Recently, more and more images are compressed and sent to the back-end devices for machine analysis tasks ( e.g. , object detection) instead of being purely watched by humans. However, most traditional or learned image codecs are designed to minimize the distortion of the human visual system without considering the increased demand from machine vision systems. In this work, we propose a preprocessing enhanced image compression method for machine vision tasks to address this challenge. Instead of relying on the learned image codecs for end-to-end optimization, our framework is built upon the traditional non-differential codecs, which means it is standard compatible and can be easily deployed in practical applications. Specifically, we propose a neural preprocessing module before the encoder to maintain the useful semantic information for the downstream tasks and suppress the irrelevant information for bitrate saving. Furthermore, our neural preprocessing module is quantization adaptive and can be used in different compression ratios. More importantly, to jointly optimize the preprocessing module with the downstream machine vision tasks, we introduce the proxy network for the traditional non-differential codecs in the back-propagation stage. We provide extensive experiments by evaluating our compression method for several representative downstream tasks with different backbone networks. Experimental results show our method achieves a better trade-off between the coding bitrate and the performance of the downstream machine vision tasks by saving about 20% bitrate.

Radiometric normalization, as an essential step for multi-source and multi-temporal data processing, has received critical attention. Relative Radiometric Normalization (RRN) method has been primarily used for eliminating the radiometric inconsistency. The radiometric transforming relation between the subject image and the reference image is an essential aspect of RRN. Aimed at accurate radiometric transforming relation modeling, the learning-based non-linear regression method, Support Vector machine Regression (SVR) is used for fitting the complicated radiometric transforming relation for the coarse-resolution data-referenced RRN. To evaluate the effectiveness of the proposed method, a series of experiments are performed, including two synthetic data experiments and one real data experiment. And the proposed method is compared with other methods that use linear regression, Artificial Neural Network (ANN) or Random Forest (RF) for radiometric transforming relation modeling. The results show that the proposed method performs well on fitting the radiometric transforming relation and could enhance the RRN performance.

Collision accidents involving construction vehicles and workers frequently occur at construction sites. Computer vision (CV) technology presents an efficient solution for collision-risk pre-warning. However, CV-based methods are still relatively rare and need an enhancement of their performance. Therefore, a novel three-stage collision-risk pre-warning model for construction vehicles and workers is proposed in this paper. This model consists of an object-sensing module (OSM), a trajectory prediction module (TPM), and a collision-risk assessment module (CRAM). In the OSM, the YOLOv5 algorithm is applied to identify and locate construction vehicles and workers; meanwhile, the DeepSORT algorithm is applied to the real-time tracking of the construction vehicles and workers. As a result, the historical trajectories of vehicles and workers are sensed. The original coordinates of the data are transformed to common real-world coordinate systems for convenient subsequent data acquisition, comparison, and analysis. Subsequently, the data are provided to a second stage (TPM). In the TPM, the optimized transformer algorithm is used for a real-time trajectory prediction of the construction vehicles and workers. In this paper, we enhance the reliability of the general object detection and trajectory prediction methods in the construction environments. With the assistance afforded by the optimization of the model’s hyperparameters, the prediction horizon is extended, and this gives the workers more time to take preventive measures. Finally, the prediction module indicates the possible trajectories of the vehicles and workers in the future and provides these trajectories to the CRAM. In the CRAM, the worker’s collision-risk level is assessed by a multi-factor-based collision-risk assessment rule, which is innovatively proposed in the present work. The multi-factor-based assessment rule is quantitatively involved in three critical risk factors, i.e., velocity, hazardous zones, and proximity. Experiments are performed within two different construction site scenarios to evaluate the effectiveness of the collision-risk pre-warning model. The research results show that the proposed collision pre-warning model can accurately predict the collision-risk level of workers at construction sites, with good tracking and predicting effect and an efficient collision-risk pre-warning strategy. Compared to the classical models, such as social-GAN and social-LSTM, the transformer-based trajectory prediction model demonstrates a superior accuracy, with an average displacement error of 0.53 m on the construction sites. Additionally, the optimized transformer model is capable of predicting six additional time steps, which equates to approximately 1.8 s. The collision pre-warning model proposed in this paper can help improve the safety of construction vehicles and workers.

In recent years, graph-partition-based clustering algorithms have attracted increasing attention. These algorithms first construct a graph over the data points and then partition this graph, regarding each connected subgraph in the partitioned graph as a cluster. However, traditional graph-partition-based clustering algorithms face challenges when clustering datasets with highly imbalanced density distributions. This is because, during the process of graph partitioning, they mainly rely on edge lengths and largely ignore local density variations. To address this issue, we propose the Adaptive Graph Partitioning (AGP) clustering algorithm. AGP integrates local density information into the partitioning process and adaptively normalizes the magnitudes of edge weights in both sparse and dense regions. This enhancement allows AGP to avoid the over-partitioning of dense clusters, effectively addressing a common problem in traditional graph-partition-based clustering algorithms. Additionally, the mechanism of connectivity domain differences is introduced into AGP, further enhancing the algorithm’s ability to discriminate between neighboring clusters that are difficult to separate. Extensive experiments on 13 benchmark datasets show that AGP achieves the best clustering performance on 7 datasets and performs competitively on the others, especially when density imbalance is severe. Moreover, scalability experiments on datasets with up to 100,000 data points, together with complexity analysis, demonstrate that AGP enjoys favorable time and memory efficiency compared with representative baselines.