AUTHOR=Lu Wei , Du Rongting , Niu Pengshuai , Xing Guangnan , Luo Hui , Deng Yiming , Shu Lei TITLE=Soybean Yield Preharvest Prediction Based on Bean Pods and Leaves Image Recognition Using Deep Learning Neural Network Combined With GRNN JOURNAL=Frontiers in Plant Science VOLUME=Volume 12 - 2021 YEAR=2022 URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2021.791256 DOI=10.3389/fpls.2021.791256 ISSN=1664-462X ABSTRACT=Soybean yield is a highly complex trait determined by multiple factors such as genotype, environment, and their interactions. The earlier the prediction during the growing season the better. Accurate soybean yield prediction is important for germplasm innovation and planting environment factors improvement. But until now, soybean yield is determined by weight measurement manually after soybean plant harvest which is time consumption, high cost and low precision. The paper proposed a soybean yield in-field prediction method based on bean pods and leaves image recognition using deep learning algorithm combined with generalized regression neural network (GRNN). Faster Region- Convolutional Neural Networks (R-CNN), Feature Pyramid Network (FPN), Single Shot MultiBox Detector (SSD) and You only look once (YOLOv3) were employed for bean pods recognition which recognition precision and speed were 86.2%, 89.8%, 80.1%, 87.4%, 13 Frames Per Second (FPS), 7 FPS, 24 FPS and 39 FPS, respectively. Therefore, YOLOv3 was selected considering both recognition precision and speed. For enhancing detection performance, YOLOv3 was improved by changing IoU loss function, anchor frame clustering algorithm and partial neural network structure which recognition precision increased to 90.3%. In order to improve soybean yield prediction precision, leaves were identified and counted, moreover, pods were further classified as single, double, treble, four and five seeds types by improved YOLOv3 because each type seed weight varies. In addition, soybean seed number prediction models of each soybean planter were built using PLSR, BP and GRNN with the input of different type pod numbers and leave number which prediction results were 96.24%, 96.97% and 97.5%, respectively. Finally, soybean yield of each planter was obtained by accumulating weight of all soybean pod types and the average accuracy is up to 97.43%. The results show that it’s feasible to predict soybean yield of plant in-situ in high precision by fusing the number of leaves and different type soybean pods recognized by deep neural network combined with GRNN which can speed up germplasm innovation and planting environmental factors optimization.