Dahua WizColor 5x59-PRO and SmartLight 5x59-IL new series

For those using 0731 and have problems with false detection

Can you please try update to 0827 from dahua website.
It has improved AI. It may fix the false alerts.
 
  • Like
Reactions: Clark Griswald
Disable AcuPick.. camera will reboot. then in IVS -> Global settings enable Large Scale AI Models..

recheck how IVS works in this mode... and describe here is there any difference...

First day of testing the AS 3.6mm with Large Scale AI Models enabled, in my most difficult scene/placement

So far very happy with critter detection (0731 FW)


View attachment Recording 2025-10-20 062547.mp4

View attachment Recording 2025-10-20 062724.mp4

View attachment Recording 2025-10-20 062856.mp4
 
I have the 54PRO-AS 3.6mm hung. Limited testing yet. Rainy yucky overcast weather

Couple of things I'm seeing.

  • Daytime Color is indeed softer than the 5442 but I could probably live with it as I tweek settings and improve sharpness.
  • Nighttime B&W/IR is superior to the 5442 which I expected from my experience with the 3449 PRO. Better IR dispersion and overall brighter. I still think the 5442 is a slightly better quality image though
  • I'm having trouble now with IVS ( @tigerwillow1 ). Nighttime it did very well on animals. Daytime I'm having to dial it way back on sensitivity as it thinks my foot stool is a critter.
  • The transition to color when Smart Illumination kicks in is..... different. I prefer how the 3449 PRO handles it
  • Illumination Overexposure Remover is way too severe. I've turned it off. Here again the 3449 PRO does this very well and correctly with a more gradual and less severe cutting of Gain. The 54PRO severely overcompensates and has a ON/OFF effect creating a lot of "dancing pixels" when off. They really need a slider sensitivity control for this as opposed to On/Off
  • My biggest problem right now is that the IVS rules are not synching with the NVR no matter what I do. Active schedule time and Recording not working right. Camera shows the tripping of the rule, Record is set, but not being recorded (as a separate clip with hash mark) on the NVR... grrrrrrr


Example of the softness. Focus on Mrs bigredfish's hair (I made sure to snap both when she was perfectly still and have repeated similar comparisons)
Also see the "jaggies" on the furthest 3 boards of the deck on the 54PRO

192.168.1.110_Backdoor-5459H-ASE-PV-PRO_main_20251018095029_@1.jpg

192.168.1.110_Backdoor-5459H-ASE-PV-PRO_main_20251020095310_@1.jpg




A lot of work to do on night image and light transitions.

192.168.1.110_Backdoor-5459H-ASE-PV-PRO_main_20251020064604_@1.jpg 192.168.1.110_Backdoor-5459H-ASE-PV-PRO_main_20251020064551_@1.jpg 192.168.1.110_Backdoor-5459H-ASE-PV-PRO_main_20251020064526_@1.jpg 192.168.1.110_Backdoor-5459H-ASE-PV-PRO_main_20251020064521_@1.jpg

View attachment 192.168.1.110_ch4_20251020064514_20251020064606.mp4
 
Disable AcuPick.. camera will reboot. then in IVS -> Global settings enable Large Scale AI Models..
Thanks Steve. I had finally figured this out yesterday, and would have sooner but for how the UI behaves and how I make assumptions. On the UI, when I turned off the AccuPick switch, the IVS switch got grayed out, leading me to assume that AccuPick has to be on to use IVS. I finally got past that and downloaded the 8/27 firmware and am running the large models. As is so typical, animals have quit walking by the camera just when I'm trying to test it. On some days, it's almost like there's a portable zoo walking by. The 5442 with dumb IVS did capture one animal missed by the 54PRO, and I don't expect the PRO will ever trigger on a passing bird. This one blew undetected through 4 tripwire lines.


Capture3.JPG Capture1.JPG Capture.JPG
 
  • Like
Reactions: bigredfish
Toggling Accupick off requires a reboot to activate IVS and you doing that as part of the update likely was the reason and not the firmware update itself.
 
  • Like
Reactions: duplo