基於神經網絡的端到端黃瓜自動化檢測系統

Abstract

精準農業和智慧農業近年來因耦合突破而變得越來越重要 機器視覺中的深度學習算法。本文旨在開發一種端到端的自動化農業食品 分級系統基於其視覺外觀。這裡考慮的目標對像是黃瓜，因為它是 世界上許多國家都可以種植的蔬菜。特別是，開發的系統包含 軟件和硬件組件，其中傳送帶上移動黃瓜的幾何特性 皮帶可以計算。具體來說，使用工業相機來捕捉黃瓜的圖像。那麼，三 執行黃瓜識別、幾何特性近似和缺陷的單個檢測系統 檢測，設計。最後，如果發現黃瓜有缺陷，PLC電機控制將被激活以分離 將黃瓜放入另一個容器中。結果，所提出的算法在以下情況下產生了有希望的性能 在一個自收集的數據集上進行實驗，即“Cuc-70”，它總共包含 4620 張圖像。黃瓜 識別產生的平均 WIoU 為 93%，體積逼近準確率為 98%，缺陷檢測 WIoU 為 92%。此外，還進行了綜合分析，以驗證所提出系統的穩健性和 所報告的定量和定性結果可以證明所執行的令人信服的績效。將來， 該系統可以集成到在線自動分揀和分級中，以實現有效的製造和生產。

Precision agriculture and smart farming have been gaining importance in recent years due to the coupled breakthrough of deep learning algorithms in machine vision. This paper aims to develop an end-to-end automatic agricultural food grading system based on its visual appearance. The target object considered herein is cucumber as it is one of the vegetables that can be grown in many countries around the world. Particularly, the developed system incorporates both the software and hardware components, in which the geometric properties of a moving cucumber on a conveyor belt can be computed. Concretely, an industrial camera is employed to capture the image of a cucumber. Then, three individual detection systems that perform the cucumber identification, geometry properties approximation, and defect detection, are designed. Finally, if the cucumber is found defective, the PLC motor control will be activated to separate the cucumber into an alternative container. As a result, the proposed algorithms yield promising performances when experimenting on a self-collected dataset, namely “Cuc-70” that consists of a total of 4620 images. The cucumber identification generates an average WIoU of 93%, volume approximation accuracy of 98%, and defect detection WIoU of 92%. In addition, comprehensive analysis is conducted in order to validate the robustness of the proposed system and the compelling performance executed can be evidenced from the quantitative and qualitative results reported. In the future, this system can be integrated into online automatic sorting and grading for effective manufacturing and production.