Quality control of apples by means of convolutional neural networks - Comparison of bruise detection by color images and near-infrared images

Abstract Different stages of logistic process can affect apples, which often results in illusive bruises - making it extremely hard for the personal who has to sort out these fruits quickly before packing. We aim to provide a solution for bruise detection. In this contribution, we use state of the art convolutional neural network architectures for the task. Simultaneous input is taken from color CMOS and near-infrared (with a bandwidth filter) cameras, illuminated with a special light source. We achieved an accuracy above 97% for bruise detection in both cases: Colored and near-infrared images, indicating that both options are equally suitable.

[1]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[2]  Michael Freitag,et al.  Deep Learning-Based Pose Estimation of Apples for Inspection in Logistic Centers Using Single-Perspective Imaging , 2019, Processes.

[3]  Chu Zhang,et al.  Detection of Subtle Bruises on Winter Jujube Using Hyperspectral Imaging With Pixel-Wise Deep Learning Method , 2019, IEEE Access.

[4]  Kilian Q. Weinberger,et al.  Densely Connected Convolutional Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[5]  Li Fei-Fei,et al.  ImageNet: A large-scale hierarchical image database , 2009, CVPR.

[6]  Jeremy Howard,et al.  fastai: A Layered API for Deep Learning , 2020, Inf..

[7]  Ning Wang,et al.  Early detection of apple bruises on different background colors using hyperspectral imaging , 2008 .

[8]  Ning Wang,et al.  Bruise Detection of Apples Using Hyperspectral Imaging , 2010 .

[9]  R. Lu,et al.  Predicting bruise susceptibility of ‘Golden Delicious’ apples using hyperspectral scattering technique , 2016 .

[10]  Lei Zhang,et al.  Gradient Centralization: A New Optimization Technique for Deep Neural Networks , 2020, ECCV.

[11]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[12]  Hamid R. Arabnia,et al.  Apple classification based on surface bruises using image processing and neural networks , 2002 .

[13]  Kamran Rousta,et al.  Life cycle assessment of supermarket food waste , 2017 .

[14]  P. Baranowski,et al.  Detection of early bruises in apples using hyperspectral data and thermal imaging , 2012 .

[15]  Yuzhen Lu,et al.  Detection of Surface and Subsurface Defects of Apples Using Structured- Illumination Reflectance Imaging with Machine Learning Algorithms , 2018 .

[16]  G. M. Hyde,et al.  Non-contact bruise detection in apples by thermal imaging , 2003 .

[17]  Leslie N. Smith,et al.  A disciplined approach to neural network hyper-parameters: Part 1 - learning rate, batch size, momentum, and weight decay , 2018, ArXiv.

[18]  Wouter Saeys,et al.  Real-time pixel based early apple bruise detection using short wave infrared hyperspectral imaging in combination with calibration and glare correction techniques , 2016 .

[19]  Geoffrey E. Hinton,et al.  Rectified Linear Units Improve Restricted Boltzmann Machines , 2010, ICML.

[20]  S. Karthikeyan,et al.  Survey of Grading Process for Agricultural Foods by Using Artificial Intelligence Technique , 2018, 2018 Second International Conference on Electronics, Communication and Aerospace Technology (ICECA).

[21]  Wenqian Huang,et al.  Detection of early bruises on peaches (Amygdalus persica L.) using hyperspectral imaging coupled with improved watershed segmentation algorithm , 2018 .

[22]  Natalia Gimelshein,et al.  PyTorch: An Imperative Style, High-Performance Deep Learning Library , 2019, NeurIPS.

[23]  Wenqian Huang,et al.  Development of a multispectral imaging system for online detection of bruises on apples , 2015 .

[24]  Liyuan Liu,et al.  On the Variance of the Adaptive Learning Rate and Beyond , 2019, ICLR.

[25]  Zenghui Wang,et al.  Deep Convolutional Neural Networks for Image Classification: A Comprehensive Review , 2017, Neural Computation.

[26]  Geoffrey E. Hinton,et al.  Lookahead Optimizer: k steps forward, 1 step back , 2019, NeurIPS.

[27]  Karthik Kuchangi Jothi Prakash,et al.  Spoilage Detection in Raspberry Fruit Based on Spectral Imaging Using Convolutional Neural Networks , 2018 .

[28]  Ning Wang,et al.  Detecting chilling injury in Red Delicious apple using hyperspectral imaging and neural networks , 2009 .

[29]  Michael Freitag,et al.  Orientation detection of fruits by means of convolutional neural networks and laser line projection for the automation of fruit packing systems , 2020 .