A New Pest Detection Method Based on Improved YOLOv5m

Dai, Min; Dorjoy, Md Mehedi Hassan; Miao, Hong; Zhang, Shanwen

doi:10.3390/insects14010054

Cited by 44 publications

(15 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Specifically, YOLOv5 models excelled in localizing lesions within frames, with the YOLOv5m model performing the best with an mAP@0.5 of 0.92 and mAP@0.5:0.95::0.05 of 0.48. It strikes a balance between speed and accuracy, making it suitable for real-time applications where both these factors are crucial, as seen in [ 26 , 27 , 28 ].…”

Section: Discussionmentioning

confidence: 99%

Determining the Presence and Size of Shoulder Lesions in Sows Using Computer Vision

Bery,

Brown-Brandl,

Jones

et al. 2023

Animals

View full text Add to dashboard Cite

Shoulder sores predominantly arise in breeding sows and often result in untimely culling. Reported prevalence rates vary significantly, spanning between 5% and 50% depending upon the type of crate flooring inside a farm, the animal’s body condition, or an existing injury that causes lameness. These lesions represent not only a welfare concern but also have an economic impact due to the labor needed for treatment and medication. The objective of this study was to evaluate the use of computer vision techniques in detecting and determining the size of shoulder lesions. A Microsoft Kinect V2 camera captured the top-down depth and RGB images of sows in farrowing crates. The RGB images were collected at a resolution of 1920 × 1080. To ensure the best view of the lesions, images were selected with sows lying on their right and left sides with all legs extended. A total of 824 RGB images from 70 sows with lesions at various stages of development were identified and annotated. Three deep learning-based object detection models, YOLOv5, YOLOv8, and Faster-RCNN, pre-trained with the COCO and ImageNet datasets, were implemented to localize the lesion area. YOLOv5 was the best predictor as it was able to detect lesions with an mAP@0.5 of 0.92. To estimate the lesion area, lesion pixel segmentation was carried out on the localized region using traditional image processing techniques like Otsu’s binarization and adaptive thresholding alongside DL-based segmentation models based on U-Net architecture. In conclusion, this study demonstrates the potential of computer vision techniques in effectively detecting and assessing the size of shoulder lesions in breeding sows, providing a promising avenue for improving sow welfare and reducing economic losses.

show abstract

Section: Discussionmentioning

confidence: 99%

Determining the Presence and Size of Shoulder Lesions in Sows Using Computer Vision

Bery,

Brown-Brandl,

Jones

et al. 2023

Animals

View full text Add to dashboard Cite

show abstract

“…Our technique is effective in performance and report measurements. Drawing upon an extensive review of relevant academic literature, a majority of the research conducted on images [28][29][30][31][32][33][34][35][36][37][38][39]50] indicates a lack of sufficient attention towards the establishment of reliable analytical techniques for prevention and control [40][41][42][43][44][45] and accurately identifying pests through the analysis of acoustic signals produced by these pests [46][47][48][49]. The summarized findings of these reviews can be found in Table 1.…”

Section: Literature Reviewmentioning

confidence: 99%

Multi-Features and Multi-Deep Learning Networks to identify, prevent and control pests in tremendous farm fields combining IoT and pests sound analysis

Ali,

Sharma,

Dhanaraj

2024

Preprint

View full text Add to dashboard Cite

The agriculture sectors, which account for approximately 50% of the worldwide economic production, are the fundamental cornerstone of each nation. The significance of precision agriculture cannot be understated in assessing crop conditions and identifying suitable treatments in response to diverse pest infestations. The conventional method of pest identification exhibits instability and yields subpar levels of forecast accuracy. Nevertheless, the monitoring techniques frequently exhibit invasiveness, require significant time and resources, and are susceptible to various biases. Numerous insect species can emit distinct sounds, which can be readily identified and recorded with minimal expense or exertion. Applying deep learning techniques enables the automated detection and classification of insect sounds derived from field recordings, hence facilitating the monitoring of biodiversity and the assessment of species distribution ranges. The current research introduces an innovative method for identifying and detecting pests through IoT-based computerized modules that employ an integrated deep-learning methodology using the dataset comprising audio recordings of insect sounds. This included techniques, the DTCDWT method, Blackman-Nuttall window, Savitzky-Golay filter, FFT, DFT, STFT, MFCC, BFCC, LFCC, acoustic detectors, and PID sensors. The proposed research integrated the MF-MDLNet to train, test, and validate data. 9,600 pest auditory sounds were examined to identify their unique characteristics and numerical properties. The recommended system designed and implemented the ultrasound generator, with a programmable frequency and control panel for preventing and controlling pests and a solar-charging system for supplying power to connected devices in the networks spanning large farming areas. The suggested approach attains an accuracy (99.82%), a sensitivity (99.94%), a specificity (99.86%), a recall (99.94%), an F1 score (99.89%), and a precision (99.96%). The findings of this study demonstrate a significant enhancement compared to previous scholarly investigations, including VGG 16, VOLOv5s, TSCNNA, YOLOv3, TrunkNet, DenseNet, and DCNN.

show abstract

“…Teng et al 6 proposed a combination of multi-scale super-resolution feature enhancement module named MSR and multi-location detection module named SI to detect 26 common pests, which improved the accuracy to 67.4%. Dai et al 7 proposed to introduce SWinTR and C3TR mechanisms into the YOLOv5m network to capture more global features and increase the receiving domain of image features. Finally, the detection accuracy of 10 common insect pests, such as vertical rice leaf rollers and vertical rice leaf rollers, was achieved at 95.7%.…”

Section: Introductionmentioning

confidence: 99%

Research on lightweight target detection algorithm of farmland insect pests based on YOLO-PPLCBot

Lü

Zhou

2023

J. Electron. Imag.

View full text Add to dashboard Cite

In the agricultural field environment, in order to meet the strict requirements of modern intelligent agriculture on the accuracy and speed of pest detection algorithms, we proposed a lightweight farmland pest detection algorithm based on an improved YOLO v5. First, we constructed a new backbone by combining PPLCNet and BotNet to reduce network parameters and enhance the ability of image feature extraction. Second, the soft non-maximum suppression (Soft-NMS) algorithm is used instead of the NMS algorithm to improve the detection of target feature areas and minimize duplicate detection of the same areas. Then, the scylla intersection over union is applied to replace the original loss function complete intersection over union in the model to obtain higher quality anchors. Besides, mosaic data enhancement, scaling, translation, rotation, and color transformation methods are used to enhance the data volume, and letterbox is employed to unify data set sizes. Finally, we performed performance analysis experiments on a dataset containing 15 species of agricultural pests. From a large number of experimental results, we can conclude that our model improves the precision by 4.3% compared with YOLOv5n, while reducing the reasoning speed and model size by 10.1% and 50%. Furthermore, compared with faster R-CNN, cascade R-CNN, CenterNet, RetinaNet, YOLOv3, YOLOv4, YOLOv5, and YOLOx, our model shows the highest detection accuracy and the fastest reasoning speed. The

show abstract

A New Pest Detection Method Based on Improved YOLOv5m

Cited by 44 publications

References 34 publications

Determining the Presence and Size of Shoulder Lesions in Sows Using Computer Vision

Determining the Presence and Size of Shoulder Lesions in Sows Using Computer Vision

Multi-Features and Multi-Deep Learning Networks to identify, prevent and control pests in tremendous farm fields combining IoT and pests sound analysis

Research on lightweight target detection algorithm of farmland insect pests based on YOLO-PPLCBot

Contact Info

Product

Resources

About