Deep Learning AI Imaging + Photonic System
Mechatronic Control + Agricultural Applications
CiPAR has four core technologies and cooperates with the OEM in the agricultural vehicle industry. The first technology is intellectualization of electric farming unmanned vehicles. The second is the controlling and positioning of an unmanned vehicle that executes repeated labor tasks and drives itself on a pre-planned route. The third is image recognition and stereo modeling that utilizes the features of plants provided by an expert system comprising agricultural research institutes. The fourth is agri-photonic equipment, including robot arms and laser systems that perform the required functions on the target objects.
Multi-Functional Agricultural Robots: An agricultural unmanned ground vehicle (UGV) automatically patrol a farm using a combination of smart agricultural machinery for pest control and quality agriculture. After acquisition of images & data, the results of the analysis can help smart agricultural machinery perform farm jobs.
By constructing a 3D model of the farm using image sensors, the SLAM technique can be used to do the real-time positioning and plan the path the vehicle will go through.
The control system module drives the vehicle along the planned path. The object identification system can instantly identify obstacles, and the obstacle avoidance technology guides the vehicle to bypass obstacles and return to the original route. In order to make the electric vehicle more adaptable to the topography, adjustable modular and 4WD techniques are developed to create higher degrees of freedom.
ØSmart Agricultural Machinery
The deep learning technology based on an agricultural database can also be realized in an embedded system, where real-time recognition is achieved via high speed computation. The databases for pests and weeds have been constructed, and the accuracy related to recognizing mature pests, weeds, and fruit thinning are good for practical use. The RS485 communicates with the coaxially corrected laser scanning system, which can be useful in terms of pest control. In addition, an IoT module is integrated to execute data analysis. Plant diseases can be effectively predicted by the analysis of the all-day data stored in the cloud. Precise fertilizing schedules and environmental control can be achieved for disease prevention and treatment.
Unmanned Aerial Vehicle (UAV) Precision Positioning Combined with Economical Hyperspectral: To monitor plant growth and fruit ripeness, the UAV is equipped with economical hyperspectral imaging to obtain normalized difference vegetation indexes or spectral images. The object position retrieved by the drone will then cooperate with the back-end analysis to work on the harvest route planning.