Design And Implementation Of A Miniature Vision Sensing System For Waterborne Walking Robots

Visual sensing system is very significance for Water Strider Robot to realize water quality monitoring, environmental monitoring and military reconnaissance. The emerging of Water Strider Robot using circular footpads makes robot airborne weight up to 53g, so that more hardware choices are available for system implementation and the foundation will be laid for Water Strider Robot to realize visual sensing. Due to the limited load capacity of Water Strider Robot, therefore, the airborne micro vision system which can be applied to Water Strider Robot must meet the requirements:small size, low power consumption, light weight and so on. This paper has important theoretical research significance and practical value in micro visual system integration and compressed sensing measurement matrix research which is suitable for low power consumption microprocessor.This paper research related issues about micro vision sensing system, the main study contents are as follows:1. The paper reviews and analysis the research achievement on existing micro vision sensing system in domestic and foreign, Use CMOS image sensor OV7670 and dynamic memory chip AL422B as image acquisition equipment, low-power processor MSP430F149 as the main control chip, RF chip nR24L01+ as the wireless communication module architecture design and realize a miniature vision sensing system. Experimental results show that the micro vision sensing system fulfill expected requirements.2. Discussed the problems about exciting traditional wireless transmission method applied to the micro vision system, introduced compressed sensing theory, and focuses on how to construct measurement matrix easy to hardware implementation. An improved Partial Hadamard measurement matrix construction method is proposed for the reconstruction performance excellent Partial Hadamard observation matrix, Which is named as sequence Partial Hadamard measurement matrix in this paper name it sequence Partial Hadamard measurement matrix. Simulation results show that the reconstruction results are further improved by the proposed sequence Partial Hadamard measurement matrix at a lower sampling rate. Its matrix structure is relatively simple and has recursive characteristic, and it can be easily implemented in hardware, but the measurement matrix sampling signal length is limited, the length of signal must be N=2K K=1,2,3..3. In order to overcome the signal applied length limitation, on the basis of study on Pseudo-Random Sequence, this paper proposes a method to construct measurement matrix by using Pseudo-Random Sequence, namely recycled pseudo-random sequence measurement matrix, in which consists of the recycled M sequence and recycled Gold sequence measurement matrix, and presents the specific measurement matrix construct process. Though it is inferior to the sequence Partial Hadamard measurement matrix, the reconstruction of the recycled pseudo-random sequence measurement matrix excels Gaussian random measurement matrix, and also overcomes the sequence Partial Hadamard measurement matrix’s 2k-length limitation. This kind of measurement matrix is easy to be implied into hardware. Besides, it can avoid the uncertainty of the random matrix and also overcome the wasting storage of random matrix. Therefore, they have good practical application values.