Quantifying biological damage to plant leaves by convolutional neural networks

• Main Authors: EGGERS TILL, CLOUKAS, CHRISTOPH, BEREARTUA-PEREZ ANDRES, NAVARRA-MEISTER, RAFAEL, ECHAZALA HUGUET, JES ¨ 2 S, SPANGLER CHRISTOPHER MICHAEL, PICON RUIZ, ANTOINE
• Format: Patent
• Language: Chinese; English
• Published: 04.11.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; PHYSICS

To quantify biological damage to leaves of a crop plant, a computer receives (701A) a leaf image taken from a particular crop plant. The leaf image shows at least one of the leaves of the specific crop plant. The computer processes the leaf image using a first convolutional neural network (CNN, 262) to derive a segmented leaf image (422) that is a set of contiguous pixels that entirely show a main leaf of the particular plant. The first CNN has been trained by a leaf image (601A) annotated by a plurality of leaves, where the leaf image is annotated to identify a primary leaf (461). The computer processes the single leaf image by regression using the second CNN (272) to obtain a degree of damage (432). 为了量化作物植物的叶子的生物损害，计算机接收(701A)从特定作物植物拍摄的叶子图像。叶子图像示出特定作物植物的叶子中的至少一个叶子。计算机使用第一卷积神经网络(CNN，262)处理叶子图像以导出分割的叶子图像(422)，该叶子图像(422)是完全示出特定植物的主叶子的连续像素的集合。第一CNN已经由多个叶子注释的叶子图像(601A)训练，其中，叶子图像被注释以识别主叶子(461)。计算机使用第二CNN(272)通过回归处理单叶图像以获得损害程度(432)。