Opium Poppy Image Detection Based on Improved YOLOV5

In the past, the use of drones to collect poppy images and perform detection and recognition mainly relied on remote sensing technology, but the limitations of this method are the complex computational process and low recognition efficiency. In recent years, the rise of deep learning technology has greatly boosted the scope of image detection and recognition applications. However, there are certain challenges in applying deep learning to remote sensing images of poppies taken aerially by unmanned aerial vehicles(UAVs), including problems such as insufficient information on image details and interference from diverse planting backgrounds. Therefore, how to form an efficient poppy detection method has become an urgent and exploratory problem. In this paper, we take poppy image recognition in complex planting environments as the research object, based on the yolov5 model, we propose a multiscale attention mechanism applicable to small target detection, through the multiscale convolution operation in the channel dimension, the model can capture the information of multiple semantic levels at the same time, so as to better adapt to the target's changes in different scales and complexity, and then improve the accuracy of target detection.