- 博客(29)
- 资源 (8)
- 收藏
- 关注
原创 xHiveAI Jetson NX盒子:软件开发指南
群蜂技术开发的Nvidia Jetson Xaiver NX盒子面向边缘AI应用需求,提供了软硬一体的解决化方案。。
2022-07-20 15:52:58
712
原创 xHiveAI Jetson NX盒子:jpeg图片编码
在实际的AI处理中,经常有保存和上传jpeg报警图片的操作。Nvidia提供的jpegencoder库中只能对YUV图片来操作,我们盒子中的libjpegenc.so库提供了图片转换和jpeg压缩一体化的操作,一个API接口可以完成整个操作流程,大大简化了上层软件的操作。...
2022-07-20 15:45:46
1083
原创 xHiveAI Jetson NX盒子:图片尺寸,像素格式等转换
在Nvidiarelease的图片转换库的基础上,盒子提供了更加简单易用的libpicconverter.so库,简单的4个API接口就可以完成绝大部分的操作,方便客户专注于上层AI软件的开发。
2022-07-20 15:14:33
759
原创 xHiveAI-A311D:How to upgrade the firmware images(Internal Yocto)
The following guidance is ONLY for internal development purpose!!!Please follow the steps to upgrade the firmware yocto images on your board:Step1: Setup the tftp server in your Host PCStep2: Download the upgrade images from below linkageGet the u
2021-05-25 14:39:59
254
原创 xHiveAI-A311D:Questions and Answers
1) Failed to boot into the ubuntu system, the error message shows in the serial console as belowSolution:execute below commands one by onexhiveai# abboot resetxhiveai# reset<enter into the Ubuntu OS>do command: boot_util -m2) HDMI LCD M
2021-04-19 11:26:51
297
原创 xHiveAI-A311D:AI Algorithm - Develop the Facial Recognition Algorithm
The example code is under git repository: a311d/ai_examples/face_recog_callback. It realizes the functions as below:1) Detect the human face in the video2) Match the detected face against the pictures in the facial library3) If matching, the registe
2021-04-19 11:25:01
356
原创 xHiveAI-A311D:AI Algorithm - Run Preset Test Cases
The testxhiveai is provided to run some preset test cases. Please follow the below steps to try some testing:Step 1: Modify the test cases configurationsvi /etc/test/testxhiveai.xml Facial Detection Display: Show the video in the window and draw b
2021-04-19 11:24:00
301
原创 xHiveAI-A311D:Graphics System V2.0
The standard DirectFB v1.7.7 is used for the graphics system on the board. Some example code are provided for the quick start./usr/bin/df_XXXX:df_andi(most popular testing)/usr/bin/dfbtest_XXXXdfbtest_layer
2021-04-19 11:21:24
283
原创 xHiveAI-A311D:How to use the hardware JPEG encoder
A311D SOC integrated the hardware jpeg encoder to encode the raw YUV data into the jpeg format picture.The libjpegenc.so is developed to use the hardware JPEG encoder in an easy way.The example code is under git repository: a311d/jpeg_encoderStep 1:
2021-04-19 11:17:22
293
原创 xHiveAI-A311D:How to resize or convert the picture color space
A311D SOC integrated the dedicated hardware module to proceed the picture resizing and picture color space conversion. It supported multiple channels processing in parallel and the maximum channels depend on the size of configured DMA buffer in kernel.Th
2021-04-19 11:16:22
221
原创 xHiveAI-A311D:How to use the hardware video decoder
A311D SOC support 1 channel 4K@60fps VP9,H265,H264 video decoding OR maximum 4 channels of 1080P@25fps video decoding in the codec format: H265,H264,MPEG-1/2/4,VP9,WMV。The libamldec.so is developed to use the A311D hardware video decoder in a convenient
2021-04-19 11:12:33
1576
1
原创 xHiveAI-A311D:Development Environment Preparation V2.0
Step 1:Prepare the host PC/Server with Ubuntu 16.04 or 18.04 installedStep 2:git clonehttps://github.com/apoidea-xhiveai/a311d.git git checkout v2.0 目录说明: toolchains:compiling toolchain for Amlogic A31...
2021-04-19 10:32:58
532
原创 xHiveAI-A311D:How to upgrade the firmware images
Please follow the steps to upgrade the firmware images on your board:Step1: Setup the tftp server in your Host PCStep2: Download the upgrade images from below linkageThe latest image:http://47.89.187.107:8080/fw/2021-04-06-170641/ubuntu_a311d_img_
2021-04-19 10:31:41
1000
原创 xHiveAI-A311D:Decide the document version
The development is still going on and the xHiveAI software architecture was upgraded accordingly. So there are 2 versions of the document in the system.Please follow the below steps to decide which version you should use: login the board shell ca
2021-04-19 10:30:28
285
原创 xHiveAI-A311D:How to use the development kit
第一步:请按照下面图中的指示来连接开发板:第二部:在开发主机上面配置串口开发主机可以为安装Ubuntu/Centos的Linux服务器,也可以为Windows PC机。如果使用 Linux服务器,请首先安装minicom软件,然后配置minicom到对应的串口:Ubuntu机器:sudo apt-get install minicomCentos机器:sudo yum install minicom lrzsz如果能够识别串口,在系统的/dev目录下会找到设备:ls ...
