Academic Journal of Engineering and Technology Science, 2024, 7(3); doi: 10.25236/AJETS.2024.070306.
Chen Jiyao1,2, Long Dongping1,2, Yang Shaozhou1,2
1Hunan Provincial Key Laboratory of High Efficiency and Precision Machining of Difficult-to-Cut Material, Hunan University of Science and Technology, Xiangtan, Hunan, China
2School of Mechanical Engineering, Hunan University of Science and Technology, Xiangtan, Hunan, China
China as the world's leading corn producer, consumer and exporter, corn peeling corn mechanized harvesting has become the main development direction of China's corn production. Due to the low bract stripping rate during the mechanical peeling process of the corn harvester, too many bracts remain on the corn ears, making it impossible for the moisture within the corn ears to be discharged in a timely manner, which can easily lead to bacterial infection and mold growth. In addition, due to the high moisture content of corn ears in China, corn ears are susceptible to mold and mildew during storage, leading to a decline in corn quality. In summary, the stripping device of the corn ears harvester has outstanding problems such as low bract stripping rate and high damage rate of corn kernels, which seriously affects the development of China's corn industry and the realization of the goal of increasing production and income. Solving the problem of sorting the residual corn (unpeeled bracts and damaged corn kernels) after peeling is a key technical problem that needs to be solved urgently. To address the above problems this paper proposes an improved corn cob defect detection algorithm for YOLOv7. Firstly, the Explicit Visual Center Block (EVCBlock) is introduced into the head network, which improves the model's ability of recognizing the small target of corn kernel breakage. Then the Receptive field enhancement module (RFEM) is introduced to enhance the feature pyramid's ability to extract defective features of corn cob. The experiments show that the mean average accuracy of the improved YOLOv7 model is 88.1%, which is 12.2 percentage points higher than that of the original YOLOv7 model, realizing the need for more accurate corn cob defect detection.
Defect detection; YOLOv7; Deep learning; Corn ears; Object detection
Chen Jiyao, Long Dongping, Yang Shaozhou. Research on corn ears defect detection algorithm based on improved YOLOv7. Academic Journal of Engineering and Technology Science (2024) Vol. 7, Issue 3: 39-47. https://doi.org/10.25236/AJETS.2024.070306.
[1] Shi J,et al.Study on a detection method for crop diseases and insect pests based on YOLO v5s improved model[J].Jiangsu Agricultural Sciences,2023,51(24):175-183.
[2] Girshick B R ,Donahue J ,Darrell T , et al. Rich feature hierarchies for accurate object detection and semantic segmentation. [J]. CoRR, 2013, abs/1311.2524.
[3] Girshick B R . Fast R-CNN. [J]. CoRR, 2015, abs/1504.08083.
[4] Shaoqing R ,Kaiming H ,Ross G , et al. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. [J]. IEEE transactions on pattern analysis and machine intelligence, 2017, 39 (6): 1137-1149.
[5] Liu W ,Anguelov D ,Erhan D , et al. SSD: Single Shot MultiBox Detector. [J]. CoRR, 2015, abs/1512.02325.
[6] Tsung-Yi L ,Priya G ,Ross G , et al. Focal Loss for Dense Object Detection. [J]. IEEE transactions on pattern analysis and machine intelligence, 2020, 42 (2): 318-327.
[7] Zhou X ,Wang D ,Krähenbühl P . Objects as Points. [J]. CoRR, 2019, abs/1904.07850.
[8] M. T ,R. P ,Q.V. L . EfficientDet: Scalable and efficient object detection [J]. Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2020, 10778-10787.
[9] Tian Z ,Shen C ,Chen H , et al. FCOS: Fully Convolutional One-Stage Object Detection. [J]. CoRR, 2019, abs/1904.01355
[10] Redmon J ,Farhadi A . YOLO9000: Better, Faster, Stronger. [J]. CoRR, 2016, abs/1612.08242.
[11] REDMON J,FARHADI A.Yolov3:an incremental improve-ment[J].arXiv:1804.02767,2018.
[12] BOCHKOVSKIV A,WANG C Y,LIAO H Y M.YOLOv4:optimalspeed and accuracy of object detection[J].ArXiv Preprint arXiv,2020,2004:10934.
[13] GLENN J.YOLOv5[EB/OL].(2020).https://github.com/ultralytics/yolov5.
[14] WANG C Y, BOCHKOVSKIY A, LIAO H Y M. YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[EB/OL]. (2022-06-06)[2023-08-06].http://arxiv.org/ abs/2207. 02696.
[15] Zhang F,et al.Screening Method of Abnormal Corn Ears Based on Machine Vision[J].Transactions of the Chinese Society for Agricultural Machinery,2015,46(S1):45-49.
[16] MENG F,et al.Design and Experiment of Real-time Detection and Sorting Device for Maize Seeds[J].Transactions of the Chinese Society for Agricultural Machinery,2021,52(3):153-159,177.
[17] LI Q,WANG K,QIANG H,et al.Classification and recognition method of abnormal corn ears based on color and texture features[J].Jiangsu Journal of Agricultural Sciences,2020,(01):24-31.[doi:doi:10. 3969/j.issn.1000-4440.2020.01.004].
[18] Yu Q ,Dong Z ,Liyan Z , et al. Centralized Feature Pyramid for Object Detection. [J]. IEEE transactions on image processing : a publication of the IEEE Signal Processing Society, 2023, PP.
[19] Yu Z,Huang H,Chen W,et al.YOLO-FaceV2:A Scale and Occlusion Aware Face Detector[J].arXiv preprint arXiv:2208.02019,2022.