YOLOv8 Video Stream Inference on Milk-V Duo 256M Fails with Sensor Config Errors and Missing Camera Nodes

Body:
Hi all,
I’m trying to deploy YOLOv8 for video stream inference on a Milk-V Duo 256M device using the milkv-duos-musl-riscv64-sd image. However, I encounter two critical issues:

Sensor Config Parsing Failure
The application fails to parse /mnt/data/sensor_cfg.ini with the following log:

[SAMPLE_COMM_SNS_ParseIni]-2219: Parse /mnt/data/sensor_cfg.ini
[parse_source_devnum]-1812: devNum = 1
[parse_sensor_name]-1893: sensor = GCORE_GC2083_MIPI_2M_30FPS_10BIT
[parse_sensor_busid]-1922: bus_id = 2
[parse_sensor_i2caddr]-1933: sns_i2c_addr = 37
[parse_sensor_mipidev]-1944: mipi_dev = 0
[parse_sensor_laneid]-1955: Lane_id = 1, 0, 2, -1, -1
[parse_sensor_pnswap]-1966: pn_swap = 0, 0, 0, 0, 0
Failed to get senor infomation from ini file (/mnt/data/sensor_cfg.ini).

Missing Camera Device Nodes
The system reports missing nodes:
:x: Device node missing: /dev/video0
:x: Device node missing: /dev/video1
However, /dev/cvi-vi and /dev/cvi-vpss exist.
Any advice on debugging steps or configuration adjustments would be greatly appreciated!

3 Likes

It can be important to specify which version you’re using.

Hi guys,

I’m having the same problem. I’m using module 01 with SDK version 2, following the steps described in the documentation.

https://milkv.io/docs/duo/application-development/tdl-sdk/tdl-sdk-introduction

1 Like

I am currently using the following firmware image on my Milk-V Duo S device:milkv-duo256m-musl-riscv64-sd_2025-0604-0928.img

1 Like

I’ve never tried running any Duo 256M firmware on Duo S. Why exactly are you trying this?

Sorry,my mistake,the device is Duo256M。

1 Like

That’s fine. What’s more important: have you built the image yourself? If yes, it is important which repo you cloned (v1 or v2), and which branch and state (commit hash) it was built on — the official releases supply differently named files.

I myself don’t have the camera and can’t test such a setup.

Thank you very much for your response. I used the official Buildroot SDK V2 method to build the milkv-duos-musl-riscv64-sd image file via the one-step compilation approach after obtaining the SDK. The documentation link is as follows: Introduction | Milk-V.
Even after trying the official face recognition implementation, the same error persists. I have attempted multiple methods but still cannot resolve this issue. The camera only runs with the camera-test.sh script, yet the following error still occurs. I suspect it might be a problem with library files, but I can confirm the relevant libraries exist on my development board.


[root@milkv-duo]~# camera-test.sh
[SAMPLE_COMM_SNS_ParseIni]-2219: Parse /mnt/data/sensor_cfg.ini  
[parse_source_devnum]-1812: devNum =  1  
[parse_sensor_name]-1893: sensor =  GCORE_GC2083_MIPI_2M_30FPS_10BIT  
[parse_sensor_busid]-1922: bus_id =  2  
[parse_sensor_i2caddr]-1933: sns_i2c_addr =  37  
[parse_sensor_mipidev]-1944: mipi_dev =  0  
[parse_sensor_laneid]-1955: Lane_id =  1, 0, 2, -1, -1  
[parse_sensor_pnswap]-1966: pn_swap =  0, 0, 0, 0, 0  
MMF Version:6b03c2762-64bit  
Create VBPool[0], size: (3110400 * 5) = 15552000 bytes  
Create VBPool[1], size: (1382400 * 5) = 6912000 bytes  
Create VBPool[2], size: (2764800 * 3) = 8294400 bytes  
Total memory of VB pool: 30758400 bytes  
Initialize SYS and VB  
Cannot open '/dev/cvi-vo': 2, No such file or directory  
Initialize VI  
ISP Vipipe(0) Allocate pa(0x8d056000) va(0x0x3fe1b90000) size(311584)  
stSnsrMode.u16Width 1920 stSnsrMode.u16Height 1080 30.000000 wdrMode 0 pstSnsObj 0x3fe264a3e8  
[SAMPLE_COMM_VI_StartMIPI]-494: sensor 0 stDevAttr.devno 0  
awbInit ver 6.9@2021500  
0 R:1400 B:3100 CT:2850  
1 R:1500 B:2500 CT:3900  
2 R:2300 B:1600 CT:6500  
Golden 1024 1024 1024  
WB Quadratic:0  
isWdr:0  
ViPipe:0,===GC2083 1080P 30fps 10bit LINE Init OK!===
********************************************************************************  
cvi_bin_isp message  
gerritId:      NULL           commitId:      6b03c2762      
md5:           1b6dd6bec5dfd417b8d1136b9ca0aa67  
sensorNum      1              
sensorName0    2083           

