In this work, we developed and proposed an auto‐classification technique for concealed weapon detection (CWD) for passive millimeter‐wave (PMMW) imaging systems. This technique has the ability to detect, classify and image hidden objects beneath the cloth of human targets. The algorithm was based on segmentation of normalized gray‐scale images for raw passive millimeter‐wave captured images and employing a decisive criterion for the classification of the targets. First, we tested our algorithm with the simulated data that were created numerically. Next, we examined our CWD technique with the measured data that were collected with PMMW radiometric imaging system at Marmara Research Center of TÜBİTAK. After producing the raw passive radar images, we have applied our passive CWD technique to the measured raw images to assess the performance of the algorithm. Produced edge‐detected final images of the concealed objects provide successful operation of the proposed technique.

No abstract

In this study, an algorithm to the detection and imaging of hidden arms for passive millimeter-wave (PMMW) imaging systems is proposed. This technique is; in fact, an improved version of our previously developed auto-classification algorithm by extending it by exploiting the Otsu’s multi-level thresholding method. The detailed derivation and the brief steps of the proposed algorithm are given. The proposed algorithm is tested and validated by real PMMW images obtained by a real radiometric imaging system. Resultant measured images are obtained with the employment of signal and image processing procedures of the suggested technique. It is demonstrated by the constructed PMMW images that proposed technique successfully detects a concealed metal threat and also predicts its size by drawing the shape outline based on Otsu’s multi-level thresholding routine that was specially tailored to our auto-classification technique.

No abstract

No abstract