Based on DJI drones, the PSDK provides a complete set of solutions.
我公司基于上云api开发了大疆系机场飞行控制平台。
具体介绍如下：
基于大疆上云API开发的高可用SAAS系统。
支持机场1与机场2的接入与控制。
支持遥控器（pilot）上云。
支持成智喊话器。
支持面状航线素材生成正射影像并使用。
支持搭载ESDK、PSDK开发的边缘计算控制盒。
支持内网部署（已有实际部署到公安内网案例）。
支持样本标注、算法训练、模型使用等全流程AI目标检测解决方案。
支持AI识别到的异常数据上报和展示。
支持航线的在线编辑，KMZ 的导入及导出。
全平台低延时，视频直播毫秒级延时，AI叠加视频秒级延时。
专业团队保障更新迭代以及bug修复。
欢迎注册试用：/yunmu/index
使用过程中有疑问或问题，及商务合作请在 个人中心-》联系我们，获取联系方式
我们有一整套的无人机解决方案，从设备端到服务端，使用5G链路进行报文传输，设备端即硬件盒子加上psdk代码，服务端以jar包的形式提供，接口清晰明了，使得用户可以快速的开展起业务，并不需要关注实现过程，简单20分钟的开发，就可以通过一条简单的http请求使飞机起飞。
盒子内置psdk代码实现了图传、手动飞行、云台控制、航线控制、飞行数据回传等等基础功能，使用私有协议与后端进行通信，经过长期测试，延时低（指令毫秒级，图传受网络影响波动较大，我们会根据网络速度实时的自适应推流分辨率和码率，以保证延迟小于500ms，稳定200ms），稳定性好。
盒子询价采购和基于大疆无人机的sdk开发，欢迎发邮件交流：zhangxiaolan940601@gmail.com
ps: 其实这套盒子不一定非要用在飞机上，一些无人车、船、枪机或球机摄像头的前端识别、一些物联网设备等等场景都可以使用，内置了大疆psdk的环境便于二次开发，系统是ubuntu系统，方便进行一些轻算法的开发工作，甚至可以采购回去当做微型电脑用，等等
Here is a translated version of the document generated based on ChatGPT. If there are any ambiguities, feel free to email us for further discussion: zhangxiaolan940601@gmail.com
We are a team specializing in the development and application of DJI drone PSDK and OSDK. Our team is based in China, and we have successfully produced a hardware device (referred to as "box" below) as a replacement for the official one. The box consists of an NX board and a 5G chip, featuring excellent engineering design to ensure stable hardware quality and powerful performance.
The box is based on the NX board and has a certain computing power, capable of running some lightweight models.
The box is compatible with DJI drones such as M300, M30t, M350, which are widely used in various industries.
We offer a complete drone solution, from the device side to the server side,Using a 5G link for message On the device side, the hardware box is combined with PSDK code, while the server side provides the functionality in the form of a jar package. The interface is clear and concise, allowing users to quickly develop their own businesses without worrying about the implementation details. With just a simple HTTP request, the aircraft can take off within 20 minutes of easy development.
The built-in PSDK code in the box provides basic features such as image transmission, manual flight, gimbal control, flight route control, flight data feedback, and more. It communicates with the backend using a proprietary protocol based on Java-Netty. After extensive testing, it has low latency (instruction-level in milliseconds, while image transmission can be affected by network fluctuations). We dynamically adjust the streaming resolution and bitrate based on network speed to keep the delay below 500ms and maintain stability at around 200ms.
We have also developed an app based on MSDk, using the same protocol as PSDK, which supports all the basic functionalities of flight control.
Additionally, we have developed a flight control system called "Airframe Library" based on DJI's cloud API, which supports almost all the functions provided by DJI.
We welcome anyone interested to contact us for further discussion. Attached below is the image of our product's appearance.
The box is based on the NX board and has a certain amount of computing power, allowing it to run lightweight models. Feel free to email us for further discussion at zhangxiaolan940601@gmail.com.
PS: Actually, this box doesn't necessarily have to be used on an aircraft. It can be utilized in various scenarios such as front-end recognition for unmanned vehicles, boats, gun or dome cameras, as well as IoT devices, and more. It is equipped with the DJI PSDK environment for easy secondary development. The system is Ubuntu-based, facilitating the development of lightweight algorithms. It can even be purchased and used as a mini-computer, and so on.