PQBIN message  
gerritId:      80171          commitId:      5c9d8fc5d      
md5:           ba5a510e093ad42db6788e6c2d13169e  
sensorNum      3              
sensorName0    2053           

author:        wanqiang.he    desc:          思博慧CV1812H_GC2083_RGB_mode_V1.0.0  
createTime:    2023-08-04 16:48:08version:       V1.1           
tool Version:       v3.0.5.24           mode:      
********************************************************************************  
sensorName(0) mismatch, mwSns:2083 != pqBinSns:2053  
19700101 00:14:14.105 1929 E isp AF_GetAttr:574 pstFocusMpiAttr is NULL  
JSON_READ_ERR:NOT_EXIST 71(L) lblc  
JSON_READ_ERR:NOT_EXIST 71(L) lblcLut  
JSON_READ_ERR:NOT_EXIST 71(L) clut_hsl  
JSON_READ_ERR:DATA_TYPE 77(L) vc_motion.MotionThreshold  
JSON_READ_ERR:NOT_EXIST 71(L) teaisp_bnr  
JSON_READ_ERR:NOT_EXIST 71(L) teaisp_bnr_np  
JSON_READ_ERR:NOT_EXIST 71(L) AWBAttrEx.u16MultiLSThr  
JSON_READ_ERR:NOT_EXIST 71(L) AWBAttrEx.u16CALumaDiff  
JSON_READ_ERR:NOT_EXIST 71(L) AWBAttrEx.u16CAAdjustRatio  
JSON_READ_ERR:NOT_EXIST 71(L) AWBAttrEx.stInterference  
JSON_READ_ERR:NOT_EXIST 71(L) FocusAttr  
19700101 00:14:14.451 1929 E isp AF_SetAttr:558 pstFocusMpiAttr is NULL  
[SAMPLE_COMM_ISP_Thread]-390: ISP Dev 0 running!  
Initialize VPSS  
---------VPSS[0]---------  
Input size: (1920x1080)  
Input format: (19)  
VPSS physical device number: 1  
Src Frame Rate: -1  
Dst Frame Rate: -1  
    --------CHN[0]-------  
    Output size: (1280x720)  
    Depth: 1  
    Do normalization: 0  
        Src Frame Rate: -1  
        Dst Frame Rate: -1  
    ----------------------  
    --------CHN[1]-------  
    Output size: (1280x720)  
    Depth: 1  
    Do normalization: 0  
        Src Frame Rate: -1  
        Dst Frame Rate: -1  
    ----------------------  
------------------------  
Bind VI with VPSS Grp(0), Chn(0)  
Attach VBPool(0) to VPSS Grp(0) Chn(0)  
Attach VBPool(1) to VPSS Grp(0) Chn(1)  
Initialize VENC  
venc codec: h264  
venc frame size: 1280x720  
Initialize RTSP  
rtsp://127.0.1.1/h264  
prio:0  
Cannot open '/dev/cvi-vo': 2, No such file or directory  
anchor:-8,-8,8,8  
anchor:-16,-16,16,16  
bbox:bbox_8_Conv_dequant  
landmark:kps_8_Conv_dequant  
score:score_8_Sigmoid_dequant  
anchor:-32,-32,32,32  
anchor:-64,-64,64,64  
bbox:bbox_16_Conv_dequant  
landmark:kps_16_Conv_dequant  
score:score_16_Sigmoid_dequant  
anchor:-128,-128,128,128  
anchor:-256,-256,256,256  
bbox:bbox_32_Conv_dequant  
landmark:kps_32_Conv_dequant  
score:score_32_Sigmoid_dequant  
Enter TDL thread  
Enter encoder thread  
0 R:1165 B:3087 CT:2688  
1 R:1464 B:2327 CT:3937  
2 R:1974 B:1613 CT:7225  
Golden 1464 1024 2327  
wdrLEOnly:1  
face count: 1strong text
1 Like

I have the same error using a DuoS board. I used the latest official image V2.0.1 installed using SD card ( milkv-duos-musl-riscv64-sd_v2.0.1.img.zip.

I can run the camera-test.sh program and it works fine. However, when trying to run the person detection (The following example: yolov8-person-pet-det-example). This was unsuccessful and I get the same error: Failed to get senor infomation from ini file (/mnt/data/sensor_cfg.ini).

Also the documentation needs to be updated. For example the link to the pet-det cvimodel is dead and needs to be updated (https://github.com/sophgo/tdl_models/blob/main/cv181x/pet_det_640x384.cvimodel).

1 Like