Visual analysis tasks, including crowd management, often require resource-intensive machine learning models, posing challenges for deployment on edge hardware. Consequently, cloud computing emerges as a prevalent solution. To address privacy concerns associated with offloading video data to remote cloud platforms, we present a novel approach using adversarial training to develop a lightweight obfuscator neural network. Our method focuses on pedestrian detection as an example of visual analysis, allowing the transformation of video frames on the camera itself to retain only essential information for pedestrian detection while preserving privacy. Importantly, the obfuscated data remains compatible with publicly available object detectors, requiring no modifications or significant loss in accuracy. Additionally, our technique overcomes the common limitation of relying on labeled sensitive attributes for privacy preservation. By demonstrating the inability of pedestrian attribute recognition models to detect attributes in obfuscated videos, we validate the efficacy of our privacy protection method. Our results suggest that this scalable approach holds promise for enabling camera usage in video analytics while upholding personal privacy.

Soil compaction is a widespread problem, leading to soil degradation, yield losses, and adverse environmental impacts. Nowadays, various measurement methods exist to assess and map soil compaction, with vertical cone penetration resistance measurements being one of the most commonly used. This method is easy, rapid, inexpensive, and generally accepted. However, manual penetration resistance measurements are time-consuming, labor-intensive, and often less accurate due to inconsistent penetration speed. To address these limitations, an automated penetrometer was developed and integrated on an autonomous robot platform, paving the way for high-resolution compaction mapping as a starting point for precision subsoiling to remediate soil compaction. The performance of this setup was validated in controlled and field conditions against a hand-held penetrometer. Therefore, experiments were conducted in soil-filled cylinders and on plots of a long-term field experiment, including measurements across spraying tracks. The automated penetrometer demonstrated high correlations with the hand-held device under controlled conditions, though the correlation was somewhat lower in the field due to the soil’s heterogeneity. Deviations between the two measurement devices were likely caused by the inconsistent insertion speed of the hand-held penetrometer, particularly in soils with high penetration resistance. Both penetrometers successfully identified the plow pan at a depth of 30–40 cm but were unable to clearly show the effect of the long-term presence of spraying tracks.