Target recognition method of green pepper harvesting robot based on manifold ranking

Ji, Wei; Gao, Xiaoxiao; Xu, Bo; Chen, Guangyu; Zhao, Debin

doi:10.1016/j.compag.2020.105663

Cited by 31 publications

(10 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…When there is no target in the grid point, but In this article, the precision rate (P), recall rate (R), F1 value and average precision (AP) are used as the evaluation indexes of the model performance after training. The calculation formula is defined as follows (Cao et al, 2020;Ji, Gao, Xu, Chen, & Zhao, 2020;Sun et al, 2018).…”

Section: Loss Functionmentioning

confidence: 99%

Apple target recognition method in complex environment based on improved YOLOv4

Gao

et al. 2021

J Food Process Engineering

Self Cite

View full text Add to dashboard Cite

In order to solve the problem of accurate recognition of apples in complex environments, this article proposes an apple recognition method based on improved YOLOv4, which can accurately locate and recognize apples in a variety of complex environments. This method uses the lightweight EfficientNet‐B0 network as the feature extraction network for apple recognition and then combines with the PANet (Path Aggregation Network) network to fuse the features in the adjacent feature layers, which improves the recognition accuracy of fruit targets. In the experiment, the apples under bagging, nonbagging and night environment are identified. The results of the experiment show that the average precision of the improved YOLOv4 method is 93.42%, the recall rate is 87.64%, and the F1 value is 0.9035, which is about 2% lower than that of the original YOLOv4. However, the storage memory required for the model trained on the improved YOLOv4 is decreased by 87.8% compared with the original YOLOv4, which is only 29.8 MB, and the recognition speed is increased by 43%, reaching 63.20 frames per second. The recognition accuracy and speed are much higher than that of the two‐stage Faster‐RCNN. The recognition accuracy is equivalent to that of Mobilenetv3‐YOLOv4, but the model of our method is smaller and the recognition speed is faster. Practical Applications Achieving rapid and accurate recognition of apples in complex environments is one of the key technologies for automatic fruit picking by picking robots, and it is also an important method and means for realizing orchard yield estimation. In order to recognize apples in the orchard environment more quickly and effectively, a light‐weight YOLOv4 target recognition method based on deep learning is proposed. The method proposed in this article can significantly reduce the storage space and computing power requirements of the device. On the basis of maintaining high accuracy, it can achieve real‐time and accurate recognition of apples in complex environments, and has great application value and practical significance for picking robot.

show abstract

Section: Loss Functionmentioning

confidence: 99%

Apple target recognition method in complex environment based on improved YOLOv4

Gao

et al. 2021

J Food Process Engineering

Self Cite

View full text Add to dashboard Cite

show abstract

“…Therefore, by employing the Schur complement, it can be found that there exists a scalar γ > 0 for (ρ) < 0, such that by integrating (ρ) < 0 between 0 and T with zero initial condition, J ≥ γ T 0 T (t) (t)dt holds. This means that the dissipative performance can be satisfied according to Definition 1 and the fault detection problem can be solved by setting appropriate fault threshold by (8). Therefore, we can finish the proof.…”

Section: Resultsmentioning

confidence: 76%

Mode-Dependent Event-Triggered Fault Detection for Nonlinear Semi-Markov Jump Systems With Quantization: Application to Robotic Manipulator

Wang

2021

IEEE Access

View full text Add to dashboard Cite

This paper is studied with fault detection issue for nonlinear semi-Markov jump systems. In particular, the mode-dependent mechanism of event-triggered transmission is developed for improving communication efficiency. Furthermore, the mode-dependent quantization method is utilized to cope with the limited network bandwidth. These strategies are chosen for more effective utilization of system mode information with less conservatism. Mode-dependent filter type fault detectors are constructed and sufficient conditions are established by Lyapunov-Krasovskii functionals, such that the filter error system can achieve the desired dissipative performance. In the end, the designed fault detection method is verified through a practical example of robotic manipulator.INDEX TERMS Mode-dependent quantization, mode-dependent event-triggered, fault detection, semi-Markov jump system, robotic manipulator.

show abstract

“…Therefore, automated techniques used in the harvest and the post-harvest can be increasingly improved with the use of ML. Ji et al (2020) have referred to the difficulty of automating and mechanizing the harvest of green peppers due to the similarity of the fruit color with the background and its form with the shape of the leaf. It is supposed that ML's multidimensional view could provide better results.…”

Section: Discussionmentioning

confidence: 99%

Machine Learning for Automatic Classification of Tomato Ripening Stages Using Electrophysiological Recordings

Reissig

Oliveira

Costa

et al. 2021

Front. Sustain. Food Syst.

View full text Add to dashboard Cite

The physiological processes underlying fruit ripening can lead to different electrical signatures at each ripening stage, making it possible to classify tomato fruit through the analysis of electrical signals. Here, the electrical activity of tomato fruit (Solanum lycopersicum var. cerasiforme) during ripening was investigated as tissue voltage variations, and Machine Learning (ML) techniques were used for the classification of different ripening stages. Tomato fruit was harvested at the mature green stage and placed in a Faraday's cage under laboratory-controlled conditions. Two electrodes per fruit were inserted 1 cm apart from each other. The measures were carried out continuously until the entire fruits reached the light red stage. The time series were analyzed by the following techniques: Fast Fourier Transform (FFT), Wavelet Transform, Power Spectral Density (PSD), and Approximate Entropy. Descriptive analysis from FFT, PSD, and Wavelet Transform were used for PCA (Principal Component Analysis). Finally, ApEn, PCA1, PCA2, and PCA3 were obtained. These features were used in ML analyses for looking for classifiable patterns of the three different ripening stages: mature green, breaker, and light red. The results showed that it is possible to classify the ripening stages using the fruit's electrical activity. It was also observed, using precision, sensitivity, and F1-score techniques, that the breaker stage was the most classifiable among all stages. It was found a more accurate distinction between mature green × breaker than between breaker × light red. The ML techniques used seem to be a novel tool for classifying ripening stages. The features obtained from electrophysiological time series have the potential to be used for supervised training, being able to help in more accurate classification of fruit ripening stages.

show abstract

Target recognition method of green pepper harvesting robot based on manifold ranking

Cited by 31 publications

References 24 publications

Apple target recognition method in complex environment based on improved YOLOv4

Apple target recognition method in complex environment based on improved YOLOv4

Mode-Dependent Event-Triggered Fault Detection for Nonlinear Semi-Markov Jump Systems With Quantization: Application to Robotic Manipulator

Machine Learning for Automatic Classification of Tomato Ripening Stages Using Electrophysiological Recordings

Contact Info

Product

Resources

About