I have a Dahua DHI-TPC-BF1241, hardware version S8, 10 mm lens, installed outdoors for perimeter detection with IVS.
Model: DHI-TPC-BF1241
Hardware: S8
Thermal sensor: 256x192
Lens: 10 mm
Algorithm Version: 2.0.6
Intelligent Version: V2.006.20230923
Firmware tested:
V2.680.0000000.2.R, 2024-05-16
V2.680.0000000.3.R, 2024-10-12
V2.681.0000000.1.R, 2026-03-10
I've tried multiple firmwares, including 2026, but the IVS behavior doesn't change significantly.
The main problem is this:
At night, detection improves slightly, while during the day, especially with high temperatures and sun-heated surfaces, the IVS completely misses the person even at relatively short distances, and the thermal image seems terrible.
Dahua claims about 70 m of perimeter protection for a person with the 10 mm lens. In my scenario, however, at certain times of day:
At about 40 m, it may not detect completely;
in some problem areas, it misses the target even at 20–30 m;
the bounding box often doesn't appear at all, so it seems like a detection/segmentation problem even before human classification.
The interesting thing is that the behavior changes significantly with the temperature of the environment. A camera that appears significantly better at night may become the worst during the day, depending on the warmer background.
Visually, during the hottest hours of the day, the human silhouette actually tends to blend into the thermal background, especially in front of walls, grass, pavement, or warm vegetation.
The camera is installed at about 2.3 m above the ground.
The main question is: have you ever seen such marked behavior on the BF1241-S8 or other 256x192 thermal cameras?
I'm especially interested in understanding whether:
it's a normal limitation due to thermal crossover/low target-background ΔT;
the BF1241-S8 has a particularly weak IVS algorithm;
are there AGC/gain/DDE/NR settings that significantly improve detection;
someone has achieved reliable detection of people at 40–70 m even during the day in high temperatures.
In my case, I'm not talking about false alarms: the problem is precisely the lack of target acquisition.
I can attach day/night thermal screenshots and images of the visible scene.
Thanks in advance for any real-world feedback.
Model: DHI-TPC-BF1241
Hardware: S8
Thermal sensor: 256x192
Lens: 10 mm
Algorithm Version: 2.0.6
Intelligent Version: V2.006.20230923
Firmware tested:
V2.680.0000000.2.R, 2024-05-16
V2.680.0000000.3.R, 2024-10-12
V2.681.0000000.1.R, 2026-03-10
I've tried multiple firmwares, including 2026, but the IVS behavior doesn't change significantly.
The main problem is this:
At night, detection improves slightly, while during the day, especially with high temperatures and sun-heated surfaces, the IVS completely misses the person even at relatively short distances, and the thermal image seems terrible.
Dahua claims about 70 m of perimeter protection for a person with the 10 mm lens. In my scenario, however, at certain times of day:
At about 40 m, it may not detect completely;
in some problem areas, it misses the target even at 20–30 m;
the bounding box often doesn't appear at all, so it seems like a detection/segmentation problem even before human classification.
The interesting thing is that the behavior changes significantly with the temperature of the environment. A camera that appears significantly better at night may become the worst during the day, depending on the warmer background.
Visually, during the hottest hours of the day, the human silhouette actually tends to blend into the thermal background, especially in front of walls, grass, pavement, or warm vegetation.
The camera is installed at about 2.3 m above the ground.
The main question is: have you ever seen such marked behavior on the BF1241-S8 or other 256x192 thermal cameras?
I'm especially interested in understanding whether:
it's a normal limitation due to thermal crossover/low target-background ΔT;
the BF1241-S8 has a particularly weak IVS algorithm;
are there AGC/gain/DDE/NR settings that significantly improve detection;
someone has achieved reliable detection of people at 40–70 m even during the day in high temperatures.
In my case, I'm not talking about false alarms: the problem is precisely the lack of target acquisition.
I can attach day/night thermal screenshots and images of the visible scene.
Thanks in advance for any real-world feedback.
