Tree Trunk and Obstacle Detection in Apple Orchard Based on Improved YOLOv5s Model

Su, Fei; Zhao, Yanping; Shi, Yuankai; Zhao, Dong; Wang, Guanghui; Yan, Yinfa; Zu, Linlu; Chang, Siyuan

doi:10.3390/agronomy12102427

Cited by 16 publications

(7 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This proved that the proposed method was useful in increasing the blueberry recognition accuracy of the model. Su et al (2022) performed a tree trunk and obstacle detection method in a semistructured apple orchard environment based on the improved YOLOv5s to improve real-time detection performance.…”

Section: Discussionmentioning

confidence: 99%

Detection of artichoke on seedling based on YOLOV5 model

Kahya,

Aslan

2024

International Journal of Agriculture Environment and Food Sciences

View full text Add to dashboard Cite

Robotic systems have become essential in the industrial field today. Robotic systems used in many areas of industry enable the development of mechanization of agriculture. Researches in recent years have focused on the introduction of automatic systems and robot prototypes in the field of agriculture in order to reduce production costs. The developed smart harvest robots are systems that can work uninterrupted for hours and guarantee minimum cost and high production. The main element of these systems is the determination of the location of the product to be harvested by image processing. In addition to the programs used for image processing, deep learning models have become popular today. Deep learning techniques offer high accuracy in analyzing and processing agricultural data. Due to this feature, the use of deep learning techniques in agriculture is becoming increasingly widespread. During the harvest of the artichoke, its head should generally be cut off with one or two leaves. One main head and usually two side heads occur from one shoot. Harvest maturity degree is the time when the heads reach 2/3 of their size, depending on the variety character. In this study, classification was made by using the deep learning method, considering the head size of the fruit. YOLOv5 (nano-small-medium and large models) was used for the deep learning method. All metric values of the models were examined. It was observed that the most successful model was the model trained with the YOLOv5n algorithm, 640x640 sized images with 20 Batch, 90 Epoch. Model values results were examined as “metrics/precision”, “metrics/recall”, “metrics/mAP_0.5” and “metrics/mAP_0.5:0.95”. These are key metrics that measure the detection success of a model and indicate the performance of the relevant model on the validation dataset. It was determined that the metric data of the “YOLOv5 nano” model was higher compared to other models. The measured value was Model 1= Size: 640x640, Batch: 20, Epoch: 90, Algorithm: YOLOv5n. Hence, it was understood that “Model 1” was the best detection model to be used in separating artichokes from branches in robotic artichoke harvesting.

show abstract

Section: Discussionmentioning

confidence: 99%

Detection of artichoke on seedling based on YOLOV5 model

Kahya,

Aslan

2024

International Journal of Agriculture Environment and Food Sciences

View full text Add to dashboard Cite

show abstract

“…For instance, Singh et al [39] use YOLO object-detection algorithm results to aid the mobile robot in detecting obstacles and navigating in an indoor environment. Another example is provided by Su et al [40], where an improved YOLOv5s object detection method is used for a semi-structured apple environment.…”

Section: Related Workmentioning

confidence: 99%

Simultaneous Object Detection and Distance Estimation for Indoor Autonomous Vehicles

Azurmendi,

Zulueta,

Lopez-Guede

et al. 2023

Electronics

View full text Add to dashboard Cite

Object detection is an essential and impactful technology in various fields due to its ability to automatically locate and identify objects in images or videos. In addition, object-distance estimation is a fundamental problem in 3D vision and scene perception. In this paper, we propose a simultaneous object-detection and distance-estimation algorithm based on YOLOv5 for obstacle detection in indoor autonomous vehicles. This method estimates the distances to the desired obstacles using a single monocular camera that does not require calibration. On the one hand, we train the algorithm with the KITTI dataset, which is an autonomous driving vision dataset that provides labels for object detection and distance prediction. On the other hand, we collect and label 100 images from a custom environment. Then, we apply data augmentation and transfer learning to generate a fast, accurate, and cost-effective model for the custom environment. The results show a performance of mAP0.5:0.95 of more than 75% for object detection and 0.71 m of mean absolute error in distance prediction, which are easily scalable with the labeling of a larger amount of data. Finally, we compare our method with other similar state-of-the-art approaches.

show abstract

“…Concurrently, the burgeoning availability of cost-effective RGB-D sensors has stimulated interest in RGB-D-based object reconstruction [42] and tracking methodologies [43]. By enhancing the YOLOV5 framework, Su et al [44] enhanced tree trunk detection across a spectrum of seasonal conditions, with the position of targets being determined on the basis of depth data. However, in the working scenarios of rubber-tapping robots, these machines operate at night, rendering them unsuitable for trunk detection algorithms based on RGB images.…”

Section: Trunk Detectionmentioning

confidence: 99%

Active Navigation System for a Rubber-Tapping Robot Based on Trunk Detection

et al. 2023

View full text Add to dashboard Cite

To address the practical navigation issues of rubber-tapping robots, this paper proposes an active navigation system guided by trunk detection for a rubber-tapping robot. A tightly coupled sliding-window-based factor graph method is proposed for pose tracking, which introduces normal distribution transform (NDT) measurement factors, inertial measurement unit (IMU) pre-integration factors, and prior factors generated by sliding window marginalization. To actively pursue goals in navigation, a distance-adaptive Euclidean clustering method is utilized in conjunction with cylinder fitting and composite criteria screening to identify tree trunks. Additionally, a hybrid map navigation approach involving 3D point cloud map localization and 2D grid map planning is proposed to apply these methods to the robot. Experiments show that our pose-tracking approach obtains generally better performance in accuracy and robustness compared to existing methods. The precision of our trunk detection method is 93% and the recall is 87%. A practical validation is completed in robot rubber-tapping tasks of a real rubber plantation. The proposed method can guide the rubber-tapping robot in complex forest environments and improve efficiency.

show abstract

Tree Trunk and Obstacle Detection in Apple Orchard Based on Improved YOLOv5s Model

Cited by 16 publications

References 28 publications

Detection of artichoke on seedling based on YOLOV5 model

Detection of artichoke on seedling based on YOLOV5 model

Simultaneous Object Detection and Distance Estimation for Indoor Autonomous Vehicles

Active Navigation System for a Rubber-Tapping Robot Based on Trunk Detection

Contact Info

Product

Resources

About