Deep learning-based automatic detection of productive tillers in rice

Abstract The number of productive tillers per plant is one of the important agronomic traits associated with the grain yield of rice crop. However, manual counting of productive tillers is time-consuming, laborious and error-prone. In this study, a method for automatically detecting and counting productive tillers of rice crop was proposed based on deep learning convolutional neural network (CNN). The CNN model was trained using large amounts of in-field images taken by mobile phones from various varieties of rice crops under various environmental conditions. A Web app, integrating the trained CNN model and a Django server, was designed for fast and high-throughput detection of productive tillers. The performance of the Web app was evaluated for field-based practical applications. Results showed that the selected CNN model had a high precision and a fast detection rate. Through applying the Web app to 200 in-field images with 5 to 30 tillers per image, the number of productive tillers detected agreed well with manual counting data, regardless of rice variety or type of mobile phone used for image taking. The coefficients of determination between the Web app detection and manual counting of tillers were over 0.97 in all cases. Overall, compared to the manual counting, the accuracy of the Web app was over 99%. Furthermore, the performance of the Web app was not affected by the environmental conditions, such as illumination condition (cloudy or sunny) and water reflection in paddy fields.

[1]  Jun Hu,et al.  Automated Counting of Rice Panicle by Applying Deep Learning Model to Images from Unmanned Aerial Vehicle Platform , 2019, Sensors.

[2]  Lifang Hu,et al.  Molecular dissection of developmental behavior of tiller number and the relationship with effective panicle using indica–japonica introgression lines in rice , 2015, Molecular Breeding.

[3]  Stefano Bianchi,et al.  Seed-per-pod estimation for plant breeding using deep learning , 2018, Comput. Electron. Agric..

[4]  H. K. Mebatsion,et al.  A Fourier analysis based algorithm to separate touching kernels in digital images , 2011 .

[5]  Kaiming He,et al.  Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks , 2015, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[6]  A. Anjos,et al.  P-TRAP: a Panicle Trait Phenotyping tool , 2013, BMC Plant Biology.

[7]  Gynheung An,et al.  Mutation of ONAC096 Enhances Grain Yield by Increasing Panicle Number and Delaying Leaf Senescence during Grain Filling in Rice , 2019, International journal of molecular sciences.

[8]  Wei Liu,et al.  Towards accurate tiny vehicle detection in complex scenes , 2019, Neurocomputing.

[9]  Dan Wu,et al.  Panicle-SEG: a robust image segmentation method for rice panicles in the field based on deep learning and superpixel optimization , 2017, Plant Methods.

[10]  F. Baret,et al.  Estimation of Plant and Canopy Architectural Traits Using the Digital Plant Phenotyping Platform1[OPEN] , 2019, Plant Physiology.

[11]  Hamid Laga,et al.  Detection and analysis of wheat spikes using Convolutional Neural Networks , 2018, Plant Methods.

[12]  Q. Luo,et al.  High-throughput measurement of rice tillers using a conveyor equipped with x-ray computed tomography. , 2011, The Review of scientific instruments.

[13]  S. P. Lanning,et al.  Identification of quantitative trait loci for productive tiller number and its relationship to agronomic traits in spring wheat , 2011, Theoretical and Applied Genetics.

[14]  Zhiguo Cao,et al.  TasselNet: counting maize tassels in the wild via local counts regression network , 2017, Plant Methods.

[15]  Liu Chengliang,et al.  Measurement of Rice Tillers Based on Magnetic Resonance Imaging , 2016 .

[16]  En Li,et al.  Apple detection during different growth stages in orchards using the improved YOLO-V3 model , 2019, Comput. Electron. Agric..

[17]  T. Głąb,et al.  Application of image analysis for grass tillering determination , 2015, Environmental Monitoring and Assessment.

[18]  Lisa M. Smith,et al.  A Rice Transcription Factor Controls Grain Length through Cell Number , 2019, Plant Physiology.

[19]  Won Suk Lee,et al.  Immature green citrus fruit detection using color and thermal images , 2018, Comput. Electron. Agric..

[20]  Wanneng Yang,et al.  Rice panicle length measuring system based on dual-camera imaging , 2013 .

[21]  T. Xuan,et al.  Efficacy of N-methyl-N-nitrosourea (MNU) Mutation on Enhancing the Yield and Quality of Rice , 2019, Agriculture.

[22]  Ye Zhang,et al.  A light and faster regional convolutional neural network for object detection in optical remote sensing images , 2018, ISPRS Journal of Photogrammetry and Remote Sensing.

[23]  Scott C. Chapman,et al.  Estimation of plant height using a high throughput phenotyping platform based on unmanned aerial vehicle and self-calibration: Example for sorghum breeding , 2018 .

[24]  Jeffrey W. White,et al.  Field-based phenomics for plant genetics research , 2012 .

[25]  Haiyang Wang,et al.  An evolutionarily conserved gene, FUWA, plays a role in determining panicle architecture, grain shape and grain weight in rice. , 2015, The Plant journal : for cell and molecular biology.

[26]  Meng Zhang,et al.  Development of a drought stress-resistant rice restorer line through Oryza sativa–rufipogon hybridization , 2019, Journal of Genetics.

[27]  J. Araus,et al.  Field high-throughput phenotyping: the new crop breeding frontier. , 2014, Trends in plant science.