Enhancing Prospective Consistency for Semisupervised Object Detection in Remote-Sensing Images

Deep-learning-based object detection has recently played a vital role in both computer vision and Earth observation communities. However, the performance of modern object detectors is highly limited by the quantity and quality of manually labeled training samples. Furthermore, compared to object detection in natural scenes, remote-sensing object detection (RSOD) faces two specific critical challenges: 1) densely arranged instances: geospatial objects tend to be densely packed in remote-sensing scenarios and 2) large variations in object scale: the wide field of the bird's eye view leads to dramatic variations in object scale across various categories. The above issues bring significant difficulties to attaining manual annotations for deep-learning-based RSOD. To this end, in this article, we turn our attention from fully supervised RSOD to semisupervised RSOD and propose a novel framework based on the teacher-student paradigm, namely Prospective Consistency Teacher (PCT), which includes three crucial components, that is, weighted dense-proposal learning (WDPL), mean-consistency-based proposal pruning (MCPP), and EM-based fitting policy (EFP). Specifically, WDPL reweights the dense proposals with box confidences, while MCPP ranks the student proposals with consistency analysis to select discriminative and consistent boxes. EFP can automatically set thresholds for pseudo-labels and improve the consistent information of the teacher network. Extensive experimental results on two challenging public datasets, that is, DOTA and DIOR, have demonstrated the reduced reliance of our proposed method on large amounts of labeled data for the task of RSOD.