Scholarlink

The use of drones for spraying has been gradually gaining popularity. Compared to traditional mechanical spraying methods, utilizing drones for spraying operations offers numerous advantages. This paper investigates the speed and patterns of droplet movement during drone spraying. The experiment is designed based on the PIV (Particle Image Velocimetry) principle. Initially, video images are collected and analyzed for speed b1y comparing data from smartphones, cameras, and high-speed cameras. The results indicate that the video quality of the PCO (Pioneering in Cameras and Optoelectronics) high-speed camera surpasses the other two methods in terms of frame clarity and completion of image capture. Subsequently, image preprocessing is conducted using techniques such as histogram equalization, Laplacian operator, logarithmic transformation, gamma transformation, Contrast Limited Adaptive Histogram Equalization (CLAHE), retinex-SSR (Single Scale Retinex), and retinex-MSR (Multi-Scale Retinex) to enhance image information expression. A comparative analysis is performed on image quality, robustness, and contrast to determine the selection of the Laplacian operator. Furthermore, Otsu’s method and morphological operations are employed for image processing.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Droplet Drift Image Detection Based on Dynamic Small Targets of Drones

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Yanwei Wang,
  • Wenbin Tao,
  • Chensheng Yang

摘要

The use of drones for spraying has been gradually gaining popularity. Compared to traditional mechanical spraying methods, utilizing drones for spraying operations offers numerous advantages. This paper investigates the speed and patterns of droplet movement during drone spraying. The experiment is designed based on the PIV (Particle Image Velocimetry) principle. Initially, video images are collected and analyzed for speed b1y comparing data from smartphones, cameras, and high-speed cameras. The results indicate that the video quality of the PCO (Pioneering in Cameras and Optoelectronics) high-speed camera surpasses the other two methods in terms of frame clarity and completion of image capture. Subsequently, image preprocessing is conducted using techniques such as histogram equalization, Laplacian operator, logarithmic transformation, gamma transformation, Contrast Limited Adaptive Histogram Equalization (CLAHE), retinex-SSR (Single Scale Retinex), and retinex-MSR (Multi-Scale Retinex) to enhance image information expression. A comparative analysis is performed on image quality, robustness, and contrast to determine the selection of the Laplacian operator. Furthermore, Otsu’s method and morphological operations are employed for image processing.