S$^3$AD: Semi-supervised Small Apple Detection in Orchard Environments

TL;DR

This paper introduces S$^3$AD, a semi-supervised system for small apple detection in orchards, leveraging a new large dataset to improve accuracy and address challenges of small object detection in precision agriculture.

Contribution

The work presents a novel semi-supervised detection system and a large annotated dataset, significantly enhancing small apple detection performance over existing fully-supervised methods.

Findings

S$^3$AD outperforms fully-supervised baselines by up to 14.9%.

The dataset MAD contains 105 labeled and 4,440 unlabeled images with detailed annotations.

Analysis of apple size and occlusion effects on detection performance.

Abstract

Crop detection is integral for precision agriculture applications such as automated yield estimation or fruit picking. However, crop detection, e.g., apple detection in orchard environments remains challenging due to a lack of large-scale datasets and the small relative size of the crops in the image. In this work, we address these challenges by reformulating the apple detection task in a semi-supervised manner. To this end, we provide the large, high-resolution dataset MAD comprising 105 labeled images with 14,667 annotated apple instances and 4,440 unlabeled images. Utilizing this dataset, we also propose a novel Semi-Supervised Small Apple Detection system S$^3$AD based on contextual attention and selective tiling to improve the challenging detection of small apples, while limiting the computational overhead. We conduct an extensive evaluation on MAD and the MSU dataset, showing that S$^3$AD substantially outperforms strong fully-supervised baselines, including several small object detection systems, by up to $14.9\%$. Additionally, we exploit the detailed annotations of our dataset w.r.t. apple properties to analyze the influence of relative size or level of occlusion on the results of various systems, quantifying current challenges.