IOS 实现以图查图 类似百度查图功能

根据“感知哈希算法”来实现的。如果不清楚可以自己搜索下。这不具体讲解。

实现如下（借鉴了网上高手的资料，在这不一一列举了）

//

// checkImage.m

#import "checkImage.h"

@implementation checkImage

+(NSString *)imageSourcePath{

NSString *path=[[NSBundle mainBundle] bundlePath];

return path;
}
//1.将图片缩小到8x8的尺寸, 总共64个像素. 这一步的作用是去除各种图片尺寸和图片比例的差异, 只保留结构、明暗等基本信息.
+(UIImage *)imageToSize:(UIImage *)image toSize:(CGSize)size{
UIGraphicsBeginImageContextWithOptions(size, NO, 0.0);
[image drawInRect:CGRectMake(0, 0, size.width, size.width)];
UIImage *newImage=UIGraphicsGetImageFromCurrentImageContext();
UIGraphicsEndImageContext();

return newImage;
}

//2.将缩小后的图片, 转为64级灰度图片.
+(uint8_t *)convertTo64GreyImage:(UIImage *)image{
int kRed = 1;
int kGreen = 2;
int kBlue = 4;

int colors = kGreen;
int m_width = image.size.width;
int m_height = image.size.height;

uint32_t *rgbImage = (uint32_t *) malloc(m_width * m_height * sizeof(uint32_t));
CGColorSpaceRef colorSpace = CGColorSpaceCreateDeviceRGB();
CGContextRef context = CGBitmapContextCreate(rgbImage, m_width, m_height, 8, m_width * 4, colorSpace, kCGBitmapByteOrder32Little | kCGImageAlphaNoneSkipLast);
CGContextSetInterpolationQuality(context, kCGInterpolationHigh);
CGContextSetShouldAntialias(context, NO);
CGContextDrawImage(context, CGRectMake(0, 0, m_width, m_height), [image CGImage]);
CGContextRelease(context);
CGColorSpaceRelease(colorSpace);

uint8_t *m_imageData = (uint8_t *) malloc(m_width * m_height);
for(int y = 0; y < m_height; y++) {
for(int x = 0; x < m_width; x++) {
uint32_t rgbPixel=rgbImage[y*m_width+x];
uint32_t sum=0,count=0;
if (colors & kRed) {sum += (rgbPixel>>24)&255; count++;}
if (colors & kGreen) {sum += (rgbPixel>>16)&255; count++;}
if (colors & kBlue) {sum += (rgbPixel>>8)&255; count++;}
m_imageData[y*m_width+x]=sum/count/4;
}
}
free(rgbImage);
return m_imageData;
}

//3.计算图片中所有像素的灰度平均值

+(uint8_t)avgGreyPixel:(uint8_t *)data{
int sum = 0;
for (int i = 0; i < 64; i++) {

sum += data[i];
}
uint8_t avg = sum/64;
return avg;
}

// 4.将每个像素的灰度与平均值进行比较, 如果大于或等于平均值记为1, 小于平均值记为0.

+(NSMutableArray *)compareGreyPixelToAvgPixel:(uint8_t *)imageData avg:(uint8_t)avgGrey{
NSMutableArray *endArr=[[NSMutableArray alloc] init];
for (int i=0; i<64; i++) {
if (imageData[i]>=avgGrey) {

int f=1;
NSNumber *ff=[NSNumber numberWithInt:f];
[endArr addObject:ff];
}else{
int f=0;
NSNumber *ff=[NSNumber numberWithInt:f];
[endArr addObject:ff];
}
}

return endArr;
}

//5.将上一步的比较结果, 组合在一起, 就构成了一个64位的二进制整数, 这就是这张图片的指纹.
+(NSMutableString *)compareWithGroup:(NSMutableArray *)arr{
NSMutableString *ms=[[NSMutableString alloc] initWithString:@""];

for (NSNumber *i in arr) {
int f=[i intValue];
[ms appendFormat:@"%d",f];
}

return ms;
}

//6.得到图片的指纹后, 就可以对比不同的图片的指纹, 计算出64位中有多少位是不一样的. 如果不相同的数据位数不超过5, 就说明两张图片很相似, 如果大于10, 说明它们是两张不同的图片.

+(int)compareToHamdDistance:(NSMutableString *)scoureImage current:(NSMutableString *)current{
int distance=0;

for (int i=0; i<64; i++) {
unichar s=[scoureImage characterAtIndex:i];
unichar c=[current characterAtIndex:i];

if (s!=c) {
distance++;
}
}

return distance;
}

@end

运行效果如图：点击选取照片
650) this.width=650;" src="http://img1.51cto.com/attachment/201303/222654991.png" border="0" alt="" />

选取第二行第三张照片去找相识图片

650) this.width=650;" src="http://img1.51cto.com/attachment/201303/222706888.png" border="0" alt="" />

找到以下几张相似图片

650) this.width=650;" src="http://img1.51cto.com/attachment/201303/222716502.png" border="0" alt="" />

650) this.width=650;" src="http://img1.51cto.com/attachment/201303/222727173.png" border="0" alt="" />

本人能力有限，如果bug欢迎拍砖。源码可去http://down.51cto.com/data/696207下载，或邮箱至478043385@qq.com免费索取。
本文出自 “做fashion的IT人” 博客，请务必保留此出处http://kyoworkios.blog.51cto.com/878347/1149733