MIoP-NMS: Perfecting crops target detection and counting in dense occlusion from high-resolution UAV imagery

Jiang, Quan; Huang, Zonghao; Xu, Gang; Su, Yingying

doi:10.1016/j.atech.2023.100226

Cited by 3 publications

(3 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Compared to satellite imagery, which is limited by weather conditions and spatial resolution, UAV remote sensing offers greater flexibility [55]. In the context of tree species classification, remote sensing-based semantic segmentation tasks have already been successfully used for large-scale mapping of fruit trees such as coconut palms [4], citrus [6], and bananas [56] from UAV-based RGB images. The combination of high-spectral and multi-spectral imagery with RGB imagery has not demonstrated significant advantages in many tasks [6,57].…”

Section: Semantic Segmentationmentioning

confidence: 99%

“…Combined with visible light (RGB) and multispectral data, as well as multitemporal analysis, a method was employed to detect chestnut vegetation coverage using a canopy height model (CHM) and vegetation index thresholds [58,59]. CHM determines the position of tree crowns based on height information and can detect the center point of chestnut tree crowns [56]. However, due to potential mismatches between the top point of chestnut trees and the center of their crowns, there may be slight displacements of the highest point relative to the center.…”

Section: Semantic Segmentationmentioning

confidence: 99%

See 1 more Smart Citation

Detection and Mapping of Chestnut Using Deep Learning from High-Resolution UAV-Based RGB Imagery

Sun,

Hao,

Guo

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

The semantic segmentation method based on high-resolution RGB images obtained by unmanned aerial vehicle (UAV) provides a cost-effective way to improve the accuracy of detection and classification in forestry. Few studies have explored the impact of sample distribution patterns on deep learning model detection accuracy. The study was carried out using the data from the 4.78 km2 RGB image of a chestnut (Castanea mollissima Blume) plantation obtained by the DJI Phantom 4-RTK, and the model training was conducted with 18,144 samples of manually delineated chestnut tree clusters. The performance of four semantic segmentation models (U-Net, DeepLab V3, PSPNet, and DeepLab V3+) paired with backbones (ResNet-34, ResNet-50) was evaluated. Then, the influence of chestnut data from different planting patterns on the accuracy and generalization performance of deep learning models was examined. The results showed that the combination of DeepLab V3 with ResNet-34 backbone gives the best performance (F1 score = 86.41%), while the combination of DeepLab V3+ with ResNet-50 backbone performed the worst. The influence of different backbone networks on the detection performance of semantic segmentation models did not show a clear pattern. Additionally, different spatial distribution patterns of chestnut planting affected the classification accuracy. The model MIX, trained on comprehensive training data, achieves higher classification accuracies (F1 score = 86.13%) compared to the model trained on single training data (F1 score (DP) = 82.46%; F1 score (SP) = 83.81%). The model performance in complex scenario data training is superior to that of the model in simple scene data training. In conclusion, comprehensive training databases can improve the generalization performance of chestnut classification with different spatial distribution patterns. This study provides an effective method for detecting chestnut cover area based on semantic segmentation, allowing for better quantitative evaluation of its resource utilization and further development of inventories for other tree species.

show abstract

Section: Semantic Segmentationmentioning

confidence: 99%

Section: Semantic Segmentationmentioning

confidence: 99%

Detection and Mapping of Chestnut Using Deep Learning from High-Resolution UAV-Based RGB Imagery

Sun,

Hao,

Guo

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…Zhang et al [19] introduced dilated convolution based on the Faster R-CNN model to solve the problem of small-sized rice panicle target and used ROIAign instead of ROIPooling to optimize and improve the average detection accuracy of the model for rice panicles. Jiang et al [20] proposed an improved NMS-based max intersection over portion (MIoP-NMS) algorithm and implemented it in the YOLOv4 network framework for singlestage target detection, and estimated the number of banana trees in dense occluded banana forests with about 98.7% accuracy. Bao et al [21] designed a lightweight convolutional neural network simple net, which is constructed using convolution and reverse residual blocks, and combined it with the convolutional attention mechanism CBAM module, which can be used for automatic recognition of wheat ear diseases on mobile terminals.…”

