Embedded vision system for mobile robot navigation

Abstract

We developed an embedded vision system that can accelerate the basic image processing functions for mobile robot navigation with compact hardware featuring low power consumption. The system is composed of a Digital Signal Processor (DSP) and a dedicated LSI for low-level image processing, specifically for spatial filtering, feature extraction, and block matching operations. The image processing LSI has a dedicated systolic array processor consisting of 64 processing elements to accelerate basic block operations for image feature calculation and correlation-based image matching. The power consumption is only 10 W, about one-seventh that of a typical Pentium 4 personal computer, but the processing speed for correlation matching is roughly three times faster than such a system. With this vision system, we implemented a stereo-visionbased navigation algorithm on our mobile service robot and performed a visual navigation experiment in a building hallway.

Development of the Embedded Vision System: A Survey

“…They collaborate to realize the needed functions. The collocations of processors include SIMD (Single Instruction Multiple Data) processor with micro-controller [19], [22] , DSP with FPGA [12]- [14] , FPGA with ANN (Artificial Neural Network) [18] , and so on. The cost of the EVS based on this architecture is higher than the singular processor architecture, as the design circuit is complicated.…”

“…An EVS was proposed in [12] , which is composed of a DSP and a dedicated LSI (Large-scale integration). Some low-level image processing is implemented by LSI, such as spatial filtering, feature extraction and block matching operations.…”

“…Figure 14 gives the results, green and blue points are feature points that match with the map, yellow points are feature points that do not match with the map, the red point is an estimated current position. The system operates at 86 MHz, and the power consumption is only 10 W. Figure 12: System architecture [12] Figure 13: Map and path data [12] Figure 14: Navigation experiment results [12] An EVS based on DSP and FPGA was described in [13] , where the DSP is used as the main processor and the FPGA is used as the coprocessor. Both the DSP and FPGA are driven in parallel for the execution of crucial parts of the vision algorithms.…”

An Embedded Real-Time Red Peach Detection System Based on an OV7670 Camera, ARM Cortex-M4 Processor and 3D Look-Up Tables

“…Additionally, in [8], an intelligent embedded vision system to implement different image filters and perform image correlation and transformation at up to 667 frames per second was proposed. In [9], an embedded vision system was proposed specifically for mobile robot navigation and low power consumption. This embedded system accelerates the basic image processing algorithms required in a mobile robot application, such as low-level image processing, spatial filtering, feature extraction, and block matching operations.…”

Edge detection using structured laser pattern and vision for mobile robot navigation

“…However, unambiguous evaluation of the environment is computationally costly since most vision systems (pinhole camera) operate by severely collapsing the 3D world into 2D [5]. While it does not pose a serious limitation on mobile robots that can accommodate suitable vision hardware [6] [7][8] [9], it is a challenge for mobile robots with small footprint to equip sophisticated image processing hardware for real time application [10] [11] [12][13] [14]. Some methods to alleviate computational overheads include using biologically inspired compound-eye like vision systems [5][15] [16].…”