Design and Application of Blackfin Processor in Industrial Image Processing
Because the processor (PC) can provide a low-cost, small-size, scalable image processing system, they are more attractive than other similar systems with higher power consumption and higher prices. The clock frequency of the company's processor series is, including an enhanced structure that facilitates efficient data movement and processing. This structure is similar to the double multiply-accumulate unit and provides a high-speed parallel interface for video and data channels. The following will introduce some commonly used image processing and video processing algorithms to illustrate how these characteristics of the processor play their role in today's image processing systems.
1Blackfin processor has many features suitable for image applications
2 The Blackfin processor has two units, so it can generate two output points in one cycle, which is equivalent to doing one × convolution every cycle.
As shown in 1, the level near the core processor is the fastest, but usually the capacity is too small to have a large image cache. This is one of the reasons why direct memory access is so important. In the mode, the data frame can store and retrieve data from the fast memory, and at the same time, the data from the video peripherals is loaded into a large off-chip buffer. The controller works independently of the core processor, and the core cycle is only used to provide an interrupt when the transfer is complete. Each algorithm described below utilizes some framework-based types to achieve the best possible performance. These algorithms are chosen because they are suitable for a wide range of applications.
3 × convolution
3 × mask, but the mask cannot be too small, otherwise the edge of an image will not be detected.
2 shows a matrix: an input matrix, a × mask matrix and an output matrix. By properly inputting data to it, the processor unit can process two output points at a time in one processor cycle, and simultaneously read multiple data that are executed in parallel with the operation. This method allows efficient calculation of dual output points for each cycle repetition period or per pixel period.
3 × convolution kernel to approach the horizontal and vertical edges. The first matrix detects changes in vertical edges, while the second matrix detects changes in horizontal edges.
The Blackfin processor adopts the method of preserving the larger value of the two gradients of the × pixel image, and it takes about one frame to process. At the same time, it uses a two-dimensional data frame based on to access data from the fast processor memory.
Several features of Blackfin processor
On Blackfin processors, even if the processor has level channels, the processing of a pre-branch can be as small as one cycle. This is considerable for processors with signal processing capabilities, and this obviously helps to shorten the calculation time in this case.
The signal processing function of the Blackfin processor can complete two multiplication operations in one cycle, and access the memory twice to read the accumulated table value. Facts have proved that this method can complete the fixed-point Hough transform well, and the effect is equivalent to floating-point calculation.
// loop over the values
a0 + = a1; // add the results
The Blackfin processor provides a vector instruction to find two maximum values ​​from two operand pairs. This processing method can effectively find an N
In addition to the MAC and arithmetic logic unit, the processor has four additional ones that can be applied to single-cycle instructions. These four can process four sets of bytes at the same time such as adding, deleting and averaging. Very useful in including motion estimation between image frames.
The 8-bit subtraction absolute value accumulation instruction subtracts four pairs of values, takes the absolute value of each difference, and accumulates each result into an accumulator to discriminate the motion of the object. In short, the processor has a variety of methods to effectively detect the motion of adjacent image frames.
FFT
(FFT) is a fast algorithm for calculating discrete Fourier transform. When calculating two-dimensional data, its main uses include fast convolution filtering, fast correlation, image enhancement and object recognition. × The two-dimensional size of the image should also be ×. The rotation factor is usually calculated before the running time.
FFT algorithm efficient bit inversion and butterfly addition and subtraction instructions. In order to complete the two-dimensional bit inversion, the × input image is expanded into a one-dimensional vector of size 2, mainly because the transposed matrix generated by the one-dimensional vector of bit inversion is equal to the two-dimensional bit inversion.
The number of cycles required for 16 × composite 2D includes overhead. This code can actually be used to calculate real numbers by setting the imaginary part of the input array element to zero. For a more efficient real number execution method using complex code, two real number matrices can be packaged as a complex input in two dimensions of the complex number. This method, called "packaging and packaging" or "mass production", requires back-end processing to separate the output and two images to transform. But this is usually not a problem in fast convolution and fast correlation, because the two transformations always need to be calculated.
The architecture of the Blackfin processor helps to make it a very useful processor in industrial image processing systems. In addition, the company also provides a common "image processing toolbox" kernel for processor-based image processing and analysis. The latest sample—integrating Ethernet, controller, and interface on a single chip enhances peripheral device functions. With these tools, it will greatly expand the range of industrial image processing applications.
All-Integer Hough Transform: Performance EvaluaTIon, "
Numerical Recipes in C: The Art of ScienTIfic CompuTIng 2nd EdiTIon, Chapter 12. New York: Cambridge University Press 1999.
The 10KV/11KV medium voltage variable speed drives are mainly used to drive various high-voltage motors, such as power stations, metallurgical industry, petrochemical, cement manufacturing, water supply and sewage treatment, mining, papermaking, etc. FGI`s high voltage drives adopt modular design, multi-stage power unit series technology, with low input harmonic content and high power factor. The Medium Voltage Inverter is small in size, easy to install and debug, and easy to maintain.
10Kv And 11Kv Medium Voltage Drives
Vfd Inverter,High Voltage Inverter,High Voltage Power Inverter,10Kv And 11Kv Medium Voltage Drives,High Watt Voltage Converter,Medium Voltage VFD Specification
FGI SCIENCE AND TECHNOLOGY CO., LTD , https://www.fgi-tech.com