A novel green apple segmentation algorithm based on ensemble U-Net under complex orchard environment

Li, Qianwen; Jia, Weikuan; Sun, Meili; Hou, Sujuan; Zheng, Yuanjie

doi:10.1016/j.compag.2020.105900

Cited by 75 publications

(39 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, deep learning methods are gradually being applied to agricultural research because they can automatically learn the deep feature information of images, and their speed and accuracy levels are greater than those of traditional algorithms [27][28][29][30]. Deep learning has also been applied to the detection of plant diseases from visible light images.…”

Section: Introductionmentioning

confidence: 99%

Plant Disease Recognition Model Based on Improved YOLOv5

Wu²,

et al. 2022

View full text Add to dashboard Cite

To accurately recognize plant diseases under complex natural conditions, an improved plant disease-recognition model based on the original YOLOv5 network model was established. First, a new InvolutionBottleneck module was used to reduce the numbers of parameters and calculations, and to capture long-distance information in the space. Second, an SE module was added to improve the sensitivity of the model to channel features. Finally, the loss function ‘Generalized Intersection over Union’ was changed to ‘Efficient Intersection over Union’ to address the former’s degeneration into ‘Intersection over Union’. These proposed methods were used to improve the target recognition effect of the network model. In the experimental phase, to verify the effectiveness of the model, sample images were randomly selected from the constructed rubber tree disease database to form training and test sets. The test results showed that the mean average precision of the improved YOLOv5 network reached 70%, which is 5.4% higher than that of the original YOLOv5 network. The precision values of this model for powdery mildew and anthracnose detection were 86.5% and 86.8%, respectively. The overall detection performance of the improved YOLOv5 network was significantly better compared with those of the original YOLOv5 and the YOLOX_nano network models. The improved model accurately identified plant diseases under natural conditions, and it provides a technical reference for the prevention and control of plant diseases.

show abstract

Section: Introductionmentioning

confidence: 99%

Plant Disease Recognition Model Based on Improved YOLOv5

Wu²,

et al. 2022

View full text Add to dashboard Cite

show abstract

“…To accurately locate the tomato fruit under the complex scenes, Liu improved the YOLOv3 one-stage target detection model to predict the circular areas ( Liu et al, 2020 ), and the precision and times of YOLO-Tomato were 94.75% and 54 ms. Jia optimized the Mask R-CNN to adapt to the detection of apple targets, where the residual neural network (ResNet) and dense convolutional network (DenseNet) were combined as the feature extraction network of the original model that the precision and recall rate have reached 97.31 and 95.70%, respectively ( Jia et al, 2020b ). Li proposed an ensemble U-Net segmentation model, and the high-level semantic features of U-Net and the edge features of Edge were integrated to retain multi-scale contextual information and realize efficient segmentation of target fruit ( Li et al, 2021 ), where the recognition rate reached 95.11% and the recognition speed was 0.39 s. A modified YOLOv3 model based on clustering optimization is designed, and the influence of front-lighting and backlighting is clarified to detect and recognize banana fruits, inflorescence axes, and flower buds by Wu et al (2021) . To recognize and detect plant diseases, Chen et al (2022) proposed an improved plant disease-recognition model based on the YOLOv5 network model via a new involution bottleneck module, an SE module, and an efficient intersection over union loss function to optimize the performance of target detection, where mean average precision reached 70%.…”

Section: Introductionmentioning

confidence: 99%

Fast Location and Recognition of Green Apple Based on RGB-D Image

Sun

Luo

et al. 2022

Front. Plant Sci.

Self Cite

View full text Add to dashboard Cite

In the process of green apple harvesting or yield estimation, affected by the factors, such as fruit color, light, and orchard environment, the accurate recognition and fast location of the target fruit brings tremendous challenges to the vision system. In this article, we improve a density peak cluster segmentation algorithm for RGB images with the help of a gradient field of depth images to locate and recognize target fruit. Specifically, the image depth information is adopted to analyze the gradient field of the target image. The vorticity center and two-dimensional plane projection are constructed to realize the accurate center location. Next, an optimized density peak clustering algorithm is applied to segment the target image, where a kernel density estimation is utilized to optimize the segmentation algorithm, and a double sort algorithm is applied to efficiently obtain the accurate segmentation area of the target image. Finally, the segmentation area with the circle center is the target fruit area, and the maximum value method is employed to determine the radius. The above two results are merged to achieve the contour fitting of the target fruits. The novel method is designed without iteration, classifier, and several samples, which has greatly improved operating efficiency. The experimental results show that the presented method significantly improves accuracy and efficiency. Meanwhile, this new method deserves further promotion.

show abstract

“…The proposed fusion is effective in the system structure and the training process on classification results ( Woźniak and Połap, 2018 ). Li et al (2021) optimized the U-Net model by combining the spatial pyramid pooling (ASPP) structure and merging U-Net’s edge features and advanced functions. In addition, this model obtained the semantic boundary information of object fruit images by integrating the residual module and closed convolution, which effectively improved the segmentation accuracy of the object fruit ( Li et al, 2021 ).…”

Section: Introductionmentioning

confidence: 99%

“… Li et al (2021) optimized the U-Net model by combining the spatial pyramid pooling (ASPP) structure and merging U-Net’s edge features and advanced functions. In addition, this model obtained the semantic boundary information of object fruit images by integrating the residual module and closed convolution, which effectively improved the segmentation accuracy of the object fruit ( Li et al, 2021 ). Xiong et al (2020) used the YOLOv2 model to detect green mango images based on mango images collected by UAVs, and the detection accuracy is only 1.1% compared with the manual measurement error.…”

Section: Introductionmentioning

confidence: 99%

YOLOF-Snake: An Efficient Segmentation Model for Green Object Fruit

et al. 2022

View full text Add to dashboard Cite

Accurate detection and segmentation of the object fruit is the key part of orchard production measurement and automated picking. Affected by light, weather, and operating angle, it brings new challenges to the efficient and accurate detection and segmentation of the green object fruit under complex orchard backgrounds. For the green fruit segmentation, an efficient YOLOF-snake segmentation model is proposed. First, the ResNet101 structure is adopted as the backbone network to achieve feature extraction of the green object fruit. Then, the C5 feature maps are expanded with receptive fields and the decoder is used for classification and regression. Besides, the center point in the regression box is employed to get a diamond-shaped structure and fed into an additional Deep-snake network, which is adjusted to the contours of the target fruit to achieve fast and accurate segmentation of green fruit. The experimental results show that YOLOF-snake is sensitive to the green fruit, and the segmentation accuracy and efficiency are significantly improved. The proposed model can effectively extend the application of agricultural equipment and provide theoretical references for other fruits and vegetable segmentation.

show abstract

A novel green apple segmentation algorithm based on ensemble U-Net under complex orchard environment

Cited by 75 publications

References 32 publications

Plant Disease Recognition Model Based on Improved YOLOv5

Plant Disease Recognition Model Based on Improved YOLOv5

Fast Location and Recognition of Green Apple Based on RGB-D Image

YOLOF-Snake: An Efficient Segmentation Model for Green Object Fruit

Contact Info

Product

Resources

About