

Camelids have many unique reproductive features that considerably differ from those of other domestic species. Females are induced ovulators with subsequent development of a corpus luteum (CL) with a short lifespan. Plasma progesterone concentration starts to increase on day 4, peaks on day 8–9 and, in non-pregnant animals, basal concentration is reached around day 10–11 post-induction of ovulation. Luteolytic pulses of prostaglandin F_2α (PGF_2α) are firstly detected on day 7 or 8 (approximately on day 5–6 after ovulation), with maximal luteolytic peaks observed between days 9 and 11 post-mating, in coincidence with a high endometrial expression of cyclooxygenase 2, a limiting enzyme in prostaglandins synthesis. Unlike other species, oxytocin seems not to be involved in the luteolytic process in these species. The CL is the main source of progesterone secretion, and its function is required to support pregnancy. Despite constant research efforts, aspects of reproduction and maternal recognition of pregnancy in camelids remain not fully understood. A transient decrease and subsequent recovery in plasma progesterone concentration are observed after day 9 post-mating in pregnant animals in association with a pulsatile release of PGF_2α and a transitory decrease in CL vascularization. Thus, embryo recognition should occur between days 8 and 12 post-mating. In camels, conceptus tissues exhibit aromatizing activity with the capacity to synthesize large amounts of oestradiol. Similarly, llama blastocysts secrete oestradiol-17β during the preimplantation stage, with a higher production during the elongation period. An increase in the endometrial expression of oestrogen receptor α is also observed on day 12 post-mating. All these evidences suggest that oestrogen could be the signal released by the embryo at the time of its recognition in camelids. Besides, nearly 98% of pregnancies are carried out in the left horn. A decrease in the endometrial expression of mucin 1 and 16 genes has been reported, suggesting that these changes are crucial for successful embryo implantation; however, no differences have been observed between horns. Thus, maternal recognition of pregnancy in camelids is a particularly complex process that must occur in a concise time to allow the rescue of the CL and embryo survival.  相似文献   

Pd Pattern Recognition Based on Linear Discriminant Analysis in GIS     

ZHANG Xiao-xing  TANG Ju  SU Cai-xin  XU Zhong-rong  ZHOU Qian 《保鲜与加工》2006,(10):1-4

According to the character of PD in GIS, the authors design four kinds of GIS defection models. The GIS gray intensity images are constructed based on mass specimens gathered by the ultra - high frequency and high speeds systems, Aiming at the PD characteristics and its defections, A PCA-FDA method is put forward based on PD images. The principal component analysis is employed to condense the dimension of PD images, then the optimal sets of statistically uncorrelated discriminant vectors are extracted, and the minimum distance classifier is constructed as classifier. The identified results show that this method can effectively elevated the discrimination of the four kinds of defects in GIS PD.  相似文献   

Vehicle Type Recognition by Using Support Vector Machine SVM     

XlAO Han-guang  CAI Cong-zhong  WANG Wan-lu 《保鲜与加工》2006,(1):61-65

  相似文献   

A QUICK THINNING AND SEGMENTATION ALGORITHM FOR HAND-PRINTED CHINESE CHARACTERS     

Liu Ping 《保鲜与加工》1991,(4):78-85

  相似文献   

Developing a Recognition System of Gun-code     

YANG Bo  PENG Jian  YE Jun yong  LIU Jian sheng  WANG Tong qing 《保鲜与加工》2001,(5):139-141148

Intelligent recognition system of gun code is designed , which achieves online recognition of gun code. We applied video camera to get image, and some special technique in this deformed character recognition. As well, we applied dynamic template matching to get a better result.  相似文献