Convolution
In mathematics and, in particular, functional analysis, a convolution is a mathematical operator that transforms two functions f and g in a third function that in some sense represents the extent to which overlap and transferred fy version g. inverted A convolution is a very general moving average, as can be seen if one of the functions we take as the characteristic function of an interval.
The convolution of and is denoted. Integral is defined as the product of two functions after one is reversed and displaced a distance τ.
The range of integration depends on the domain over which the functions are defined. In the case of a finite integration range, f and g are often considered as extended, regularly in both directions, such that the term g (t-τ) does not imply a violation in the range. When we use these domains periodic convolution is sometimes called cyclic. Of course it is also possible to extend with zeros domains. The name used when we stakes in these domains "zero-extended" or the infinite is the linear convolution, especially in the discrete case we will present below.
If X and Y are two independent random variables with probability density functions f and g, respectively, then the probability density of the sum X + Y is given by the convolution f * g.
For discrete functions you can use a discrete form of convolution. That is,
When we multiply two polynomials, the coefficients of the product are given by the convolution of the original sequences of coefficients, in the sense given here (using extensions with zeros as mentioned).
generalize the previous cases la convolución puede ser definida para cualesquiera dos funciones cuadrado-integrables definidas sobre un grupo topológico localmente compacto. Una generalización diferente es la convolución de distribuciones
Mediante la convolución calcularemos la respuesta de un sistema (y(t)) a una entrada arbitraria (x(t)).
Dos condiciones para realizar la convolución:
• Sistema LTI.
• La respuesta al impulso del sistema es h(t).
Basándonos en el principio de superposición y en que el sistema es invariante en el tiempo:
Si L{δ (t )} = h( t ) L {K * δ ( t - t0)} = K * h( t – t0).
Una arbitrary input signal x (t) can be expressed as an infinite train of pulses. To do this, we divide x (t) in rectangular strips of width and height ts x (k ts). Each strip replaced it with an impulse whose amplitude is the area of \u200b\u200bthe strip: ts * x (kts) × δ (t-ct).
Convolution Theorem
The convolution theorem states that under certain circumstances, the Fourier transform of a convolution is the product of the transformed point. In other words, convolution in one domain (eg time domain) is equivalent to the dot product (or internal) in the other domain (ie the spectral domain). Sean
two functions f and g whose convolution is given by f * g. (Note that the asterisk denotes convolution in this context and no multiplication is sometimes also used the symbol). Is the operator of the Fourier transform, so that and are the Fourier transforms of f and g, respectively. Then
where • indicates dot product. Also arguable that:
Applying inverse Fourier transform, we can write:
.
Using Convolution and related operations are found in many applications of engineering and mathematics.
• In statistics, as stated above, a weighted moving average is a convolution. • In theory
of probability, the probability distribution of the sum of two independent random variables is the convolution of each of their probability distributions.
• In optics, many kinds of "spots" are described in convolutions. A shadow (eg the shadow on the table when we hand between it and the light source) is the convolution of the shape of the light source that creates the shadow and the object whose shadow is being projected. A blurry photograph is the convolution of the right image with the blur circle formed by the iris diaphragm.
• In acoustics, an echo is the convolution of the original sound with a function that represents the various objects that lo reflejan.
• En ingeniería eléctrica y otras disciplinas, la salida de un sistema lineal (estacionario o bien tiempo-invariante o espacio-invariante) es la convolución de la entrada con la respuesta del sistema a un impulso (ver animaciones).
• En física, allí donde haya un sistema lineal con un "principio de superposición", aparece una operación de convolución.
Tipos de Convolución
• Convolución Discreta
Cuando se trata de hacer un procesamiento digital de señal no tiene sentido hablar de convoluciones aplicando estrictamente la definición ya que solo disponemos de valores en instantes discretos de tiempo. Es necesario, pues, a numerical approximation. To perform the convolution between two signals, will assess the area of \u200b\u200bthe function. For this, we have samples of both signals at the instants of time, we call y (where n and k are integers). The area is therefore
The discrete convolution is determined by a sampling interval t = 1:
• Circular Convolution gT
When a function is periodic with a period of T, then the functions f such as f * gT exist, their convolution is also periodic i equal to:
Where is chosen arbitrarily. The sum is called a periodic extension of function f.
If gT is a periodic extension of another function, g, then f * gT is known to be circular, cyclical, or periodic convolution of fi g.
method to calculate the circular convolution:
We have 2 circles, one outside and one inside. We turn the inner circle i adding their values. If the two circles have different sizes, then the smallest we add "0" at the beginning, end or the beginning and end.
[L> = L1 + L2-1]
properties of the different properties of convolution operators are 12
Commutativity
Note: This property may be lost if not asked to "turn around" to a function. Associativity
Distributivity
Associativity with scalar multiplication
For any real or complex number a. Derivation rule
where Df denotes the derivative of Buddha, in the discrete case, the difference operator
. Convolution theorem
where denotes the Fourier transform of f. This theorem is also true with the Laplace Transform.
0 comments:
Post a Comment