2020-12-02 17:26:45
1386
原创 xHiveAI-A311D:USB的使用
开发板上有1个3.0 USB和3个2.0 USB接口。1、USB3.0 测试1.1、插入u盘到3.0接口//查看u盘信息 fdisk -l /dev/sd*root@Apoidea:~# fdisk -l /dev/sd*Disk /dev/sda: 117.2 GiB, 125829120000 bytes, 245760000 sectorsUnits: sectors of 1 * 512 = 512 bytesSector size (logical/physical)
2020-11-30 13:39:32
551
原创 xHiveAI-A311D:TF_CARD的使用
TF卡卡槽在板子背面(J403)处。1、插入TF卡,观察串口打印//串口有下面的打印 即识别tf卡成功root@Apoidea:/# [ 5668.085995@2] meson-mmc: card IN2、挂在TF卡root@Apoidea:/# mount /dev/mmcblk1p1 /mnt/tf/root@Apoidea:/# ls /mnt/tf/config jdjq.wav jdjq.wav-1 kernel rootfs.ext4 u-boot...
2020-11-30 11:19:36
341
原创 xHiveAI-A311D:SPEAKER的使用
板子上可以接两个喇叭。1、播音测试1.1、命令配置//导出p1_4端口echo 406 > /sys/class/gpio/export //设置为输出echo out > /sys/class/gpio/gpio406/direction //使能功放echo 1 > /sys/class/gpio/gpio406/value //关闭功放echo 0 > /sys/class/gpio/gpio406/value 1.2、查看音频输入设.
2020-11-30 11:13:03
2551
原创 xHiveAI-A311D:PHONE(MIC IN)的使用
使用3.5mm四级插头接入到耳机接口进行音频测试。1、测试播音1.1、配置命令//导出 p1_0 端口echo 402 > /sys/class/gpio/export//设置为输出echo out > /sys/class/gpio/gpio402/direction//使能耳机echo 1 > /sys/class/gpio/gpio402/value//关闭耳机echo 1 > /sys/class/gpio/gpio402/value1.
2020-11-30 10:54:10
438
原创 xHiveAI-A311D:GPIO的使用
测试板上有18个GPIO引脚(下图),测试时将18个GPIO分为两组,将相连的那组GPIO分别设置为一个输入和一个输出。参考命令:1、导出gpio//GPIO_01 - GPIO_16echo 378 > /sys/class/gpio/exportecho 379 > /sys/class/gpio/exportecho 380 > /sys/class/gpio/exportecho 381 > /sys/class/gpio/exporte.
2020-11-30 10:27:56
642
1
原创 xHiveAI-A311D:视频解码的实现
A311D芯片支持4K@60fps VP9,H265,H264视频解码或者最多4路1080P@25fps的多格式视频解码(H265,H264,MPEG-1/2/4,VP9,WMV)。xHiveAI-A311D实现了基于gstreamer的媒体处理平台,方便客户快速定制和开发自己的应用程序,而不必关心底层硬件的实现逻辑。第一节:快速体验视频解码1)将开发板与具有HDMI接口的显示器通过HDMI线连接2)上电后启动板子,登陆到shell3)执行下面命令 gst-launch-...
2020-11-12 11:12:21
1696
原创 xHiveAI-A311D:图片color space的转换和resize操作
A311D芯片集成了硬件模块来实现图片的color space转换和resize操作。可以实现多路并发地处理,最大处理路数取决于内存DMA buffer的大小(在kernel中配置)。目前板子配置的DMA buffer大小为256MB。按照下面的步骤来实现你的图片color space的转换和resize操作:1)请参考文章:开发环境准备2)请参照代码树/pic_convert/src 下的color_space.example来实现color space转换的功能 参考代码如下:...
2020-11-11 17:47:23
295
原创 xHiveAI-A311D:JPEG Encoder的实现
A311D芯片集成了一个硬件JPEG Encoder来将yuv格式的图片压缩成jpeg格式的图片。按照下面的步骤来实现你的jpeg encoder功能:1) 请参考文章:开发环境准备2) 请参照代码树/jpeg_encoder下的example.c来实现你的jpeg encoder代码 -- 因为只有一个硬件JPEG Encoder,对于多路视频解码后,调用jpeg encoder来生成jpeg图片的情况,配置参数中的config.multi_proc是用来决定多线程调用还是多进...
2020-11-11 16:20:08
724
原创 xHiveAI-A311D:H264 Encoder的实现
A311D芯片集成了一个H.264/H.265 encoder,可以配置它的bit rate,frame rate等参数。按照下面的步骤来实现你的第一个h264 encoder:1) 请参考文章:开发环境准备2)h264 encode的示例程序在目录 encoder下面 cdencoder/src make <正常编译后,会生成可执行文件:test_h264encode> scptest_h264encode root@&...
2020-11-11 16:14:14
671
1
原创 xHiveAI-A311D:开发环境准备
第一步:准备一台安装了Ubuntu系统的服务器,我这边的系统版本是16.04第二步:git clonehttps://github.com/apoidea-xhiveai/a311d.git git checkout v1.0 目录说明: toolchains:编译工具链第三步:准备编译工具链 export PATH=$PATH:<root directo...
2020-11-11 16:07:54
952
原创 xHiveAI-A311D:AI开发套件
群蜂技术的xHiveAI开发套件采用Amlogic的A311D六核(Cortex-A73x4 +Cortex-A53x2) 64位处理器,集成了多核MaliTM-G52 GPU,计算能力达到5TOPS@int8, 12.8GFLOPS@FP16 or 6.4 GFLOPS@FP32的NPU AI Neural network accelerator (NNA),性能优异。板载 HDMI接口,MIPI摄像头接口 ,WIFI/Bluetooth模块,声音相关接口。支持 Linux/Ubuntu 系统,开放源代.
2020-11-11 16:02:56
3854
2
Linux System Programming 2nd edition 2013
2014-01-07
Android技术内幕.系统卷
2012-02-08
Fundamentals of digital television
2010-07-12
空空如也
TA创建的收藏夹 TA关注的收藏夹
TA关注的人