Section: Introductionmentioning

confidence: 99%

Research and Implementation of Millet Ear Detection Method Based on Lightweight YOLOv5

Qiu,

Li,

Gao

et al. 2023

Sensors

View full text Add to dashboard Cite

As the millet ears are dense, small in size, and serious occlusion in the complex grain field scene, the target detection model suitable for this environment requires high computing power, and it is difficult to deploy the real-time detection of millet ears on mobile devices. A lightweight real-time detection method for millet ears is based on YOLOv5. First, the YOLOv5s model is improved by replacing the YOLOv5s backbone feature extraction network with the MobilenetV3 lightweight model to reduce model size. Then, using the multi-feature fusion detection structure, the micro-scale detection layer is augmented to reduce high-level feature maps and low-level feature maps. The Merge-NMS technique is used in post-processing for target information loss to reduce the influence of boundary blur on the detection effect and increase the detection accuracy of small and obstructed targets. Finally, the models reconstructed by different improved methods are trained and tested on the self-built millet ear data set. The AP value of the improved model in this study reaches 97.78%, F1-score is 94.20%, and the model size is only 7.56 MB, which is 53.28% of the standard YoloV5s model size, and has a better detection speed. Compared with other classical target detection models, it shows strong robustness and generalization ability. The lightweight model performs better in the detection of pictures and videos in the Jetson Nano. The results show that the improved lightweight YOLOv5 millet detection model in this study can overcome the influence of complex environments, and significantly improve the detection effect of millet under dense distribution and occlusion conditions. The millet detection model is deployed on the Jetson Nano, and the millet detection system is implemented based on the PyQt5 framework. The detection accuracy and detection speed of the millet detection system can meet the actual needs of intelligent agricultural machinery equipment and has a good application prospect.

show abstract

Satin Bowerbird Optimization With Convolutional LSTM for Food Crop Classification on UAV Imagery

2023

View full text Add to dashboard Cite

Food crop classification and identification are crucial aspects of modern agriculture. With progression of drones or unmanned aerial vehicles (UAVs), crop detection from RGB images goes through a paradigm shift from traditional image processing methods to deep learning (DL) methods due to effective breakthroughs in convolutional neural networks (CNN). Drone images are reliable for identifying different crops because of its higher spatial resolution. Food crop classification utilizing deep learning on drone images includes machine learning techniques for distinguishing and identifying different types of crops in images captured by UAVs. It is beneficial for various applications, like crop monitoring and precision agriculture. This paper presents a new Satin Bowerbird Optimization with deep learning for Food Crop Classification (SBODL-FCC) technique on UAV images. The presented SBODL-FCC technique exploits DL models with hyperparameter optimizers for food crop classification on UAV images. To accomplish this, the presented SBODL-FCC technique employs adaptive bilateral filtering technique for image preprocessing. Besides, the SBODL-FCC technique uses MobileNetv2 feature extractor with Bayesian optimization (BO) algorithm for parameter optimization. Moreover, the food crop classification process is performed through convolutional long short-term memory (ConvLSTM) model. Furthermore, the hyperparameter tuning of the ConvLSTM method is accomplished through SBO algorithm. The experimental validation of the SBODL-FCC technique is validated on UAV image database and the results are inspected under different aspects. The simulation outcomes inferred that the SBODL-FCC technique reaches better performance over other models in terms of several performance measures.

show abstract

MIoP-NMS: Perfecting crops target detection and counting in dense occlusion from high-resolution UAV imagery

Cited by 3 publications

References 21 publications

Detection and Mapping of Chestnut Using Deep Learning from High-Resolution UAV-Based RGB Imagery

Detection and Mapping of Chestnut Using Deep Learning from High-Resolution UAV-Based RGB Imagery

Research and Implementation of Millet Ear Detection Method Based on Lightweight YOLOv5

Satin Bowerbird Optimization With Convolutional LSTM for Food Crop Classification on UAV Imagery

Contact Info

Product

Resources

About