Detection of Apple Lesions in Orchards Based on Deep Learning Methods of CycleGAN and YOLOV3-Dense
原文链接下载
提出了一种基于深度学习的炭疽病检测方法。首先,针对苹果病害的随机发生导致图像数据不足的问题,在传统图像增强技术的基础上,采用CycleGAN深度学习模型完成数据的增强。在图像数据增强的基础上,利用densely connected neural network(DenseNet)对YOLO-V3模型中分辨率较低的特征层进行优化。实验证明,改进的模型超过带有VGG16的Faster R-CNN网络、原始YOLO-V3模型和其他三个最先进的检测网络。该方法可以很好地应用于果园苹果表面炭疽病的检测。
Inappropriate farming decisions result in the waste of labors and resources, and plant diseases cannot be well controlled.
In our previous work, a BP neural network improved by genetic algorithm was applied to realize multithreshold image processing. The region of green apple in the image was segmented, and the lesion area was extracted by subsequent SVM method. The diseased apple was further identified [10]. This method has been well applied in the images collected in orchards but could not achieve realtime image processing.
However, because Faster RCNN [21] and other R-CNN models include two procedures of region proposal generation and classification, the model cannot realize real-time detection.
In our previous work [26], DenseNet was utilized to substitute the low resolution feature layers in YOLO network. The detection performance was therefore improved.
Because the incidence of apple disease is relatively random, it is very difficult to acquire a large number of specific disease images. To overcome this deficiency, CycleGAN deep learning method is adopted to expand the datasets in this paper.
CycleGAN can learn the features of one type of images and transplant them to another. By this means, features of apple lesion images can be extracted and transplanted to healthy apple images. Hence, the dataset of diseased apple images is enlarged.
使用两类数据,500张健康苹果和140张患病苹果(anthracnose)。采用传统的图像增强技术和CycleGAN深度学习方法对患病苹果图像进行预处理和扩展。最终将患病类别图像扩展到传统方法即颜色、角度和亮度变换来处理原始图像。
为了更好地比较不同检测模型的性能,将最终的训练集图像制成Pascal VOC的格式。
YOLOV3-dense模型
在神经网络的训练过程中,卷积和降采样操作减小了特征层的大小,导致特征丢失。为了更有效地利用这些特征,并使其损失最小化,DenseNet采用前馈的方法将每一层连接到每一层。对原文Darknet-53架构进行了改进。采用DenseNet代替尺度较小的特征层。当特性传输到这些层时,它们将被DenseNet中的多个特性层重用,从而减少损失。
评价指标 | 描述 |
---|---|
Precision-Recall (P − R) Curve and F1 Score | 用于评估目标检测模型的性能 |
IoU | 用于评价bounding box精确度的参数 |
Loss Value | 用于评估神经网络在训练时的收敛情况 |
Average Detection Time | 分析其实时性 |
传统数据增强(color/bright/angle)
CycleGAN增强结果
这项工作通过使用CycleGAN做数据增强,对其网络没有改动。使用改进的YOLO-V3检测模型,替换了特征提取层为DenseNet。优点:对比实验完整。
可供参考的对CyclGAN原论文的解读
对抗神经网络CycleGAN论文解读
GAN学习历程之CycleGAN论文笔记
Q. Li, M. Wang, and W. Gu, “Computer vision based system for apple surface defect detection,” Computers and Electronics in Agriculture, vol. 36, no. 2-3, pp. 215–223, 2002.
本文发布于:2024-01-28 11:25:01,感谢您对本站的认可!
本文链接:https://www.4u4v.net/it/17064123077085.html
版权声明:本站内容均来自互联网,仅供演示用,请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系,我们将在24小时内删除。
留言与评论(共有 0 条评论) |