Abstract: X/γ rays can directly interact with CMOS sensors, thus detecting nuclear radiation based on the widespread surveillance cameras is possible. The objective of this paper is to validate this with an un-covered CMOS camera under the condition of static scenes. Static scene videos (1 000 frame images) were acquired using a Logitech C170 camera with different distances of 7.4×10⁸ Bq ⁹⁹Tc^m source to the camera. It is observed that bright blotches appear in these videos randomly. A simple algorithm to distinguish whether a frame contained a bright blotch was developed by combing the inter-frame differencing, Gaussian smoothing and threshold judgement. Comparison measurements for dose rate were carried out using TERRA-P personal dosimeter. The experimental results show that the bright blotches are generated by the nuclear radiation, and the number of images with bright blotches per thousand images has a linear relationship with the radiation dose rate (correlation coefficient R is 0.99). The proposed algorithm in this paper can effectively identify the radioactive events in the static scene videos acquired by an un-covered CMOS camera, i.e. realizing nuclear radiation detection. It is expected to achieve low-cost radiation monitoring using the large number of surveillance cameras by further developing the processing algorithm.