Fruit Detection Using CenterNet
In today's world, agriculture automation is becoming more and more important. This thesis is based on how to detect fruits from digital images. The relevant approaches are split into two parts, machine learning-based methods, and deep learning-based methods. After introducing those methods and comparing different deep learning-based methods, CenterNet is chosen as our model to settle this fruit detection problem. Three CenterNet models with different backbones were introduced, the backbones are ResNet-18, DLA-34, and Hourglass. A fruit dataset with four classes and 1,690 images was collected for this research work. By comparing those models with different backbone, according to the results, the deep learning-based model with DLA-34 was chosen as the final model to detect fruits from an image, the performance is excellent. In this thesis, the contribution is that we deploy a model based on CenterNet for object detection to settle the problem of fruit detection. Meanwhile, our dataset with four classes and 1,690 images were collected. By evaluating the performance of the model, we eventually design a CenterNet based on DLA-34 to detect multiclass fruits from our images. The performance of this method is better than the existing ones in fruit detection.