Defect detection is the most important step in the postpartum reprocessing of kiwifruit. However, there are some small defects difficult to detect. The accuracy and speed of existing detection algorithms are difficult to meet the requirements of real-time detection. For solving these problems, we developed a defect detection model based on YOLOv5, which is able to detect defects accurately and at a fast speed. The main contributions of this research are as follows: (1) a small object detection layer is added to improve the model’s ability to detect small defects; (2) we pay attention to the importance of different channels by embedding SELayer; (3) the loss function CIoU is introduced to make the regression more accurate; (4) under the prerequisite of no increase in training cost, we train our model based on transfer learning and use the CosineAnnealing algorithm to improve the effect. The results of the experiment show that the overall performance of the improved network YOLOv5-Ours is better than the original and mainstream detection algorithms. The mAP@0.5 of YOLOv5-Ours has reached 94.7%, which was an improvement of nearly 9%, compared to the original algorithm. Our model only takes 0.1 s to detect a single image, which proves the effectiveness of the model. Therefore, YOLOv5-Ours can well meet the requirements of real-time detection and provides a robust strategy for the kiwi flaw detection system.

No abstract

Imagers with pre-processing functions, such as image recognition and classification, contrast enhancement, and noise reduction, play a critical role in the neuromorphic visual system. Optoelectronic plasticity is a prerequisite to achieve these functions. In this study, we demonstrate a nonvolatile reconfigurable broadband photodetector based on a ferroelectric heterostructure composed of BP (black phosphorus)/α-In2Se3. The plasticity of the device comes from the ferroelectric polarization of α-In2Se3 that can tune the built-in potential of the p–n junction. As a result, the rectification ratio and responsivity increase almost one order when changing the gate voltage pulse from +16 V to −16 V. Due to the introduction of BP, the device has a wide spectral response covering 473–1550 nm. In addition, our devices show excellent performance in terms of a high responsivity of up to 4.73 × 104 A/W, a large specific detectivity of ∼2.09 × 1012 Jones, a high external quantum efficiency of 9.21 × 106%, and a notable photo-on-off ratio of 4.82 × 103. Due to its high performance, reconfigurability, and broadband response, our device shows considerable potential in neuromorphic visual systems even in the infrared region.