|
|
基于Faster RCNN模型的鸡胴体表皮破损缺陷高光谱图像检测方法研究
|
Abstract:
鸡胴体缺陷自动检测十分必要。针对鸡胴体表皮划伤、表皮剥落、断骨露出等体表破损类型缺陷难以通过机器视觉有效辨别的问题，提出一种高光谱技术结合深度学习的检测方法。先用竞争性自适应重加权算法(Competitive Adaptive Reweighted Sampling algorithm, CARS)对鸡胴体高光谱图像进行降维并提取特征波段。接着对所有特征波段图像进行主成分分析(Principal Component Analysis, PCA)，选贡献率前三的主成分图像合成假彩图并构造鸡胴体伪彩图数据集，构建基于Faster RCNN模型的鸡胴体缺陷检测方法，并利用鸡胴体数据集对模型进行训练和测试。结果表明，鸡胴体表皮划伤、表皮剥落、断骨露出三种缺陷检测的Precision分别为84.5%、83.7%、87.7%，模型的mAP为82.7%，鸡胴体表皮破损类缺陷的高光谱检测方法是可行的。
Automatic detection of chicken carcass defects is highly necessary. In response to the difficulty in effectively identifying surface damage types such as scratches, skin peeling, and exposed bone fractures on chicken carcasses through machine vision, a detection method combining hyperspectral technology with deep learning is proposed. Firstly, the Competitive Adaptive Reweighted Sampling algorithm (CARS) is used to reduce the dimensionality of hyperspectral images of chicken carcasses and extract feature bands. Then, principal component analysis (PCA) is applied to all feature band images, and the top three principal component images with the highest contribution rates are synthesized into pseudo-color images to construct a dataset for chicken carcass pseudo-color images. A Faster RCNN model for detecting defects in chicken carcasses is constructed and trained and tested using the chicken carcass dataset. The results indicate that the precision for detecting scratches, skin peeling, and exposed bone fractures on chicken carcasses are 84.5%, 83.7%, and 87.7% respectively. The model achieves an mAP of 82.7%. The hyperspectral detection method for skin damage defects on chicken carcasses is deemed feasible.
| [1] | Nyalala, I., Okinda, C., Makange, N., Korohou, T., Chao, Q., Nyalala, L. and Chao, L. (2021) On-Line Weight Estimation of Broiler Carcass and Cuts by a Computer Vision System. Poultry Science, 100, 101474. https://doi.org/10.1016/j.psj.2021.101474 |
| [2] | 吴江春, 王虎虎, 徐幸莲. 基于机器视觉的鸡胴体原发性皮炎快速检测[J]. 食品科学, 2023, 44(20): 350-356. |
| [3] | Okinda, C., Lu, M., Liu, L., Nyalala, I., Muneri, C., Wang, J. and Shen, M. (2019) A Machine Vision System for Early Detection and Prediction of Sick Birds: A Broiler Chicken Model. Biosystems Engineering, 188, 229-242. https://doi.org/10.1016/j.biosystemseng.2019.09.015 |
| [4] | Mortensen, A.K., Lisouski, P. and Ahrendt, P. (2016) Weight Prediction of Broiler Chickens Using 3D Computer Vision. Computers and Electronics in Agriculture, 123, 319-326. https://doi.org/10.1016/j.compag.2016.03.011 |
| [5] | Wu, X., Liang, X., Wang, Y., Wu, B. and Sun, J. (2022) Non-Destructive Techniques for the Analysis and Evaluation of Meat Quality and Safety: A Review. Foods, 11, Article 3713. https://doi.org/10.3390/foods11223713 |
| [6] | Signoroni, A., Savardi, M., Baronio, A. and Benini, S. (2019) Deep Learning Meets Hyperspectral Image Analysis: A Multidisciplinary Review. Journal of Imaging, 5, Article 52. https://doi.org/10.3390/jimaging5050052 |
| [7] | Kang, R., Yang, K., Zhang, X.X., Wu, W. and Chen, K.J. (2016) Development of Online Detection and Processing System for Contaminants on Chicken Carcass Surface. Applied Engineering in Agriculture, 32, 133-139. https://doi.org/10.13031/aea.32.11200 |
| [8] | Lim, J., Lee, A., Kang, J., Seo, Y., Kim, B., Kim, G. and Kim, S.M. (2020) Non-Destructive Detection of Bone Fragments Embedded in Meat Using Hyperspectral Reflectance Imaging Technique. Sensors, 20, Article 4038. https://doi.org/10.3390/s20144038 |
| [9] | LeCun, Y., Bengio, Y. and Hinton, G. (2015) Deep Learning. Nature, 521, 436-444. https://doi.org/10.1038/nature14539 |
| [10] | Al-Sarayreh, M.M. Reis, M., Qi Yan, W. and Klette, R. (2018) Detection of Red-Meat Adulteration by Deep Spectral—Spatial Features in Hyperspectral Images. Journal of Imaging, 4, Article 63. https://doi.org/10.3390/jimaging4050063 |
| [11] | Campos, R.L., Yoon, S.-C., Chung, S. and Bhandarkar, S.M. (2023) Semisupervised Deep Learning for the Detection of Foreign Materials on Poultry Meat with Near-Infrared Hyperspectral Imaging. Sensors, 23, Article 7014. https://doi.org/10.3390/s23167014 |
| [12] | Pu, H., Yu, J., Sun, D.-W., Wei, Q., Shen, X. and Wang, Z. (2023) Distinguishing Fresh and Frozen-Thawed Beef Using Hyperspectral Imaging Technology Combined with Convolutional Neural Networks. Microchemical Journal, 189, Article ID: 108559. https://doi.org/10.1016/j.microc.2023.108559 |
| [13] | Ye, M.C., Ji, C.X., Chen, H., Lei, L., Lu, H.J. and Qian, Y.T. (2020) Residual Deep PCA-Based Feature Extraction for Hyperspectral Image Classification. Neural Computing & Applications, 32, 14287-14300. . https://doi.org/10.1007/s00521-019-04503-3 |
| [14] | Khojastehnazhand, M., Khoshtaghaza, M.H., Mojaradi, B., Rezaei, M., Goodarzi, M. and Saeys, W. (2014) Comparison of Visible-Near Infrared and Short Wave Infrared Hyperspectral Imaging for the Evaluation of Rainbow Trout Freshness. Food Research International, 56, 25-34. https://doi.org/10.1016/j.foodres.2013.12.018 |
| [15] | Fan, N.Y., Liu, G.S., Zhang, J.J., Zhang, C., Yuan, R.R. and Ban, J.J. (2021) Hyperspectral Model Optimization for Protein of Tan Mutton Based on Box-Behnken. Spectroscopy and Spectral Analysis, 41, 918-923. |
| [16] | Wang, H.H., Zhang, S.L., Li, K., Cheng, S.S. and Zhang, X. (2017) Non-Destructive Detection of Ready-to-Eat Sea Cucumber Freshness Based on Hyperspectral Imaging. Spectroscopy and Spectral Analysis, 37, 3632-3640. |
