Surveillance system and operation method thereof

An operation method of a surveillance system, comprising:

• detecting a movement in a scene based on signals sensed by a motion-detection engine included in an auxiliary image sensor module being an always-on sensor that integrates a low-resolution image sensor which is in full-time operation for capturing low-resolution images of the scene;
• determining one or more ROIs based on the low-resolution images;
• extracting features in the ROI based on the images;
• recognizing a human in the scene by a human-detection engine included in the auxiliary image sensor module according to the features extracted from the low-resolution images;
• wherein human-detection engine detects whether or not the human enters the scene based on the movement information generated by the motion-detection engine via different ROIs in one low-resolution image in which one of the ROIs covers a full body of the human within the scene at a distant location, and/or another ROI covers part of the human at a near location;
• outputting the recognized human in the scene;
• detecting a moving object in the scene based on the movement to be detected;
• and
• activating a main image sensor module from a power-saving mode for performing surveillance function when the moving object is detected by the auxiliary image sensor module, in which the main image sensor module is in the power-saving mode when there is no human to be detected in the scene.

A surveillance system, comprising:

• an auxiliary image sensor module being an always-on sensor that integrates an auxiliary processor, and a low-resolution image sensor that is in full-time operation for taking low-resolution images and that outputs sensed signals from a scene;
• a motion-detection engine that is included in the auxiliary image sensor module and is used to detect a movement in the scene based on the sensed signals outputted from the low-resolution image sensor of the auxiliary image sensor module;
• a human-detection engine that is included in the auxiliary image sensor module and is used to detect a human in the scene based on the sensed signals outputted from the low-resolution image sensor after the movement is detected by the motion-detection engine;
• wherein human-detection engine detects whether or not the human enters the scene based on the movement information generated by the motion-detection engine via different ROIs in one low-resolution image in which one of the ROIs covers a full body of the human within the scene at a distant location, and/or another ROI covers part of the human at a near location;
• a controller receiving signals sent by the human-detection engine;
• and
• a main image sensor module, implemented by a main sensor being a high-resolution sensor and a main processor, that is in a power-saving mode when there is no human to be detected in the scene, and that is activated to perform a surveillance function for collecting surveillance data, performing object recognition, human facial recognition, and/or recording surveillance data in the scene when the human is detected by the human-detection engine;
• wherein, the auxiliary processor generates an actuation signal when any moving object is detected by the low-resolution image sensor;
• the controller activates the main image sensor module from the power-saving mode if the controllers receives the actuation signal and acknowledges that a movement or the moving object is found.