1. 1. Smoke and Fire Detection Algorithm
    1. Function Description

Smoke and fire detection refers to presetting a monitoring area in a surveillance scene, and triggering an alarm when smoke or fire occurs within that area. The detection distance is 10-20 meters, and the resolution of smoke and fire in a 1080p image should be ≥ 30*30 pixels.

1. 1. 1. Recommended Scenarios

Scene requirements: Can be used indoors and outdoors, avoid direct sunlight or backlight conditions; only supports full-color images at night, if the camera does not support full-color mode, please use white light for illumination.

Target distance requirements: Can detect events beyond 3 meters from a normal 3-meter pole height.

Target pixel requirements: The resolution of the detected target in a 1080P image should be ≥ 30*30 pixels.

Flame colors: Supports detection of red, orange, and yellow flames.

Smoke colors: Supports detection of irregular smoke such as off-white and dark gray.

Recall rate: 90%

1. 1. Crowd Gathering Detection Algorithm
    1. Function Description

Crowd gathering detection refers to presetting a monitoring area in a surveillance scene, and triggering an alarm when the number of people in the set area exceeds a certain threshold. The resolution of the detected target in a 1080p image should not be less than 30*30 pixels.

1. 1. 1. Recommended Scenarios

Target distance requirements: Can detect events beyond 3 meters from a normal 3-meter pole height.

Target pixel requirements: The resolution of the detected target in a 1080P image should be greater than or equal to 30*30 pixels.

Recall rate: 95%

1. 1. Lane Occupancy Detection Algorithm
    1. Function Description

Lane occupancy detection refers to presetting a monitoring area in a surveillance scene, detecting whether there are vehicles or two-wheeled vehicles in the area, custom setting a dwell time, and triggering an alarm promptly. The resolution of the detected target in a 1080p image should not be less than 30*30 pixels.

1. 1. 1. Recommended Scenarios

Target distance requirements: Can detect events beyond 3 meters from a normal 3-meter pole height.

Target pixel requirements: The resolution of the detected target in a 1080P image should be greater than or equal to 30*30 pixels.

Recall rate: 95%

1. 1. Unattended Object Detection Algorithm
    1. Function Description

Unattended object detection refers to presetting a monitoring area in a surveillance scene, detecting whether there are various items such as luggage, containers, furniture, trolleys, traffic cones, cardboard boxes, trash cans, mops, etc., in the area, custom setting a dwell time, and triggering an alarm promptly. The resolution of the detected target in a 1080p image should not be less than 30*30 pixels.

1. 1. 1. Recommended Scenarios

Target distance requirements: Can detect events beyond 3 meters from a normal 3-meter pole height.

Target pixel requirements: The resolution of the detected target in a 1080P image should be greater than or equal to 30*30 pixels.

Recall rate: 95%

1. 1. Hard Hat and Reflective Vest Detection Algorithm
    1. Function Description

Hard hat and reflective vest detection refers to presetting a monitoring area in a surveillance scene, and triggering an alarm promptly when personnel in the area are not wearing hard hats or reflective vests. The resolution of the detected target in a 1080p image should not be less than 30*30 pixels.

1. 1. 1. Recommended Scenarios

Target distance requirements: Can detect events beyond 3 meters from a normal 3-meter pole height.

Target pixel requirements: The resolution of the detected target in a 1080P image should be greater than or equal to 30*30 pixels.

Recall rate: 95%