Research On Multi-spectral Image Acquisition Control And Data Processing Technology

Multi-spectral images are images of multiple spectral bands acquired by the camera through the tunable filter of the same object.Compared with ordinary images,they contain more physical information,which is ideal for use in scientific research.Therefore,in agricultural production,mineral exploration,space exploration,and other fields have a wide range of applications.In this paper,for multi-spectral image acquisition control and data processing,the following work is carried out:1.In order to solve the problem of being susceptible to external interference when using Unmanned Aerial Vehicle(UAV)for multi-spectral image acquisition.It is proposed to block the disturbance of the UAV through the pan-tilt-zoom(PTZ),and use Firefly-RK3288 Reload as the experimental platform.After the experiment,the software is transplanted to the custom control panel.In order to enable the server to control the PTZ and know the real-time status of the PTZ,the data format designed to communicate with the server through the SOCKET.Write the I2 C driver so that the development board can get the data in the MPU6050 sensor through the I2 C interface.Since the collected acceleration and angular velocity data are noisy,the Mahony complementary filtering algorithm is used to fuse the two parts of the data to obtain more accurate PTZ attitude data.However,since the YAW angle cannot be corrected by the accelerometer,an improved Mahony algorithm is proposed to correct the deviation according to the characteristics of the YAW angle.Experiments show that the improved Mahony algorithm can effectively suppress the integral drift of the YAW angle.The cascading PID algorithm is used to control the motor through the PWM and GPIO interfaces.Compared with the single-loop PID,it can improve the control quality of the system.2.Aiming at the problem that the multi-spectral image is large and the transmission time is too long,the FPGA development board is used as the experimental platform to compress the multi-spectral image.After the test is passed,the program is transplanted to the customized board.In order to solve the USB communication problem between the server and the FPGA development board,the firmware program and driver of the USB chip are designed and tested,and experiments show that the two can transmit data correctly and efficiently.Encapsulate the DDR2 IP controller,making it easier and faster to read and write data.According to the inter-spectral correlation and spatial correlation of multi-spectral images,it is proposed to use the combination of reversible component transform and JPEG-LS to eliminate spatial redundancy and inter-spectral redundancy of multi-spectral images,and effectively reduce the volume of multi-spectral images.3.Test the PTZ control module and image processing module.Tests have shown that the PTZ can suppress interference in a short period of time and enhance system stability.The image processing module can perform a lossless compression of multi-spectral images with high quality and has a higher compression ratio and less time than JPEG2000.