A CNN-based lightweight ensemble model for detecting defective carrots

Carrot grading plays a crucial role in producing high-competitive carrot products. However, carrot grading mainly depends on manual work nowadays, and this unreliable situation leads to high labor consumption, low efficiency, and unstable standard. In this study, a lightweight model (CarrotNet) based on machine vision with DCNN was proposed with the inspiration of several classic CNNs. The optimizations were conducted for the influential hyper-parameters in CarrotNet through comparative analysis. And the partial layers of the network were removed by the ablation study to obtain a more efficient structure. Furthermore, ensemble learning was adopted in the model to further improve the model accuracy. In the test set, the proposed model CarrotNet gained an accuracy of 97.04%, the modeling time of 1.42 h, the model size of 8.18 MB, and the detection speed of about 80 images per second. The robust performance of CarrotNet in the carrot dataset indicates that it can be used in on-line detection and grading of carrot external quality.

[1]  Forrest N. Iandola,et al.  SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <1MB model size , 2016, ArXiv.

[2]  Di Wang,et al.  Improved Deep CNN with Parameter Initialization for Data Analysis of Near-Infrared Spectroscopy Sensors , 2020, Sensors.

[3]  Weijun Xie,et al.  Research on Carrot Grading Based on Machine Vision Feature Parameters , 2019, IFAC-PapersOnLine.

[4]  Deng Limiao,et al.  A Carrot Sorting System Using Machine Vision Technique , 2019, Computer Vision-Based Agriculture Engineering.

[5]  Shifeng Chen,et al.  Salient Object Detection via Random Forest , 2014, IEEE Signal Processing Letters.

[6]  U. Mn.  Image processing method for detection of carrot green-shoulder, fibrous roots and surface cracks , 2013 .

[7]  Stephen W. Searcy,et al.  Estimation of tip shape for carrot classification by machine vision , 1992 .

[8]  Xiong Zhang,et al.  Online Sorting of the Film on Cotton Based on Deep Learning and Hyperspectral Imaging , 2020, IEEE Access.

[9]  Quoc V. Le,et al.  Searching for Activation Functions , 2018, arXiv.

[10]  Sam Kwong,et al.  G-MS2F: GoogLeNet based multi-stage feature fusion of deep CNN for scene recognition , 2017, Neurocomputing.

[11]  Rui Gao,et al.  Bayesian network classifiers based on Gaussian kernel density , 2016, Expert Syst. Appl..

[12]  Jian Sun,et al.  Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification , 2015, 2015 IEEE International Conference on Computer Vision (ICCV).

[13]  F. Hahn,et al.  Carrot Volume Evaluation using Imaging Algorithms , 2000 .

[14]  Mohammed Eunus Ali,et al.  Identification and recognition of rice diseases and pests using convolutional neural networks , 2018, Biosystems Engineering.

[15]  Stephen W. Searcy,et al.  Fresh market carrot inspection by machine vision , 1991, Other Conferences.

[16]  S. W. Searcy,et al.  REFLECTANCE CHARACTERISTICS OF FRESH-MARKET CARROTS , 1990 .

[17]  Stephen W. Searcy,et al.  Computer vision determination of the stem/root joint on processing carrots , 1989 .

[18]  Giles M. Foody,et al.  Multiclass and Binary SVM Classification: Implications for Training and Classification Users , 2008, IEEE Geoscience and Remote Sensing Letters.

[19]  Andreas Kamilaris,et al.  Deep learning in agriculture: A survey , 2018, Comput. Electron. Agric..

[20]  H. Ali,et al.  Symptom based automated detection of citrus diseases using color histogram and textural descriptors , 2017, Comput. Electron. Agric..

[21]  Zhao-qi Zheng,et al.  Recognition of Defective Carrots Based on Deep Learning and Transfer Learning , 2021, Food and Bioprocess Technology.

[22]  P. Vithu,et al.  Machine vision system for food grain quality evaluation: A review , 2016 .

[23]  Lei Song,et al.  The Rotate Stress of Steam Turbine Prediction Method Based on Stacking Ensemble Learning , 2019, 2019 IEEE 19th International Symposium on High Assurance Systems Engineering (HASE).

[24]  Geoffrey E. Hinton,et al.  ImageNet classification with deep convolutional neural networks , 2012, Commun. ACM.

[25]  Hu Wang,et al.  Motion Estimation From Surface Electromyogram Using Adaboost Regression and Average Feature Values , 2019, IEEE Access.

[26]  Yang Tao,et al.  Automatic inspection machine for maize kernels based on deep convolutional neural networks , 2019, Biosystems Engineering.

[27]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[28]  Xiaofeng Zhu,et al.  Efficient kNN Classification With Different Numbers of Nearest Neighbors , 2018, IEEE Transactions on Neural Networks and Learning Systems.

[29]  Arthur Zanatta Da Costa,et al.  Computer vision based detection of external defects on tomatoes using deep learning , 2020 .

[30]  Sepp Hochreiter,et al.  Fast and Accurate Deep Network Learning by Exponential Linear Units (ELUs) , 2015, ICLR.

[31]  Guy Shani,et al.  Using Deep Learning for Image-Based Potato Tuber Disease Detection. , 2019, Phytopathology.

[32]  F. Barba,et al.  High pressure processing of carrot juice: Effect of static and multi-pulsed pressure on the polyphenolic profile, oxidoreductases activity and colour. , 2020, Food chemistry.

[33]  Deyong Yang,et al.  Research on Carrot Surface Defect Detection Methods Based on Machine Vision , 2019, IFAC-PapersOnLine.

[34]  Oyebade K. Oyedotun,et al.  Intelligent Grading System for Banana Fruit Using Neural Network Arbitration , 2017 .

[35]  Zihao Liu,et al.  Soft-shell Shrimp Recognition Based on an Improved AlexNet for Quality Evaluations , 2020 .

[36]  Deng Limiao,et al.  Identifying carrot appearance quality by transfer learning , 2019, Journal of Food Process Engineering.

[37]  Andrew Zisserman,et al.  Very Deep Convolutional Networks for Large-Scale Image Recognition , 2014, ICLR.

[38]  David H. Wolpert,et al.  Stacked generalization , 1992, Neural Networks.