Researchers from the National University of Singapore and Yonsei University (Korea) have developed a method for detecting hidden cameras indoors using a conventional smartphone equipped with a ToF (Time of flight) sensor. It is noted that currently a hidden camera can be bought for a little more than a dollar and such cameras are 1-2 millimeters in size, which makes it difficult to find them indoors. In South Korea, more than 6800 incidents related to the placement of hidden cameras in hotel rooms or bathrooms were recorded per year.
The LAPD (Laser-Assisted Photography Detection) method proposed by the researchers makes it possible to detect hidden cameras using modern smartphones equipped with a depth sensor (ToF), used to estimate the distance to objects when focusing the camera and in augmented reality applications. Examples of smartphones that use such sensors are Samsung S20 and Huawei P30 Pro. The sensor builds a depth map by scanning the surrounding area with a laser and calculating the distance based on the arrival delay of the reflected beam.
The method of detecting hidden cameras is based on the detection of anomalies when lenses and lenses are illuminated by a laser, which create specific glare on the resulting depth map. Anomalies are detected using a machine learning algorithm that can distinguish camera-specific glare. The authors of the study intend to publish a finished application for the Android platform after solving some problems with API restrictions.
The total time it takes to scan a room is estimated at 30-60 seconds. In an experiment conducted with 379 volunteers, hidden cameras were detected using the LAPD method in 88.9% of cases. For comparison, only 46% of the participants in the experiment could find cameras by eye, and the efficiency of using the specialized K18 signal detector was 62.3% and 57.7%, depending on the selected scanning mode. The LAPD method also showed a lower false positive rate - 16.67% versus 26.9% / 35.2% for K18 and 54.9% for eye searches.
LAPD detection accuracy depends on the hidden camera entering the 20-degree viewing angle of the sensor and being at the optimal distance from the sensor (if the camera is too close, the glare from the camera is smeared, and if it is too far away, it disappears). To improve accuracy, it is proposed to use sensors with a higher resolution (in the smartphones that researchers have, the resolution of the ToF sensor is 320 Γ 240, i.e. the size of the anomaly in the image is only 1-2 pixels) and depth detail (now there are only 8 for each pixel). depth levels).
Other methods for assessing the presence of a hidden camera are wireless traffic analyzers that determine the presence of video streaming over a wireless network, as well as electromagnetic radiation scanners.
Source: opennet.ru