中文说明：条码识别，MATLAB7.0下运行，仅支持一维条码[1] 剪切图像，并作旋转一定的角度。(2007-11-1，9:42)；[2] 剪切出第一个条形块。 (2007-11-2)；[3] 测试左斜的条形码，得出来的数据基本正确。(2007-11-2)[4] 采用两种方式对条形码进行锐化，一种噪声比较大，另一种噪声比较小；[5] 重新设计算法，先是进行边缘检测，提取条形码区域；然后对条形码区域进行锐化、光滑等预处理，再做识别；经过比较，采用prewitt边缘检测方法比较好。[6] 对12张图进行剪切，对剪切函数不断的修改！(2007-11-08)[7] 完成条形码区域的提炼，比较拉普拉斯和直方图均匀化增强方法，发现直方图均匀化效果比较好！(2007-11-09)[8] 比较矩不变量和大津二值化算法，大津算法效果比较好！->改进小区域剪切方法，先剪切上下后剪切左右！->增加二次剪切，解决旋转后造成的冗余量。(2007-11-09)[9] 调节pixelV门槛值，以及改正imagefilter函数；修改剪切函数，测试33张图片，除了w1/w2/w5/w25/w40/w41/w42因亮度不均、w39_w有些模糊等8张图片不能正确剪切出条形码外(同时说明大津算法值得改进)，其余的可以正确剪切出条形码，以待进一步处理和识别。2007_11_10[10] 字符能够识别，但是12个数据会有个把数据是错误的；主要是前面的预处理没有处理好，
English Description：
Bar code recognition, matlab7.0 run, only support one-dimensional bar code [1] cut image, and rotate a certain angle. (2007-11-1, 9:42); [2] cut out the first bar block. (2007-11-2); [3] test the left oblique bar code, and the data is basically correct. (2007-11-2) [4] sharpen the bar code in two ways, one is more noisy, the other is less noisy; [5] redesign the algorithm, first carry out edge detection, extract the bar code area; then sharpen and smooth the bar code area, and then do recognition; after comparison, it is better to use Prewitt edge detection method. [6] Cut 12 pictures and modify the cut function continuously! (2007-11-08) [7] complete the extraction of bar code region, compare the Laplacian and histogram homogenization enhancement method, found that the histogram homogenization effect is better! (2007-11-09) [8] comparing moment invariants and Otsu binarization algorithm, Otsu algorithm is better! -&Gt improve the small region cutting method, first cut up and down, then cut left and right! -&The second shearing is added to solve the redundancy caused by rotation. (2007-11-09) [9] adjust the pixel V threshold, and correct the imagefilter function; modify the clipping function, and test 33 images, except for W1 / W2 / W5 / W25 / W40 / w41 / W42 due to uneven brightness and W39_ W some fuzzy and other 8 pictures can not be correctly cut out of the bar code (at the same time, it shows that the Otsu algorithm is worth improving), the rest can be correctly cut out of the bar code for further processing and recognition. 2007_ 11_ 10 [10] characters can be recognized, but one of the 12 data will be wrong; the main reason is that the previous preprocessing is not handled well,