摘要
为了快速准确地诊断小麦病害,及时采取防治措施,提高小麦产量和质量,构建了基于多神经网络的小麦病害在线诊断系统。用户通过Android手机采集病害图像,在具备网络覆盖的地方将病害图片发送至诊断平台。诊断平台采用多神经网络模型对病害诊断,将多个神经网络预测值平均作为病害诊断结果,并根据多个神经网络预测值的方差计算出诊断结果的可信度。测试结果表明,该系统实现了病害的及时准确诊断,可信度参数具有提示作用,满足农田小麦病害诊断的实际需要。
To diagnose wheat diseases quickly and accurately, a system for online diagnosis of wheat diseases through multiple neural networks is proposed, which helps take precautions promptly and increase both output and quality of wheat. Users can collect images of diseases through the Android mobile phones and send the images to the diagnosis platform via the network.The platform can diagnose the wheat diseases using the multiple neural network model and compute the mean of multiple predictive values from the neural networks as the final diagnosis result. The variance of these predictive values is also calculated to measure the reliability of the diagnosis result. Test results show that the proposed system is capable of quickly and accurately diagnosing wheat diseases and the reliability parameter is informative. Thus, the proposed system can meet the practical need for wheat disease diagnosis in the farmland.
出处
《科技通报》
北大核心
2016年第10期59-62,共4页
Bulletin of Science and Technology
基金
国家自然科学基金(61302118)
河南省高校青年骨干教师资助计划(2010GGJS-114)
